JFS: Food Engineering and Physical Properties

Rheological Characterization of
Skim Milk Stabilized with Carrageenan
at High Temperatures
A.D. ANDERSON, C.R. DAUBERT, B.E. FARKAS

ABSTRACT:Superpositioning principles were applied twice to model temperature (25 to 125 8 C) and concentration
(0.005 to 0.040% w/w) effects on skim milk and carrageenan solutions. Samples were analyzed using a controlled
stress rheometer equipped with a pressurized sealed cell, permitting measurements well above standard boiling
conditions. Individual samples were sheared between 10 and 160 s -1, and predictive equations were developed to
predict Newtonian viscosity as a function of temperature and carrageenan concentration. The superpositioning
technique coupled with advancements in rheological instrumentation permits high temperature measurements and
offers a strategy for viscosity determination for thermal processing unit operations.
Keywords: concentration, temperature, superposition, rheology, carrageenan

Introduction and others (1994) collapsed rheological data for pectin disper-

Food Engineering and Physical Properties

A DVANCEMENTS IN RHEOLOGICAL INSTRUMENTATION, PERMITTING
product assessment at temperatures above sample boiling,
present new opportunities for the rheologist to simulate thermal
sions obtained between 5 and 65 8C to a reference temperature
of 25 8C. To evaluate the independent effects of concentration
and temperature, Partal and others (1997) applied time-tem-
processing conditions using benchtop methodologies. Predic- perature and time-concentration superpositioning principles
tive equations developed from modeling rheological data collect- independently to model the flow behavior of stabilized emul-
ed over broader temperature ranges may be applied during pro- sion systems.
cess development for calculating product residence times during The objective of the research is to present a procedure for
heating. Better viscosity prediction insures adequate thermal modeling temperature and concentration effects of skim milk
treatment and limits opportunities for product overheating. stabilized with a carrageenan blend at processing temperatures
Viscosity changes caused by temperature are typically mod- up to 125 8C. According to ingredient specifications, the blend se-
eled for Newtonian fluids using an Arrhenius relationship (Stef- lected is typically used to stabilize chocolate milks, and concen-
fe 1996). In addition to modeling temperature effects, model- trations were selected as representative of typical industrial use
ing concentration effects is often desirable, especially for levels to accommodate different functional purposes. Carrag-
predicting viscosity for solutions with varying hydrocolloid con- eenan fulfills many roles when stabilizing food systems, and it’s
tent. Davey (1997) and Velez-Ruiz and Barbosa-Canovas (1997) functionality depends on the ratio of carrageenan type: kappa,
developed predictive equations modeling the effects of tem- lambda, or iota. Interactions between carrageenan types with
perature and concentration on viscosity by adding terms to the dairy ingredients was not a focus of this research. Nevertheless,
Arrhenius relationship accounting for concentration effects. the blend selected for this study was comprised of greater than
However, temperatures above standard temperature-pressure 50% k-carrageenan. Advancements in rheological instrumenta-
(stp) boiling consistent with higher temperature processing tion accommodate data collection for low viscosity fluids at high
have not been modeled, most likely due to instrumental con- temperature conditions. Superpositioning principles, normally
straints and limitations. applied once to account for either temperature or concentration
Time-temperature superpositioning (Ferry 1970), a com- effects, are herein performed in combination. Empirical equa-
monly-used empirical technique incorporating shift factors to tions derived from variables of the superpositioning process pre-
account for property changes, is a method often used for mod- dict viscosity as a function of temperature and carrageenan con-
eling temperature effects on polymers (Macosko 1994) and re- centration and may be applied to predict viscosity for
ducing the thermal and shear effects to a single master curve. high-temperature, processing considerations.
According to the principle behind superpositioning, the effects
of time and temperature on a system have equal effects in such Materials and Methods
a way that a simple shift of the response function superimposes
data taken from different temperatures (McNulty and Flynn Materials
1979). This method may be used for modeling viscosity of food Pasteurized skim milk, processed at 79 8C for 18 sec, was ob-
products not necessarily adhering to an Arrhenius relationship tained from the North Carolina State University (Raleigh, N.C.,
with increasing temperature or by a simple Newtonian or power U.S.A.) dairy facility. A commercially available carrageenan
law model. For example, locust bean gum solutions were sub- blend (SeaKem CM611), specifically developed for use by the
jected to superpositioning to document the temperature effects dairy industry, was obtained from the FMC Corporation (Phila-
on viscoelasticity (Richardson and Norton 1998). Lopes da Silva delphia, Pa., U.S.A.).

