INTRODUCTION where a* is the thermodynamic activity of water, 7* the

activity coefficient of water, and Xw the total mole fraction

Thermodynamics provides means for predicting and of water in the solution. 7e and T are in oK. Water's
correlating: 1) equilibrium freezitg points of foods; 2) molecular weight is 18.02. AII., is the average latent heat
food's equilibrium liquid water and ice contents as of fusion between Ts and T . The ideal gas law constant R i s
1 1.
functions of temperature during freezing; 3) heat-removal 8.314 kJ (kg mole)- K- The freezing point depression
requirements for freezing; and 4) effects of pressure and (To - T) is the same in K and "C. 7* is difficult to predict
ice crystal size on freezing points. and correlate for foods. Because aqueous solutions contain
At atmospheric pressure, the equilibrium freezing point both solvent water and water, which is bound to solute
of pure water, 76, is 273.L6K (0t or 32"F). To decreases molecules and acts as part of those molecules,t3l
by 0.0074K atm-t as pressure increases.tll Water often has
to be subcooled before the ice crystal nuclei form. Once N* - n* bn"
nucleation occurs, the temperature of the ice-water mix
an - X*": N*+Ns n*-bn"*En" (2)

rises rapidly to 7 and remains there as cooled liquid

converts to ice. When only ice exists, its temperature drops where X*" is the effective mole fraction of solvent water in
below Zs and gradually approaches that of the coolant. the solution, N*, the moles of solvent water, i.e., unbound
4y'1q, water's latent heat of freezing at 7o, is water, N the moles of solute, n* the total mass fraction of
333.6 kJ kg-1 (143.4 Btu lb-1). C*, liquidwater'sheatcapa- water in the solution, z" the mass fraction of solute in the
city, averages 4. I 87 kJ kg-1 K- I or 1.00 Btu lb- 1 o F-1. At solution, the mass of water bound per unit mass of solute,
To, Ct, ice's heat capacify, is 2.093kJkg-tK-t or E : 18.021M,, and M, the solute's molecular weight. Eq. I
0.500Btu 16-r op-r. C1 decreases roughly by with X*,, given by Eq. 2, substituted for a* will be calld
0.007kJkg-1K-l per K or 0.00093BtulbllT-1 per T modified Eq. l.
as ice cools. Food solutions usually contain many solutes. Unless a
solute precipitates, the relative weight proportions of the
dissolved solutes do not change during freezing. There-
fore, constant effective E and b can usually be used for
FREEZING POINT DEPRESSION solute mixtures. Very good fits between modified Eq. 1
and experimental freezing point depressions vs. solute
Dissolved solutes depress water's fteezing point. As Fig. concentration data are obtained for single and mixed
lshows, freezing point depression increases as solute solutes when best-fit values of E and b are used. X*", E,
concentration increases.[2] As water in a solution freezes, and b can similarly be determined for moist solid foods
solute concenfration in the remaining solution increases that contain insoluble solids and water sorbed by those
except when solutes also solidify; and 7, the equilibrium solids in addition to solvent water and solute-bound water.
temperature of the ice-solution mix, decreases as ice By using 2(X*. - l)l(X*. * 1), the first term of a series
formation continues. If the solution contains a single expansion for ln(X*"), in place of ln(X*") in modified Eq. I
solute, cooling-induced ice formation.causes I to drop and noting that AHayf (TsT) - LHolTS, one obtains
until it reaches ZB, the eutectic temperature, where it
remains while the remaining water and the solute solidify. En" En"
After both completely solidifu, continued cooling will n* - bn" * 0.5En, n* - Bn" t
cause I to drop further and gradually approach the
temperature of the coolant. _t&.o2LHo(To - T)
(3)
'r
Equilibrium freezing points of aqueous solutions are RT i
governed by the following equationt2]
: {
where B (b - 0.58). Applying Eq. 3 when n: n*o,tre
ln(a*) : ln(7*X*): - t8.024,H"'(To - T)
(1) total weight fraction of water in the food prior to freezing, c
RTsT and 7: Z, its equilibrium initial freezing point, one
,C)

tM4 Encvctopediaof Asricutturat'Foord'r;:ir:::f

-il"#:iffi;{rZ
Copyright @ 2003 by Marcel Dekker, Inc. All rights reserved. U

2?0 tvfadison Avenr", o.* li,txl*Htriffi m

Thermoilynamics of Food Freezing

obtains

En, _l8.O2AHo(To-T) (4)

