Food Hydrocolloids 13 (1999) 203–210

Characteristics of foam prepared from starfruit (Averrhoa carambola L.)

puree by using methyl cellulose
A. Abd Karim*, C. Chee Wai
Food Technology Division, School of Industrial Technology, Universiti Sains Malaysia, 11800 Penang, Malaysia

Abstract
Foams were prepared from starfruit (Averrhoa carambola L.) puree by adding various concentrations of methyl cellulose (Methocel
65HG). All foams were prepared by whipping using a kitchen mixer. Overrun of foams from various concentrations of methyl cellulose were
compared. Average bubble diameter and bubble size distribution were calculated based on microphotography. Overrun and stability of foams
increased with increasing methyl cellulose concentration until maximum value was obtained at methyl cellulose concentration of 40 g/kg. At
this concentration, the overrun value of 653.2% and foam density of 0.153 g cm 23 was obtained. Higher concentration of methyl cellulose
produced larger air bubbles. The size of the air bubbles showed a wide distribution as the methyl cellulose concentration was increased, with
the average diameter ranged from 176 to 381 mm. q 1999 Elsevier Science Ltd. All rights reserved.
Keywords: Averrhoa carambola L.; Foam; Microphotography; Overrun and stability

1. Introduction are responsible for both the tendency of a liquid to foam and
the stability of the resulting dispersion of bubbles. Such a
An understanding of the principles governing foam compound is usually an existing (such as lecithin in egg
phenomena is necessary in present day food technology. yolk) or added surface-active agent that reduces the surface
From the sparkling foam bubbles of carbonated beverages tension of the liquid and promotes a surface layer of a
to the foamy meringue of whipped egg white, the foam composition different from the rest of the liquid. This
phenomenon and its mechanism is of interest to the food layer, acting as ‘‘buffer’, prevents the natural coalescence
technologist. In other words, whether foam formation or of gas bubbles dispersed in the liquid.
foam destruction, controlling and maintaining the desired Many liquid foods, when converted to stable foams, can
condition is the ultimate goal in producing the desired be dried rapidly in air to yield instant powders of good
product characteristics. quality. It has been shown that fruit and vegetable juices
A foam can be defined loosely as a two-phase system in can be succesfully dried by exposing a layer of stable foam
which a distinct gas bubble phase is surrounded by a contin- to a current of hot air under atmospheric pressure (see
uous liquid lamellar phase (see German, O’Neill, & Morgan & Ginnette, 1960; Ginnette, Graham, Miers, &
Kinsella, 1985). A consequence of this dispersion is a Morgan, Jr., 1963; Bisset, Tatum, Wagner, & Veldhuis,
very large gas–liquid interface. As interfaces between 1963; and Berry, Bisset, Wagner, Jr., & Veldhuis, 1965).
non-miscible phases are under tension, the expansion of The process, known as foam-mat drying, consists essentially
these interfaces requires energy, i.e., work is performed in of three steps, namely: (a) formation of stable foam contain-
forming them and energy is released on their relaxation, ing the product to be dried, (b) air drying of the foam to form
hence foams are highly unstable (see German et al., a thin porous sheet or mat, and (c) compression of the dried
1985). For some food products (e.g. salad cream, cake mat followed by disintegration to yield a free flowing
batter), it is desirable not only to achieve good foaming powder.
properties, but also to form stable foams. The technological The foam-mat drying process is relatively simple and
art is to control the rates of the processes which promotes inexpensive. One difficulty that was experienced previously
instability. An essential ingredient in a liquid-based foam is with this process, however, is the lack of stability of the
surface-active molecules. These reside at the interfaces and foam during the heating cycle. If the foam does not remain
stable, cellular breakdown occurs causing serious impair-
* Corresponding author. Tel.: 1 60-4-6577888; fax: 1 60-4-6573678. ment of the drying operation. Variables affecting foam
E-mail address: [email protected] (A.A. Karim) formation, density and stability have been reported, Hart,
0268-005X/99/$ - see front matter q 1999 Elsevier Science Ltd. All rights reserved.
PII: S0268-005 X( 98)00 086-1
204 A.A. Karim, C.C. Wai / Food Hydrocolloids 13 (1999) 203–210

Fig. 1. Foam density, percent overrun and drainage rate as a function of methocel concentrations.

