You are on page 1of 12

HIGH TEMPERATURE RHEOLOGY OF TOMATO PUREE AND STARCH DISPERSION WITH A DIRECT-DRIVE

VISCOMETER
M.A. RAO', H.J. COOLEY and H-J. LIAO
Department of Food Science and Technology Cornell Universiry Geneva, Ny 14456
Accepted for Publication October 12. 1998

ABSTRACT
A direct-drive concentric cylinder viscometer in a pressure chamber was f used to study the flow behavior o a I1 Brix tomato puree at several temperatures in the range: 7 to 120C. The activation energy of flow of the tomato 6 puree, based on apparent viscosity at 50 s-I, was: 9.4 W mol-I. The apparent viscosity (qJ versus temperature profiles of a 4.0% wary rice (WR) starch dispersion were determined during continuous heating from 30 to llOC at six shear rates in the shear range: 114 to 644 s-'; beginning at about 95C, significant increase in q,, was observed at 114 and 160 s-I. The said profiles collapsed to a single curve when 7, values were converted to reduced viscosity,
q = {) & q;.

Values o q, of the starch dispersion were lower than those o q* f f

at about the same shear rates, ?;=5.7 s" and w=6.3 rad s-'.
INTRODUCTION
Many fluid foods are subjected to temperatures greater than lOOC during thermal processing. Therefore, flow properties of fluid foods at these temperatures are useful in thermal processing applications. Relatively few studies have been conducted on rheological behavior of foods at high temperatures ( 95C). Dail and Steffe (1990) used a pressurized tube (12.7 mm dia) viscometer to study the flow behavior of 1.82% and 2.72% waxy maize (WM) starch dispersions at 121 to 143C. This system required a large mass ( 160 kg) of test sample and, as with many tube/capillary viscometers, time-dependent rheological behavior at a fixed shear rate could not be detected. Abdelrahim et al. (1995) studied the flow behavior of WM starch dispersions with an indirect-drive

' Corresponding author


Journal of Food Process Engineering 22 (1999) 29-40. All Rights Reserved. Ocbpyright 1999 by Food di Nutrition Press, Inc.. Trumbull. Connecticut.
29

30

M.A. RAO,H.J. COOLEY and H-J. LIAO

(magnetic coupling) pressurized (D 1OO/3OO, Haake) concentric cylinder viscometer system. The uncertainty in rotational speeds due to slip of the magnetic coupling and the effect of the clutch mechanism were of concern. Nevertheless, the two studies provided valuable information on flow behavior of starch dispersions at high temperatures. It is well known that similar rheological data can be obtained on a specific food product using different flow geometries (e.g., capillary, concentric cylinder, cone-plate), when the data are obtained properly in terms of the basic units of shear rate (s-') and shear stress (Pa). A tomato puree does not undergo significant changes in composition when subjected to temperatures in the range 76 to 120C for short times so that shear rate versus shear stress data can be obtained at specific temperatures. Over the same temperature range, a starch dispersion undergoes irreversible changes due to starch gelatinization and it would be desirable to obtain rheological data as the sample is heated continuously ( D o h and Steffe 1990; Yang and Rao 1998). Our objective was to evaluate the performance of a pressurized direct-drive concentric cylinder viscometer system in obtaining shear rate versus shear stress data on a tomato puree at several fixed temperatures between 76 and 120C, and temperature versus apparent viscosity data at several shear rates on a 4% waxy rice (WR) starch dispersion during gelatinization over the temperature range: 30 to 110C.

MATERIALS AND METHODS


Materials Canned tomato puree (11 Brix) purchased at a local supermarket w s used. a WR starch donated by California Natural Products (Lathrop, CA) was used; its amylose content was found to be less than 1 % using the method for rice starch based on the strong iodine affinity of amylose (Juliano er al. 1981).To minimize settling of starch granules in the concentric cylinder geometry, a two-step procedure of Keetels and van Vliet (1993)was used: (1) using a small portion (1.5 %) of the required amount of starch, a gelatinized dispersion was created by heating to 75C in boiling water to which the remaining amount of ungelatinized starch (2.5%)was added, and (2)the entire dispersion, made up of a large amount of ungelatinized granules suspended in a medium containing a small amount of gelatinized granules, was allowed to hydrate for 1 h and then poured in to the MVI concentric cylinder geometry.