-4
nno - Bn" R-
Dividing Eq. 4 by Eq. 3, one obtains
o-6
n* - B:"
n-o-Bn,-!4
(s)
-8 To-T
-10
the equilibrium weight fraction of unfrozen water in the
-12 food at 7. The weight fraction of ice in the food is

-14
,t: Lwo - w -(*o - n*,l!:!)
"'To-T (6)

-16
Dimensionless Eqs. 5 and 6 can be used with temperatures
-18 in K, or oF. The combined weight fraction of solutes
t,
and solids in the food is n. : (1 - n*o).
-20

-22
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 VOLUME CHANGES

Weight Fraction Solutes The respective densities of liquid water and ice are
Ftg. 1 Freezing point vs. solute weight fraction for wine and
1000kgm-3 and 917kgm-3. Therefore as aqueous
juices. (From Ref. 4, p. 137.) solutions freeze, expansion occurs. Freeze-concentrated
solution may exude from spaces between the growing ice
crystals. Gas-filled pores in foods may compact as ice
forms. Therefore, extents of freezing-induced expansion in
porous foods are difficult to predict.

ENTHALPIES
90

80
350 At temperaturesT 1 Ti, enthalpies (heat contents) of foods
undergoing freezing can be obtained by summing up the
70 300 enthalpy contributions of the components involved, i.e.,
(,)
60 ll H : nnH* * n1H1 * nrH, O)
250
50 where 1, H*, Hb and H* are the enthalpy per unit mass for
200 = the food, liquid water, ice, and the combined solids and
40 o- solutes, respectively. The 1/s are measured with respect to
J
150 a reference temperature Ip, usually 233.16K ( - 40'C or
30 -F - 40'D, where each H is zerc. n* at T can be obtained
100 z from modified \. I , n1: flwo - *. If, instead , n* and ni
20 LrJ
are obtained from Eqs. 5 and 6, respectively, the relative
50 error will be at most 0.47o. H*: LHo + C*(I - 7n),
10
h: C{T - In), and I/. : C"(T - Zn). C. is the partial @
'E

0r- J0 heat capacity for the combined solutes and solids in the
-50 -44 -30 -20 -10 0 10 20 food. Using Eqs. 5 and 6, substituting for II", H*, H ns,
T(oC) z*, and ny arld rearranging, one obtains

Fig. 2 Enalpy vs. temperature for lean beef. Sod lines--data H : (T - r*) fc, *(n*o - Bn")LHo!!:-T'1
from Eqs. 6 ad 7. O---+alorimetric data by Reidel.t5]
{
.E
a
C)

2?0 Madison Aven"", n"* #,T,".|ffffl.* m

L046 Thermodynamics of Food Freezing

The heat capacity of the food in the completely frozen crystals instead of dendritic crystals. ,t, given by Eq. 11,
state is C o (n*o - bn")Ct I bn"C* * n"C". If 1 vs. 7 and C. from Eq. 10 or.Fl from Eqs. 8 and t have been used
data obtained by calorimetry are available, C, B, andT,for in modeling how 7 varies with time and position during
use in Eq. 8, can be found by best-fit methods. Above [, fre,ezing and thawing.