Graham, Ginette, and Morgan, Jr. (1963), and these include puree and mango puree have also been reported (Bates,
chemical nature of the fruit, soluble solids, pulp fraction, 1964). However, to the best of the authors’s knowledge,
type and concentration of foaming agent, and type and no work has been reported on the foaming characteristics
concentration of foam stabiliser. of starfruit (Averrhoa carambola L.) puree. Therefore, the
Of the two main categories of food macromolecules — objective of the present study was to study the foaming
carbohydrates and proteins — it is the proteins that domi- characteristics of starfruit puree by using methyl cellulose
nate in most food foams and emulsions. Some use is made of in relation to the foam characteristics and stability for foam-
modified carbohydrates, however, e.g. cellulose derivatives mat drying process.
in artificial creams and propylene glycol alginate in salad
dressings (see Stainsby, 1989). Stanley, Goff, and Smith
2. Materials and methods
(1996) reported that the hydrocolloids were effective in
increasing the stability of whipped cream through enhanced
Starfruits (Averrhoa carambola L., B10 clone) were
viscosity that prevented drainage during the storage period.
acquired fresh from the local market or from the farm
Prin (1988) has suggested that one role of hydrocolloid
around Seberang Prai. The fruits selected were free from
stabilisers in foams is to provide a yield value in the liquid
physical damage, ranged from 8 to 10 cm in length. Total
phase that keeps the bubbles in place.
soluble solids of the juice was 7.08Brix and pH in the range
Methyl cellulose was used successfully to aid in foam
of 3.80–3.95. The fruits were washed in excess water,
formation in fruit puree destined for foam-mat drying
followed by soaking for 1 min in 1 mol/l HCl and rinsing
process. Methyl cellulose is a cellulose derivative produced
with water to remove dirt and pesticide residues.
by etherification of alkaline cellulose with methyl chloride
to form the cellulose ethers (see Grover, 1984). The unique 2.1. Preparation of the starfruit puree and foam
property of the formation of completely reversible gels of
methyl cellulose upon heating the solution is responsible for The starfruits were cut with a stainless steel knife to
many of the food use applications, which include among remove the edible portion (which constituted about 91.6%
others, film formation in fried foods, foam stabilisation, of the whole fruit) and chopped into small pieces. The fruits
and stabilisation of fruit pie filling during baking. The were then blended in Moulinex 276 blender at maximum
surfactancy of methyl cellulose improves the whippability speed for 2 min to obtain a homogeneous puree. The puree
of cake batters in which a portion of the egg whites were was pasteurised at 878C for 3 min to inhibit microbial and
replaced by the methyl cellulose. The batter foam stabilisa- enzyme activity such as pectinase and polygalacturonase.
tion resulting from the surface active behaviour of the Excess heating of the puree was avoided by cooling the
methyl cellulose is further reinforced by the thermal gela- puree immediately in the ice bath after the stipulated
tion of the gum while baking (see Grover, 1984). pasteurisation time.
Foaming conditions of various tropical fruits such as Methocel 65 HG (Fluka Biochemika 64670) was used as
pineapple, passion fruit, guava puree, banana fruit, papaya a foaming agent and contains 27%–29% methocyl group. It
 A.A. Karim, C.C. Wai / Food Hydrocolloids 13 (1999) 203–210 205

Fig. 2. Photomicrographs of starfruit foam showing the changes in number and size of air bubbles at different methocel concentrations: (a) 0 g/kg; (b) 10 g/kg;
(c) 20 g/kg; (d) 30 g/kg; (e) 40 g/kg; and (f) 50 g/kg. The scale bar represents 1 mm.

has a solubility of 2% at 258C and a viscosity of 4000 cp. was done by blending the mixture in Kenwood Chef (KM
Methocel was prepared in the form of solution by dissolving 901D model) at maximum speed for 4 min at 308C.
an appropriate amount of the dry powder in 20 ml distilled
water at 308C and stirred slowly until a uniform solution was 2.2. Determination of foam density and foam overrun
obtained. Methocel solutions were prepared to give a final
concentration of 10, 20, 30, 40 and 50 g/kg. Foam density was determined using a method described
Twenty grams of methocel solution of a certain concen- by Labelle (1966). The density of methocel–puree mixture
tration was mixed with 180 g of the starfruit puree to give a was determined by weighing 100 ml of the mixture in a
final concentration of methocel in the mixture at 10, 20, 30, 100 ml measuring cylinder. For the foam, 200 ml of the
40 and 50 g/kg. Foaming of the methocel–puree mixture foam was transferred into a 250 ml measuring cylinder
206 A.A. Karim, C.C. Wai / Food Hydrocolloids 13 (1999) 203–210

Fig. 3. Average bubbles diameter (DT) and average percentage area of air bubbles at different methocel concentrations.