Pressurized Viscometer System


Figure 1 is a schematic diagram of the experimental set up of the directdrive concentric cylinder viscometer system in a pressure chamber. The drive

Pressurized Direct Drive Viscometer System


, Motor Drive Cable

Control Panel

Temperature Control System

E n
X

25.0"C

I I

0 0

Cover Doors Held Closed with C-Clamps

FIG. 1. SCHEMATIC DIAGRAM OF PRESSURIZED DIRECT DRIVE VISCOMETER SYSTEM

32

M.A. RAO, H.J. COOLEY and H-J. LlAO

motor, torque unit, and concentric cylinder unit and temperature control vessel of a Haake RV2 viscometer system (Haake Inc.) were placed in a chamber (PRC) that could be pressurized to 0.2 MPa (two atmospheres). The PRC had housed a drum drier previously. The temperature control vessel was insulated to minimize heat loss. A copper-constantan (36 gage wires) thermocouple placed in the well of the inner concentric cylinder measured the temperature of the test sample. The control unit of the viscometer was kept outside the PRC so that the desired rotational speeds could be selected. The rotational speed and the corresponding torque data, as well as the temperatures sensed by the thermocouples were recorded on a lap-top computer (Toshiba, T1200 XE) after analog to digital conversion via a data logger (Digistrip 11, Kaye Instruments). The electrical lines to the drive motor and the torque unit inside the PRC were passed through air-tight Swagelok fittings. A high-temperature bath (Model 9500, Fisher Scientific) and a lowtemperature,bath (Model MH, Julabo, Germauy), both containing silicone oil (Aldrich Chemical Co.) as the heating/cooling medium, were used to heat and cool the test sample, respectively. Insulated copper tubing was used for flow of the silicone oil between the oil baths and the viscometer temperature control vessel; the rigid tubing also helped in holding the viscometer firmly in the PRC. A set of three valves was used to either circulate the hot oil outside the PRC and keep it well agitated in the high-temperature bath or through the viscometer temperature control vessel. Another set of three valves was used with the lowtemperature oil bath. Experimental Procedure In a typical experiment, the temperatures of the high- and low-temperature baths were first set and allowed to reach equilibrium values while the silicone oil was circulated in the lines outside the PRC. Because there was considerable energy loss in the high-temperature lines, there was a substantial difference between the set temperature of the high-temperature bath and the sample temperature. Therefore, the temperature of the high-temperature bath was set about 20C higher than the desired sample temperature. The concentric cylinder system (MVI, Ri=20 mm, %=21 mm, H=60 mm) was carefully filled ( - 120 mL) with either tomato puree or starch dispersion, and secured in the temperature control vessel. Apparent Viscosity (qb Data. The data logging system was started, and after the PRC was pressurized using filtered laboratory compressed air, the test sample was heated as desired. Torque data were obtained on the tomato puree samples at selected values of RPM at the fixed temperatures: 76, 95, 100, 110

HIGH TEMPERATURE RHEOLOGY

33

and 120C. A fresh tomato puree sample was used at each temperature. The torque and the RPM values were converted to shear stress (a, Pa) and nonNewtonian shear rates (q, s-l), respectively, as described in Vitali and Rao (1984). Typical experiment time for collection of RPM versus torque data at each temperature was about 20 min. In preliminary experiments of about 40 min duration, a thin layer of slightly concentrated tomato puree due to dehydration was found at the top of the concentric cylinder system, and from the weights of the MVI geometry before and after experimentation loss of sample due to evaporation was found to be less than 2%. However, to keep evaporation losses at a minimum, total experiment time at the high temperatures (95 to 120C) was kept close to 20 min. A Teflon" cover at the top of the concentric cylinder system and a piece of wet cloth wrapped above the concentric cylinder system helped considerably in reducing moisture loss from both the tomato puree and the starch dispersion samples. With the 4.0% WR starch dispersion, in order to obtain yp versus temperature profiles as the starch dispersion was undergoing gelatinization, torque data were recorded at fixed RPM values corresponding to the Newtonian shear rates: 114, 160,228 and 322 s-' as the dispersion was heated continuously from 30C to about 1IOC. No attempt was made to correct the shear rates for the non-Newtonian behavior of the starch dispersions because the correction would be small for the narrow gap of the MVI concentric cylinder system and the difficulty in estimating the continuously changing flow behavior index, that was needed for the correction (Vitali and Rao 1984), during the transient rheological behavior studied.