© 2002 Institute of Food Technologists Vol. 67, Nr. 2, 2002—JOURNAL OF FOOD SCIENCE 649
 17503841, 2002, 2, Downloaded from https://ift.onlinelibrary.wiley.com/doi/10.1111/j.1365-2621.2002.tb10654.x by C/O Glucksman Library, Wiley Online Library on [21/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Temperature/concentration superposition . . .

Sample Preparation Table 1—Superpositioning Shift Factors.

To add carrageenan, skim milk was heated and stirred at 350 Part I Part II
rpm to create a vortex using a Thermolyne stir and heat plate Temperature Superposition Concentration Superposition
(Barnstead-Thermolyne, Dubuque, Iowa, U.S.A.). Carrageenan (5 temperatures) (4 concentrations)
was added slowly to the vortex center to avoid clumping at concen- aT, 25 8C = 0.34 aC, 0.005 w/w = 1.12
trations of 0.005, 0.010, 0.020, and 0.040% (w/w). Solutions were aT, 65 8C = 0.90 aC, 0.001 w/w = 1.07
heated to 85 8C for approximately 5 min, allowing the carrageenan aT, 85 8C = 1.00 aC, 0.020 w/w = 1.00
aT, 105 8C = 1.03 aC, 0.040 w/w = 0.89
to hydrate and solubilize. Next, each sample was cooled to room aT, 1258C = 1.19
temperature with no agitation, followed by rheological analyses.

Equipment and Rheological Testing

A StressTech, controlled stress rheometer (Rheologica Instru-
ments AB, Lund, Sweden) was equipped with a pressurized lated stress (s) for each temperature was divided by the calculat-
sealed cell to complete rheological measurements, see Figure 1. ed stress (s*) at the reference temperature (85 8C). The subse-
The rheometer had a minimum stress requirement of 25 mPa, quent value was the shift factor (aT) for each temperature
enabling accurate measurements of low viscosities at high tem-
peratures. Compressed air was applied to the sealed cell at
179.270 kPa to prevent sample boiling and excessive moisture (2)
loss at all temperatures.
Individual samples were sheared between 10 and 160 s -1 at Shift factors for a given temperature were averaged considering
25, 65, 85, 105, and 125 8C using a 25.0 mm rotating cylindrical at- each concentration to produce a T, provided in Table 1. When
tachment in a 27.5 mm cup. Samples tested above 25 8C were shear stress was plotted against shear rate divided by the shift
pre-sheared at 50 s-1 while ramping temperature (4 8C/min) to factor, all the data lined up on the reference curve - the curve at
insure adequate heat distribution and limit cooking. All rheologi- 85 8C. The equation for viscosity became:
cal data collected below the rheometer sensitivity (25 mPa) was
Food Engineering and Physical Properties

discarded, and the Taylor number was computed to guard

against secondary flows compromising the results. All tests were (3)
performed in duplicate the day of sample preparation.
Since the carrageenan blend was tested at 4 concentrations, 4
Model Development
reference curves shifted for temperature were produced (Figure
Superpositioning principles modeled temperature and con- 3). The same superposition procedure was applied again to fur-
centration effects on apparent viscosity by incorporating calcu-
ther shift data for concentration effects. The chosen reference
lated shift factors for each condition. To begin model develop- concentration was 0.02%, and the variable a C represented the
ment, shear stress was plotted as a function of rate for each shift factors for concentration,
temperature-concentration combination tested (Figure 2). A
Newtonian relationship was observed at every combination
above 25 8C: (4)

(1)

An 85 8C reference temperature was selected, corresponding to a

central temperature tested within the range. The reference vis-
cosity, m*, was the Newtonian viscosity at 85 8C for each concen-
tration. For every temperature, shear rate was calculated at 50
mPa using the corresponding Newtonian viscosity, m. The calcu-

Figure 2—Effects of temperature on the rheology of 0.01%

carrageenan in skim milk. Legend: e 25 °C, h 65 °C, n
Figure 1—Rheometer pressure cell. 85 °C, s 105 °C, 3 125 °C

