WATER ACTIVITY: aw

Research Department

TECHNICAL BULLETIN
I Editor-Donald Dubois Volume II, Issue 3 March, 1980
I
I I

WATER ACTIVITY (aw) - NEW INTEREST IN AN OLD CONCEPT

Art Davis, Ph.D.
American Institute of Baking

INTRODUCTION solution), the relative humidity would no longer be 100%.

The Proposed Good Manufacturing Practices Due to the interaction between the salt and liquid water,
published June 8,1979 repeatedly refer to measurement less water would evaporate and the relative humidity
of water activity (aw) as a control or monitoring value to would be 96.7%.
be used in determining the safety (from a microbial This example illustrates two points. First is the
growth standpoint) of foods. Although aw is not a occurrence of a change in the “freedom” of the water to
particularly new concept in the food industry, its evaporate due to mixing with salt. Second is the ability to
inclusion in the proposed GMP’s will make a basic observe this change by measuring the relative humidity
understanding of the concept important to many of a fixed volume of air with which the solution has been
persons not previously concerned with its meaning allowed to come into equilibrium.
and/or measurement.
Relative Humidity and a,:
DEFINITION The actual conversion of equilibrium relative humidity
According to Subpart A 110.3 (p) of the proposed (ERH) to aw is rather simple. For the original pure water
GMP’s, “Water Activity (aw) is a measure of the free sample with an ERH of lOO%, the a, is 1.00. For the salt
moisture in a product and is the quotient of the water solution with an ERH of 96.7%, the a, is .967. Thus, ERH
vapor pressure of pure water at the same temperature”. and a, are different by a factor of 100. The important
As suggested by this definition, water in association with point is that the use of relative humidity to determine a,
other substances does not behave in the manner of pure is dependent on the substance measured being in
water, and one measure of this difference is vapor equilibrium with the atmosphere in which the relative
pressure which may be expressed as a ratio and called humidity measurement is taken. This equilibrium
a,. requires time to occur and, as will be discussed in the
measurement systems section, is one of the major
Relative Humidity: problems with aw determinations.
The most practical way to determine a, values is to
make use of the relationship between relative humidity
Water Freedom, aw, and Microbial Growth:
and a,. Relative humidity is the ratio between the
To return to the first point made by the salt solution
amount of water vapor in the air and the maximum example, the change in the “freedom” of water to
amount the air is capable of holding at that temperature. evaporate as a result of its association with other
If a fixed volume of air is exposed to pure water, it will, substances is of major interest in food preservation as
after a period of time, reach 100% relative humidity. This this same change in “freedom” is reflected in the
means that sufficient water has evaporated to saturate availability of water for utilization by microbes.
the air with water vapor and that further additions of The ability to measure this availability provides an
water vapor will result in condensation of liquid water. immediate insight into the possibilities for microbial
If the fixed volume of air required 100 grams of water in spoilage which is much more accurate than knowing only
the form of vapor to become saturated and achieve 100% the actual quantity of water in a food. For example,
relative humidity, then an equal volume of air at the same honey averages 17.2% moisture (Pyler, 1973) with an a,
temperature which contained only 50 grams of water of about .60. It is not particularly susceptible to microbial
vapor would have a relative humidity of 50%. If the growth other than an occasional yeast specifically
original volume of air was exposed, not to pure water, adapted to grow at low aw levels. In the case of flour, a
but instead to a solution containing 58.44 grams of salt in 17% moisture level corresponds to an aw of
one liter of water (known chemically as a one molal approximately .78 (Bushuk and Winkler, 1957) which is