Complex Viscosity ( p ) Versus Temperature Profile of 4% WR Dispersion. In order to compare qa versus temperature profile with that of complex viscosity (q*) versus temperature profile, data were obtained on 4% WR dispersions with the pressurized viscometer system and a Carri-Med CSL' rheometer (TA Instruments)with a 4 cm stainless steel parallel plate system with 500 pm gap, respectively. The qa versus temperature data were obtained from about 25C to llOC at a shear rate 5.7 s-I, and the q* versus temperature data from about 55C to 95C at a dynamic frequency 6.3 rad s-', 3% strain, and a heating rate 2C min-I. The Carri-Med rheometer was operated at n o d atmospheric pressure. More details on the procedures used to obtain reliable q* versus temperature data with the Carri-Med rheometer are given in Yang and Rao (1998).

34

M:A. RAO, H.J. COOLEY and H-J. LIAO

RESULTS AND DISCUSSION

Tomato Puree
The logarithms of the shear rates plotted against the shear stresses of the tomato puree at 76, 95, 100, 110, and 120 C are shown in Fig. 2. The slight scatter in the data is about as much as observed in earlier studies at lower temperatures (Tanglertpaibul and Rao 1987) and is probably due to rearrangement of the tomato pulp particles in the 11 Brix tomato puree sample within the MVI geometry. The data at all the temperatures followed the power law model: u =K+" (1) where, n is the flow behavior index (dimensionless) and K (Pa s") is the consistency index. Magnitudes of the flow behavior index and the consistency index of the tomato puree as a function of temperature are given in Table 1. Values of K decreased noticeably with increase in temperature from 7.0 Pa s" at 76 C to 5.5 Pa s" at 120C. In contrast, the values of n showed a modest change from 0.27 at 76C to 0.23 at 120C.

1.4
m
(] I
44 -

1.4 1.3 1.2 1.1 1 1 1.5


2

2 1.3
m

1.2

1.1

1 2.5
3

log (shear rate, s-')


FIG. 2. LOG SHEAR RATE VERSUS LOG SHEAR STRESS DATA OF 11 BRIX TOMATO PUREE The data at all the temperatures followed the simple power law model.

The effect of temperature on the apparent viscosity at 50 s-' of the tomato puree w s well described (Fig. 3) by the Anhenius relationship (Eq.2): a fl,.so = fl,exp@PV (2)

HIGH TEMPERATURE RHEOLOGY

35

where, 17.. is the frequency factor, T is the temperature (K), R is the gas constant (J mol-I K-I), and E, is the activation energy (J mol-I). The shear rate 50 s-I was within the range of shear rates used and was chosen arbitrarily. From Fig. 3, the activation energy was determined to be 9.4 k mol-' and 7.. to be l 0.015 Pa s.
TABLE 1.' POWER LAW PARAMETERS OF 1 1 "BFUX TOMATO PUREE

Temperature

Shear rate range


(S-9

("0
76 95
100

K (Pa sn) 7.0


6.1

n (dimensionless)
0.27 0.26 0.25 0.26 0.23

R2

13-121 19-216 31435 27-608 39-622

0.99

0.98

5.9 5.1 5.5

0.99
0.99 0.99

110

I20

aK is consistency index, n is flow behavior index.

- 1 .os
h

-1.05
-1.1

v .