650 JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 2, 2002

 17503841, 2002, 2, Downloaded from https://ift.onlinelibrary.wiley.com/doi/10.1111/j.1365-2621.2002.tb10654.x by C/O Glucksman Library, Wiley Online Library on [21/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Temperature/concentration superposition . . .

also displayed in Table 1. The reference viscosity, m*, was modi- the superpositioning process shifted data based on temperature
fied to be the viscosity at 85 8C at a carrageenan concentration of effects. Figure 2 depicts typical temperature trends for a skim
0.02%: milk and carrageenan solution at a concentration of 0.01% (w/w).
For each concentration tested, viscosity typically diminished
with increasing temperature as expected. However, results devi-
(5) ated from this trend at certain temperatures, and these discrep-
ancies were attributed to protein denaturation and/or aggrega-
Since aT and aC vary with temperature (T) and concentration tion or sample cooking. Deviant samples were re-evaluated to
(C), respectively (see Table 1), the shift factors were modeled us- insure variations were inherent to the sample and not caused by
ing Power Law fits as opposed to Arrhenius fits (Partal and others instrumentation errors or inconsistent testing methods. Overall,
1997; Ferry 1970) sample viscosities showed minimal variation at the observed, el-
evated temperatures.
Once each concentration of carrageenan was shifted for tem-
(6)
perature, the same process was repeated to shift for concentra-
tion effects using the appropriate factor. Figure 3 represents ref-
(7)
erence curves already shifted for temperature for each
concentration of CM carrageenan. Reference viscosity, denoted
where b,d and f,g represent empirical power law constants. Sub-
by the slope of each data set, increased with increasing concen-
stituting equations (6) and (7) into equation (5) gives an empiri-
tration, as expected.
cal model for viscosity:
The final superposition plot (Figure 4) represents data collect-
ed at temperatures from 25 to 125 8C at carrageenan concentra-
(8) tions of 0.005% to 0.04%. All data was shifted to a final master
curve to account for temperature and concentration effects.
A model equation, based on equation 10, was developed from
Collapsing constant values for variables m*, b, and f yields an variables of the superpositioning process to predict viscosity of
overall constant, A:

Food Engineering and Physical Properties

Newtonian skim milk and carrageenan solutions for the range of
temperatures (8C) and concentrations (% w/w) evaluated.
(9)

(11)
The final equation for Newtonian viscosity becomes:

The ability of the skim milk and carrageenan equations to ac-

(10) curately predict viscosity is shown in Figure 5. Each temperature-
concentration combination produced a viscosity, modeled using
the prediction equation for each temperature from 65 to 125 8C
Results and Discussion
(a total of 16 combinations). A correlation between actual mea-
F IGURES 2 THROUGH 4 SHOW THE PROCESS OF SUPERPOSITIONING
data for the milk and carrageenan solutions. The first step in
sured viscosities and predicted values resulted in a slope of 1.00
and a 0.74 coefficient of determination. Slight scattering of the

Figure 3—Effects of concentration on the rheology of car-

rageenan in skim milk. Each concentration represented Figure 4—Reference curve showing combined tempera-
has already been shifted to account for temperature ef- ture and concentration effects for carrageenan in skim
fects. Legend: e 0.005%, h 0.01%, n 0.02%, s 0.04% milk. Legend: e 0.005%, h 0.01%, n 0.02%, s 0.04%

Vol. 67, Nr. 2, 2002—JOURNAL OF FOOD SCIENCE 651

 17503841, 2002, 2, Downloaded from https://ift.onlinelibrary.wiley.com/doi/10.1111/j.1365-2621.2002.tb10654.x by C/O Glucksman Library, Wiley Online Library on [21/11/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Properties of Herring Protein Hydrolysate . . .

dition above 65 8C. For the food engineer designing a thermal

process, this value is important for design considerations.

Conclusions

S UPERPOSITIONING PRINCIPLES DEVELOPED AN EMPIRICAL EQUATION

to predict viscosity of skim milk and carrageenan solutions as
a function of temperature (8C) and carrageenan concentration
(% w/w). The equation is applicable only within the ranges of ex-
perimental conditions tested. Viscosities at temperatures of 65 8C
and above were below 1.2 mPa·s for all concentrations consid-
ered, and overall variability between measured viscosity and
0.74 predicted viscosity was less than 0.1 mPa·s. The superposition-
ing technique presented, coupled with advancements in rheo-
logical instrumentation to permit high temperature measure-
ments, offers a method for viscosity prediction at thermal
processing conditions.