The ~nfotmat~on contafned I” this bulleton IS true and accurate to the best of our knowledge However. stnce condittonsare beyondourcontrol. nothmgconta~ned herein should beconstrued
as a recommendation. guarantee. or warranty. ather expressed or lmplaed by the Amewan lnstltute of Saklng Nather should the mentwn of regfstered brand names be construed as an
endorsement of that product by the American InstWte of Saklng Matenal contained !n this bulletin copynghted. 19SO. by the American lnstltute of Baking.
well within the growth range of many the energy required for vaporization, All of the above a, levels are
molds and yeasts and at least one additional energy is needed to based on other factors known to
bacteria (Troller and Christian, remove the water from its associa- affect microbial growth, such as
1978). tion with the product before it can be temperature, salt concentration, pH,
vaporized. As more water is and oxygen availability, being at an
STATE OF WATER IN FOODS removed, this additional energy optimum level or condition.
Obviously, there are differences requirements is increased to remove Additionally, the a, levels given
in the degree of association be- water in closer association with the represent survival or minimum
tween water and other materials product. As the monolayer is growth points, and generally a
in food products and these associ- reached in the water removal somewhat higher a, would be
ations are responsible for changes process, the curve is nearly flat and required for the microorganism to
in the availability of the water for great effort is required to remove the achiti;e its full potential growth rate.
metabolic use by microbes. monolayer.
This is, of course, not a complete
Monolayer: picture of drying technology.
RELATING a, AND WATER
Water in association with other However, it serves to illustrate the
CONTENT - ISOTHERMS
materials in bakery foods is generally Water activity is, of course,
various degrees of association
classified in one of three states. The related to the total water content of
between a product and the water it
first layer of water surrounding the the product. However, this
contains.
material and in intimate contact with relationship is different for each
it is referred to as the monolayer. aw AND MICROBIAL GROWTH product and will also vary with
This is theoretically a one molecule Water existing in the area between temperature. Additionally, relatively
thick layer of water molecules the monolayer and “free” water is of minor formulation changes, while
surrounding the material. Monolayer interest to food processors because, not affecting total water content,
water is probably not freezable nor for each food spoilage micro- may have significant effects on a,.
available for use in microbial organism, there exists a point of The relationship between a, and
metabolism. This water can only be water/product association within total water content for a given
removed with difficulty in drying this area where the microorganism is product is consistent enough that
processes and its removal generally unable to utilize the water in its the proposed GMP regulations allow
results in an increase in oxidative metabolism and thus is unable to the monitoring of a, through
rancidity type reactions. From a food survive. For bacteria, this level of measurement of total water content.
processing standpoint, this layer association corresponds to an aw of This is accomplished through the
may be assumed to exist as its .85 or less. For most bacteria to construction of an isotherm relating
removal is difficult and upon reproduce at an appreciable rate the two quantities (aw and total
exposure to any moisture source, it requires an aw of .90 or greater and water) at a given temperature.
will be quickly reestablished. many of the common problem The construction of an isotherm is
bacteria from a food standpoint illustrated in Figure 1 and shows
Intermediate and Free Water: require water activities greater than several interesting points. Most
The second state of water in .94. However, there are bacteria evident is the area towards the
association with other materials is a which can survive at .85 a,. center of the curve where, in fact,
much less definite area which Molds do not survive well at water two curves are shown. This
includes water from the next layer activities below .75, except for one or divergence of the relationship
outside the monolayer on out to two varieties which can survive to .70 between water content and water
where the alterations in properties a,. Yeasts have been found which activity is due to hysteresis which
due to the association are so slight can survive to aw levels of around results in different water activity
that the water behaves nearly .62. However, most are unable to values for a given water content,
identically with pure or “free” water. grow below .75 a,. depending on whether the water
Water which is indistinguishable level was reached by drying the
from pure or “free” water comprises Figure I product from a higher moisture or
the third state. wetting it from a lower moisture
Observations of a normal product level. The effect is variable from one
drying curve provides a graphic type of product to another and may
demonstration of the various states change in magnitude within a
of water in the product. With a product, depending on the tempera-
constant rate of energy input, the ture used when constructing the
rate of moisture loss will be linear so isotherm.
long as “free” water is being Although hysteresis is an inter-
removed. In this stage, only the esting facet of water activity
energy needed to vaporize water is research, it is not often much of a
required. As “free” water is removed problem to food processors. Most
and water from the second general food products consistently arrive at
state is reached, the drying &ve will their final moisure level either as a
begin to flatten as now, in addition to function of a drying or of a wetting