a
v1

L a

-1.1 -1.15

-1. 1s -1.2

@
0 .-

>

-1.2 -1.25 -1.3 2.5

d a
v

- 1.25
-1.3 103

(lm, K)
FIG. 3. ARRHENIUS PLOT OF (l/T,K) VERSUS APPARENT VISCOSITY AT 50s.' OF 1 1 BRIX TOMATO PUREE

36

M.A. RAO. H.J. COOLEY and H-J. LIAO

Starch Dispersion Apparent Viscosity Versus Temperature. The complex viscosity versus temperature profiles of corn starch dispersions were not affected significantly by the heating rates: 1.6-6.OC min-' (Yang and Rao 1998). Thus, it is useful to know magnitudes of the heating rates. From the recorded sample temperature (in the well of the inner cylinder) as a function of time, the heating rates of the starch dispersions were determined to be in the range 3.3-3.8C min-'. While the heating rates are comparable in magnitude to those used by Yang and Rao (1998) using a Carri-Med rheometer, they could not be controlled precisely as with the Peltier heating system of the Carri-Med rheometer. Apparent viscosity versus temperature data on the 4% WR starch dispersion at the shear rates 114, 161, 228, and 322 s'are shown in Fig. 4. For the sake of clarity, data at 445 and 644 s-I are not shown in Fig. 4. A common feature of the data was that qa increased rapidly with temperature from 60C to 70C, reached a plateau region in the range -70 to 105C, and then decreased at temperatures > 105C. At the shear rates 114 and 161 s-], there was a clear increase in qa towards the end of the plateau region in the range 95 to 105C. Because the shear rate was held constant, the observed increase in viscosity beginning at about 95C appears to be due to shear thickening. Shear thickening has been observed with waxy maize dispersions (Dail and Steffe 1990). Therefore, it is not surprising that shear thickening in the range 95 to 105C was observed with the 4% WR dispersions. Further, the observed shear thickening clearly tookplace towards the later stages of gelatinization. Okechukwu and Rao (1995, 1996) reported that partially gelatinized corn and cowpea starch dispersions hated at fixed low temperatures -65 to 80C exhibited shear thickening behavior. However, because of the increase in viscosity due to increase in granule size in the temperature range -60C to 70C, it is not possible to decipher in Fig. 4 if shear thickening occurred in the beginning of gelatinization. The qa, at the shear rates: 114, 161, 228, 322, 445, and 644 s' versus temperature data were reduced to a single curve by scaling the values of 9, using a reference shear rate, y = 114 s-I, as described in detail by Yang and Rao (1998). In Fig. 5, the reduced apparent viscosity (q&) plotted against temperature is shown, where q&is defined as:

The value 0.85 of the scaling exponent /3 was within the range of values reported earlier (Yang and Rao 1998). The reduced viscosity data in Fig. 5

HIGH TEMPERATURE RHEOLOGY

31

converged well to a single curve over most of the temperature range and the deviations were in data that exhibited shear thickening.
0.2
i n

2
.Y

0.1s

i ,
i n

.i >

0.1

20

40

60

80

100

120

Temperature, "C
FIG.4. APPARENT VISCOSITY (Pa s) VERSUS TEMPERATURE (C) OF 4 % WAXY RICE
STARCH DISPERSION AT SEVERAL SHEAR RATES . Note the shear thickening behavior at 113.8 and 160.9 s' at the end of the plateau regions.
h

v1

k
i
w

0.2

h
v1

. I

. I

8 > * E Ll
v1

0.15 -

0.1 -

yf
a* r
q $
A

Reference shr rate: 113.8 5'


643.9 s.'

'
(

445.4s.'
321.9s"

0.0s

z
V

L E
60
70

80

90

100

110

120

Temperature ("C)
FIG. 5 . REDUCED APPARENT VISCOSITY tla=q

{ir,

VERSUS TEMPERATURE

PLOT OF 4% WAXY RICE STARCH DISPERSION The value of the scaling exponentd was 0.85 and the reference shear rate 9, was 114 S-'.

38

M.A. RAO. H.J. COOLEY and H-J. LIAO

Complex Viscosity Versus Temperature. The q, versus temperature and the q* versus temperature profiles of the 4% WR dispersion are shown in Fig. 6 at a = 5.7 sLLand a o = 6.3 rad s-I, respectively. The dip seen in the qa profile was probably due to either settling of starch granules or breakdown in structure followed immediately by its reformation. Compared to v*, the q, values began to increase at a lower temperature, 50C versus 62C for values of q*, probably due to structure formation by the amylose that leached out of the shear-induced partially disrupted WR starch granules. The much flatter q, versus temperature profile and lower q, values in the range -70 to 95C were also due to shear-induced rupture of granules. The q* versus temperature data in Fig. 6 were obtained with minimal disturbance to the structure of the WR starch dispersion, while the qa data were obtained under continuous shear, both at about equal values of shear rate: 6.3 rad s-l and 5.7 s-I, respectively. Therefore, the difference in the areas of the two curves (AA) can be used to estimate quantitatively the effect of continuous shear of a fixed magnitude on the structure of the WR starch dispersion. In the temperature range 55 to 95C in Fig. 6, the area A1 under the complex viscosity

3.5
3
2.5
2
1.5

1 0.5
0 0

FIG. 6 . COMPLEX AND APPARENT VISCOSITY VERSUS TEMPERATURE (C) OF 4% WAXY RICE STARCH DISPERSION AT A DYNAMIC FREQUENCY 6.28 RAD s. AND A SHEAR RATE OF 5.69 s., RESPECTIVELY The differences in the profiles of the two curves are due to the nature of the shear fields used: small amplitude oscillatory versus continuous shear.