H : (Tt- ril [C,

*
(*o- Bn')A]Iol t Co(T -7) (e)
Zo-In DEVIATIONS FROM EOUILBRIUM
Cs is the heat capacity of the food in the thawed state. Fig. 2
Iplot for beef.tsl Eqs. 5, 6, 8, and 9 are
depicts an 11 vs. During freezing, heat transfers from ice growth sites,
more accurate than similar equations derived earliertal and water diffuses through water-depleted regions to
where n" was used instead of Bn". The amount of heat those sites. In cellular foods, membranes impede such
removed per unit mass in cooling a food initially at diffusion. As freezing progresses and solute concen-
temperature Tt ) Ti to subfreezing temperature 12 is tration in the residual solution increases, food solutions
Ht - Hz, where I11 is obtained by substituting T1 for T in become very viscous. Because of high viscosity, water
Eq. 9 and I12 by substituting T2 for I in Eq. 8. diffuses more slowly to ice gowth sites. Solute
concentrations at those sites rise above those in the
Effective Heat Capaci rest of the residual fluid; and the temperature falls
below the equilibrium temperature for the current
As freezing of foods takes place over a range of average bulk solute concentration. Deviations from
temperatures, the effective heat capacity C" at T for
equilibrium are greatest when freezing is carried out
T < Ti is obtained by differentiang F,q. 6 with respect to
rapidly and at I below -2O"C. Therefore, extremely
T. C.: dHldT, yielding slow cooling has to be used in calorimetric measnre-
ments intended to determine equilibrium properties of
. LHs(n*o -
Le:Lf+6 Bn,)(To - Ti)
(10)
freezing foods. Deviations from equilibrium often occur
in industrial freezing processes. Thus, the equations
presented earlier will be somewhat in error for such
Eq. l0 is valid only at 7 below Ti. As T increases, C, processes. As most heat removal and ice formation
increases and peaks sharply at Ti, a.9., for lean beef it occurs just below T and well above - 20oC, the errbr
goes from 2.2AkJkg- rK-l at 233.16K (-40'C) to involved will be tolerable for most practical purposes.
220.6Ukg- 1K- 1 at Ti,i.e.,272.16K(-1.0b'C). As soon
as Iis above ?], Eq. 10 no longer applies and C, immediately
1 1
drops to Cs, a.g.,3.456 kJ kg- K- for lean beef.
EFFECTS OF CRYSTAL SIZE

Small crystals possess more surface energy per unit mass

THERMAL CONDUCTIVITY
than larger ones. Therefore, equilibrium Z for small ice
crystals are slightly lower than that for larger crystals. The
The respective thermal conductivities of ice and in I is inversely proportional to ice crystal
water at 273.16K are kt:2.25Wm- 1K- 1 and
depression
diameter. For lpm crystals the depression is roughly
k* : 0.569Wm- I K- l. Therefore, the thermal conduc-
0.1"K, for 10 pm crystals roughly 0.0 1"K, and for 100 pm
tivity /c of a foodincreases markedly as it freezes. Thermal
crystals roughly 0.001'K. Smaller crystals tend to melt and
conductivity is a transport property, not a thermodynamic
larger crystals tend to grow in foods containing ice crystals
property. Nevertheless, Eq. 9 correlates /< for foods fairly
of mixed size. This effect, Ostwald ripening, can lead to
well at T < Ti when, as frequently occurs, dendritic (tree-
graininess in ice cream. Ostwald ripening is used to
like) ice crystals form during freezing.
produce large, readily separable ice crystals in modern
freeze-concentraon systems. t2I
k: kr *(ro - frr) +#1 (1r)
.E

ftq and t are the thermal conductivities of food in the fully REFERENCES ,j
thawed and fully frozen states, respectively. At T > Ti,
Eq. I 1 no longer applies and : h.Bq. 1 1 does not fully
H

1. Hobbs, P.J. Ice Physbs; Oxford University Press: Oxford,

account for increases in /c that occur as 7 decreases, but is 1974;346-364. e
accurate enough for most engineering purposes. Eq. 11 is 2. Schwartzberg, H.G. Food Freeze Concentration. In Biotech-
not valid for materials that contain dispersed, rounded ice nology and Freeze Concentration; IFT Basic Symposium o
{
'

.")

N{a:nr Dnxxen, INc.

270 Mad.ison Avenue, New York, New York ltfll
 Thermddynmics of Food tr.rering lM7

Series, Schwartzberg, H.G., Rao, M.A., Eds.; Marcel 4. Schwartzberg,H.G.EffectiveHeatCapacitiesfortheFreezing

Dekker, Inc.: New York, 1990; 127 -2O2. and Thawing of Foods. J. Food Scr.lyl6,4l,l52-156.
3. Riedel, L. On the hoblem of Bound Water in Meat. 5. Riedel, L. Calorimetric Investigation of the Meat Freezing
Kaltetechnik 1961,13,41-43. Process. Kaltetechnik 1957,9,38-40.

1\"
{J

{
e
{

o

o

o

MrncuDexren,lxc.
2T0 Madison Avenue, New York Ncw York 10016 m