References
Figure 5—Correlation between observed viscosity and pre- Davey KR. 1997. A new rheological model for predicting the combined effect of
dicted viscosity for skim milk and carrageenan solutions. temperature and concentration on viscosity. In: Jowitt R, editor. Engineering
& Food at ICEF 7. Part 1. Sheffield, UK: Sheffield Press. p E60-E63.
Ferry JD. 1970. Viscoelastic properties of polymers. 2nd edition. New York: John
Wiley & Sons, Inc. 292 p.
Langendorff V, Cuvelier G, Michon C, Launay B, Parker A, De kruif CG. 2000.
data reduced the correlation, attributed to the narrower range of Effects of carrageenan type on the behavior of carrageenan/milk mixtures.
Food Hydrocolloids 14:273-280.
viscosities produced at the conditions. Overall variability be- Lopes da Silva JA, Goncalves MP, Rao MA. 1994. Influence of temperature on the
tween measured viscosity and predicted viscosity was less than dynamic and steady-shear rheology of pectin dispersions. Carbohydrate Poly-
0.1 mPa·s. mers 23:77-87.
Food Engineering and Physical Properties

Macosko CW. 1994. Rheology Principles, Measurements, and Applicatons. New

In order for superpositioning to be effective, molecular relax- York: VCH Publishers, Inc. 550 p.
ation mechanisms are assumed to be similarly influenced by McNulty PB, Flynn DG. 1979. Temperature dependence of the stress relaxation
behavior of aqueous sugar glasses. Trans. ASAE:445-448.
equivalent temperature changes (Lopes da Silva and others 1994). Partal P, Guerrero A, Berjano M, Gallegos C. 1997. Influence of concentration
This consideration is valid for solutions above a critical tempera- and temperature on the flow behavior of oil-in-water emulsions stabilized by
sucrose palmitate. JAOCS 74(10):1203-1212.
ture where the carrageenan undergoes transition from a helix (or- Richardson PH, Norton IT. 1998. Gelation behavior of concentrated locust bean
dered) to a coiled (disordered) formation (Schorsch and others gum solutions. Macromolecules 31:1575-1583.
Schorsch C, Jones MG, Norton IT. 2000. Phase behaviour of pure micellar casein/
2000; Langendorff and others 2000). At 25 8C, carrageenan solu- k-carrageenan systems in milk salt ultrafiltrate. Food Hydrocolloids 14:347-
tions have gelled and display helix formation, frequently leading 358.
Steffe JF. 1996. Rheological Methods in Food Process Engineering. 2 nd edition.
to non-Newtonian rheological flows. This deviation from ideal East Lansing: Freeman Press. 418 p.
Newtonian behavior may complicate superpositioning. This situa- Velez-Ruiz JF, Barbosa-Canovas GV. 1997. Rheological characterization of con-
tion was not experienced for the described experimental design, centrated milk at various temperatures and concentrations. In: Jowitt R, edi-
tor. Engineering & Food at ICEF 7. Part 1. Sheffield, UK: Sheffield Press. p E72-
but should be considered at elevated concentrations, alternative E75.
carrageenan selection, or lower temperatures in general. MS 20000747 Submitted 7/24/00, Accepted 2/4/02, Received 2/5/02
The pressure cell attachment allowed the measurement of The authors are grateful for the financial support provided by the Southeastern Dairy
Foods Research Center. Appreciation is also expressed to the North Carolina State Univer-
rheological properties at all experimental temperatures. At high sity Dairy, Raleigh, N.C., for providing milk samples and to FMC Corporation, Philadelphia,
temperatures, low viscosities did challenge rheometer sensitivi- Pa., for providing carrageenan samples. Thank you also to Dr. Van-Den Truong for his lab
assistance.
ty, and data collection was in fact limited to a narrower shear rate
range. By considering sensitivity constraints and Taylor flow Authors are with the North Carolina State University, Department of Food
Science, Box 7624, Raleigh, N.C. 27695. Please address correspondence to
complications, the data described is accurate. A close inspection author Daubert, e-mail:[email protected]
of Figure 5 reveals the viscosity was below 1.2 mPa·s for each con-

652 JOURNAL OF FOOD SCIENCE—Vol. 67, Nr. 2, 2002