HIGH TEMPERATURE RHEOLOGY

39

versus temperature curve and A2 under the apparent viscosity versus temperature curve were 79.3 and 53.9 Pa s C, respectively. Compared to complex viscosity data, the reduction due to continuous shear: AA/A1 amounted to 32%.

CONCLUSIONS
A pressurized direct-drive concentric cylinder viscometer system provided reliable shear rate versus shear stress data on a 11 Brix tomato puree at specific temperatures between 76 and 120C, and apparent viscosity versus temperature data on a 4% waxy rice starch dispersion between 25 and 110C. In experiments on tomato puree samples at 95 to 120C, evaporation of water was not significant when the time of operation of the viscometer system was limited to about 20 min. Apparent viscosity data at several shear rates could be reduced to a single curve using a reduced apparent viscosity, q&. Due to shear-induced rupture of granules, values of qa of the starch dispersion were lower than those of q* obtained with a Carri-Med rheometer at about the same shear rates, T = 5.7 s-I and w = 6.3 rad s-', respectively.

NOMENCLATURE
A1 A2 G' G" area under the complex viscosity versus temperature curve (Pa s C) area under the apparent viscosity versus temperature curve (Pa s C) storage modulus (Pa) loss modulus (Pa) complex modulus G * = / m (Pa) consistency index (Pa s") flow behavior index (dimensionless) temperature (C)

G* K n T

Greek letters AA difference in the areas A1 and A2, AA=Al-A2 scaling exponent for reduced viscosity shear rate (s-I) apparent viscosity (Pa s)
P

P T
Tla

reduced apparent viscosity (Pa s)


Tl*

complex viscosity (Pa s), q*=(G*/w)

40

M.A. RAO. H.J. COOLEY and H-J. LIAO

shear stress (Pa) dynamic frequency (rad s-') ACKNOWLEDGMENTS

We are grateful to the USDA for NRI Grant #97-35503493 and the California Natural Products, Lathrop, CA, for donation of starch. REFERENCES ABDELRAHIM, K.A., RAMASWAMY, H.S. and VAN DE VOORT, F.R. 1995. Rheological properties of starch solutions under aseptic processing temperatures. Food Res. Int. 28, 473-480. DAIL, R.V. and STEFFE, J.F. 1990. Rheological characterization of crosslinked waxy maize starch solutions under low acid aseptic processing conditions using tube viscometry techniques. J. Food Sci. 55, 1660-1665. DOLAN, K.D. and STEFFE, J.F. 1990. Modeling rheological behavior of gelatinizing starch solutions using mixer viscometry data. J. Texture Studies. 21, 265-294. JULIANO, B.O.er al. 1981. International cooperative testing of the arnylose content of milled rice. StarchBtiirke 33, 157-162. KEETELS, C.J.A.M. and VAN VLIET, T. 1993. Mechanical properties of concentrated starch system during heating, cooling, and storage. In Food Colloidsand Polymers: Stability and Mechanical Properties, (E. Dickinson and P. Walstra, eds.) pp. 266-271, The Royal Society of Chemistry, Cambridge, UK. TANGLERTPAIBUL, T. and RAO, M.A. 1987. Rheological properties of tomato concentrates as affected by particle size and methods of concentration. J. Food Sci. 52, 141-145. VITALI, A.A. and RAO, M.A. 1984. Flow properties of low-pulp concentrated orange juice: Serum viscosity and effect of pulp content. J. Food Sci. 49,
876-88 1.

YANG, W.H. and RAO, M.A.1998. Complex viscosity-temperature master curve of cornstarch dispersion during gelatinization. J. Food Process Engineering 21, 191-207.

You might also like