Page 2
 process, which is unlikely to vary. relative humidities of the various humidities greater than the
Thus, their a,/water content ingredients are different, it is likely equilibrium relative humidity of the
isotherm will be used only in the that they will begin to exchange material. By drying the atmosphere
adsorption (wetting or moisture moisture to achieve a balance which in a closed system such as a
gaining) or the desorption (drying or may be quite different than that package, these materials cause
moisture losing) mode, rarely both. present at packaging. This gain or other substances having higher
loss of moisture by individual equilibrium relative humidities to
ingredients may lead to lumping in lose moisture to the atmosphere.
Isotherm Preparation: the package and/or loss of function- Thus, the atmosphere in a package
Preparation of an isotherm is ality of some materials. It may also acts as a transfer agent during
relatively simple in concept. alter the possibilities for microbial moisture equilibration of the
However, some equipment is growth in specific ingredients. enclosed engredients.
needed and considerable care must A number of interesting studies
be taken to insure accurate results. have been conducted to evaluate METHODS OF
Essentially, the preparation of an MEASUREMENT
systems for predicting the final
isotherm for a product in a fixed Vapor Pressure Method:
moisture distribution with a mixture
temperature chambers containing of materials based on their Three basic systems have been
atmospheres of known relative used to measure a,. The most direct
equilibrium relative humidity and
humidities. is determination of vapor pressure
initial moisture content. Essentially
Selecting a range of relative through use of a device known as a
the equations take initial moisture
humities and monitoring the weight
content, equilibrium relative vapor pressure manometer. This
of the product exposed to each until
humidity, and percentage of total system is normally constructed on
the weight stabilizes will produce a
mixture for each ingredient and an individual basis within a labora-
series of product samples each in
determine the equilibrium relative tory. Although the equipment is not
equilibrium with a different relative
humidity of the entire system. From particularly expensive, it is essenti-
humidity. Determination of the total
this value, it is possible to go back to ally an all glass system and thus
water content of each sample will
an isotherm for each ingredient and rather fragile. Additionally, as noted
then allow construction of a graph
predict its moisture content at that by Labuza (1976) and Troller (1978),
with total water content on one axis
relative humidity. the vapor pressure manometer
and the relative humidity with which
As an example, Salwin and system is extremely sensitive to
that water content was in equilibrium
Slawson (1959) determined that in a temperature variations during use.
on the other.
The equilibrium relative humidity precooked dehydrated chicken stew
is, of course, easily converted to a, mixture at 72”F, the following Fett-Vos/Isopiestic Method:
and a curve connecting the graph moisture transfers occurred. A second system involves the
points should represent the isotherm Chicken (ERH l.l), with an initial measurement of water transferred
for the product at that temperature. moisture of 1.07% (as a percent of from the sample to a compound
For each bakery food as produced, solids), gained to a level of 3.18% in whose isotherm is well-known. This
there is a relative humidity at which it the presence of the other transfer is carried out in a closed
will neither gain nor lose weight. This ingredients. Potato dice (ERH 28.4) system and the weight of the
is the product’s equilibrium relative contained 5.88% water and lost recieving compound monitored.
humidity (ERH) as produced. If the moisture to 3.64%. Lima beans (ERH When an equilibrium is reached, this
product is later exposed to relative 7.7) increased from 3.68% to 3.90% weight is used to determine moisture
humidities higher than its equilibrium moisture. The moisture in the cream content and from the compound’s
relative humidity, it will tend to gain sauce base (ERH 18.8) decreased isotherm, an a, value is determined.
moisture and weight and, as from 3.49% to 2.54% and the chicken This system is known as the
expected, the aw will increase. The soup and gravy base (ERH 37.0) isopiestic method and is also called
moisture gain or loss will be limited to decreased from 2.49% to 1.19%. For the Fett-Vos method after the
the water content which is in each ingredient, the relationship investigators who developed it. The
equilibrium with the atmosphere to between its equilibrium relative Fett-Vos isopiestic method is
which the product is being exposed. humidity and the final equilibrium particularly suitable for work with
relative humidity of the system, materials having an aw above 0.85.
which was 7.6, determined whether
PREDICTING SHELF LIFE the ingredient would gain or lose Relative Humidity Method:
CHARACTERISTICS FROM a, moisture. For each ingredient, the The third system for measure-
MEASUREMENTS final observed moisture was within ment requires determination of the
The practical application of this 0.22?, of that predicted by the relative humidity of an atmosphere in
information comes when several investigator’s equations. equilibrium with the material of
ingredients are mixed, packaged, In general, materials with lower interest. This system is probably the
and expected to maintain quality equilibrium relative humidities tend most widely used and a number of
over an extended shelf life as in to associate more strongly with commercially available instruments
prepared doughnut, bread, or other water and are able to pull water out are based on relative humidity
bakery mixes. If the equilibrium of atmospheres having relative measurement.

Page 3
 Two different principles are used hygrometry is subject to aging of the 3. LABUZA, T.P., ACOTT, K.A.,
in the various instruments designed fiber which may, in addition, exhibit TATlNl, S.R., LEE, R. Y., FLINK, J.
to measure a, by determining some hysteresis effects of its own. AND MCCALL, W. Water Activity
relative humidity. One uses a fiber These problems may require Determination A Collaborative
attached to a carefully calibrated frequent recalibration and care to Study of Different Methods. J. Food
spring. The fiber is exposed to the insure that the instrument always Science 41:910, 1976.
atmosphere which is in equilibrium comes into equilibrium from the
4. LABUZA, T.P., L.N. KREISMAN
with the sample, gaining or losing same direction in terms of moisture
gain or loss. AND C.A. HEINZ. Evaluation ofthe
moisture until it, too, is in equili- Abbeon Cup Analyzer Compared to
brium. As a function of the moisture Electrical properties of immo-
the VPM and Fett-Vos Methods for
gained or lost, the tension the fiber bolized salt solutions can be altered
Water Activity Measurement.
placed on the spring varies. The by exposure to volatiles other than
Journal of Food Processing and
spring is also attached to a dial which water vapor. Samples containing
Preservation, 1131, 1977.
reads in a, units. This type of device glycols can contaminate the sensing
requires relatively frequent recali- device and result in inaccurate 5. PYLER, E. J., Baking Science and
bration due to changes in the fiber readings. Technology, Siebel Publishing
with time. With care and understanding, Company, 1973.
The second principle uses the accurate a, determinations can be
6. SAL WIN, HAROLD AND
change in electrical conductivity achieved with any of the systems
SLA WSON, VlDA. Moisture Trans-
which occurs in a salt solution as a mentioned. In addition to the
fer in Combinations of Dehydrated
function of moisture content. The methods mentioned, there are
Foods. Food Technology, 13:715,
salt solution is immobilized in a thin several less common systems in use
1959.
layer on a probe, which is then which are reviewed by Troller and
exposed and allowed to come into Christian in their book, Water 7. STOLOFF, LEONARD. Cali-
equilibrium with an atmosphere in Activity and Food. bration of Water Activity Measuring
equilibrium with the sample. By Specific comparisons of various Instruments and Devices: Colla-
calibration against known values, systems and instruments have been borative Study. J. Assoc. Off. Anal.
conductivity is related to relative carried out in at least two collabor- Chem., 61:1166, 1978.
humidity and aw. Instruments using ative studies. In both studies (Stoloff
1978, and Labuza et al., 1976) the 8. TROLLER, JOHN A. AND
this principle usually have a strip
systems were compared on both CHRISTIAN, J.H.B. Water Actiuity
chart recorder to provide a graphic
known a, salt solutions and on a and Food, Academic Press, 1978.
print of the equilibrium process
between the atmosphere and series of foodstuffs. Differences were
sample. This curve is useful in that its found in the abilities of the systems
trend can be observed to predict to measure a, and a review of these
when equilibrium will occur. When articles is probably a valuable
the initial steep curve comes to level exercise prior to selection of an a,
and has remained so for a time, the system for a particular application.
process of equilibrium can be
assumed to be complete and a SUMMARY
datum recording made. The measurement of water
activity and use of the results in the
food industry will undoubtedly
Problem Areas in a, increase when the proposed Good
Measurement: Manufacturing Procedures are
All three methods require esta- enacted into law. A clear under-
blishment of an equilibrium before a standing of the concept and
measurement is taken. This require- importance of aw measurements,
ment is a major problem in using aw as well as the abilities and limitations
as a rapid feedback process control of current systems, will be an
system. Additionally, each method important part of the food pro-
has inherent sensitivities which cessors knowledge. This knowledge
require an understanding of both the and understanding should contri-
basic concept of water activity and bute to a continued supply of high
the specific measurement system in quality and reasonably priced food to
use. the public.
The temperature sensitivity of the
vapor pressure manometer is one
REFERENCES
example previously pointed out.
1 BUSHUK, W., AND WINKLER,
The Fett-Vot method is consider-
C.A. Cereal Chem., 34:73, 1957.
ably less accurate at aw below .85
than at higher values. The fiber- 2. Federal Register, 44(112):33238,
spring-dial system known as hair June 8, 1979.

Page 4