and weighed. The foam transferring was carried out very The foam sample was gently pressed down by a coverslip.
carefully to avoid destroying the foam structure or trapping This provided a sample of the foam about 1 mm thick.
the air voids while filling the cylinder. The determinations Approximately equal amounts of foam were placed on the
were done in triplicate for each batch of preparation and the slide in each tests. The covered portion was photomicro-
average values were reported. graped immediately at magnification 12 × with a photomi-
The foam overrun was calculated based on AOAC (1984) croscope Wild MP552 mounted with Wild MPS 46
and modified for foam: photoautomat using Kodak 100ASA film. In all cases, at
Foam density least three samples were observed and photograph from
Overrun ˆ × 100% each foam studied. Field size for counting purposes was
Density of methocel–puree mixture
1 mm. The average diameter (DT) of the air bubbles was
In order to study the effect of variation in the raw materi- determined by measuring the diameter of each bubbles in
als used, this experiment was repeated four times using the 3R photo which has a total area of 30 mm 2. The number
different batch of starfruits. Experimental design based on of air bubbles in each photograph was around 150–200.
randomised block design was employed to verify that the
effect of different sources of raw material will not interfere d 1 1 d2 1 d3 1 … 1 dn
with the interpretation of the results. Average diameter DT † ˆ ;
n
2.3. Determination of foam stability
where dn is the diameter of nth air bubble and n is the
Foam drainage method was adapted from a method number of air bubbles in the photograph.
described by Sauter and Montoure (1972). The foam was Bubble size distribution was determined by grouping the
filled into a conical-shaped plastic wire mesh (10 mesh) bubbles based on the diameter. Bubbles were classified by
which was supported by a thistle funnel and placed on diameter at 100 mm interval in the range of 150–950 mm.
250 ml graduated cylinder. The apparatus assembly was Using the aforementioned methods of foam evaluation, the
put in the mini kiln smoker (used for separate drying kinetic effect of the conditions of foam preparation on the bubble
experiment) at 708C and air flow rate of 0.12 ms 21. The size distribution, diameter and density were studied. The
liquid juice which separated from the foam as a result of relationship between bubble size and density at different
drainage was collected in the measuring cylinder. The stabiliser concentration was analysed.
amount of juice collected after a certain time interval was
recorded.
2.5. Experimental design and data analysis
2.4. Determination of average diameter and air bubble size
distribution Completely randomised block design with four replica-
tions was adopted in this study. Statistical analysis on the
The foam sample was prepared by placing suitable data was carried out with MINITAB Release 8 (Minitab
amount of foam on a clean glass slide (2.5 × 7.5 cm 2). Inc.) for Macintosh.
 A.A. Karim, C.C. Wai / Food Hydrocolloids 13 (1999) 203–210 207

3. Results and discussion concentrations of protein which prevent incorporation of

air into liquid.
3.1. Effect of methyl cellulose concentration on the density The influence of pulp particles present in the methyl
and overrun cellulose–puree mixture is difficult to predict. Prins
(1988) suggested that the presence of particles of the proper
Preliminary experiments with several types of foaming size, shape and wetting behaviour can stabilise the films
agents showed that methocel 65HG produced satisfactory through a ‘Pickering mechanism’. To ensure that they are
foam at the lowest concentration. Analysis of variance kept at the air–water interface, the surface of the particles
showed that the source of starfruit had no significant effect must be such that the contact angle lies in the range from 308
(p . 0.05) on the foam overrun as compared to the effect of to 908. When two such surfaces come close together, further
methyl cellulose. drainage of film liquid is prevented by the particles which
Methyl cellulose can reduce the surface tension and inter- act as spacers. Hart et al. (1963) reported that the amount of
facial tension in aqueous system in the range of 41–55 and soluble solids needed to produce a stable foam can be
18–28 dyn cm 21, respectively (see Greminger & Krumel, greatly reduced by the addition of a small amount of pulp.
1980). In addition, methyl cellulose aids in the formation For example, a lemon concentrate containing no pulp must
of strong film and stabilised the interfacial film. Fig. 1 shows be dried with 60 g/100 g soluble solids, whereas a concen-
the effect of methyl cellulose on the foam overrun. As the trate containing 3 g/100 g pulp can be dried with as little as
concentration of methyl cellulose in the puree increased, the 10 g/100 g soluble solids.
foam overrrun also increased significantly. Higher overrun
indicates that more air was trapped in the foam and this gave 3.2. Effect of methyl cellulose concentration on foam
rise to lower foam density. This can be seen more clearly by stability
examining the changes in density which shows that the foam
density decreased with increasing methocel concentration. Fig. 1 shows the drainage rate as a function of different
The foam overrun increased drastically at 30 g/kg methyl concentrations of methocel. Drainage rate in this context
cellulose concentration (Fig. 1). This indicates that at least reflects the water holding capacity of the foam. All liquid
30 g/kg of methyl cellulose was required to reduce the foams are subject to drainage of the liquid from between the
surface tension and interfacial tension to a level sufficiently bubbles caused by the action of gravity. Drainage is accom-
low to form the interfacial film. At lower concentration of panied by a progressive thinning of the lamellae and may,
methyl cellulose (, 30 g/kg), the instability of air bubbles therefore, enhance the probability of film collapse (see
could be attributed to non-optimal or non-uniform coverage Durian & Weitz, 1994). It can be seen that there was a direct
of the surface active agent, because of the fact that migra- relationship between foam density and the extent of drai-
tion of surface active agent from the bulk towards the nage. In other words, foams with higher concentration of
expanding surface is diffusion limited (see Stanley et al., methocel (lower density) exhibited less drainage as
1996). compared to higher density foams (lower concentration of
As the concentration of methyl cellulose was increased, methocel). This observation is in marked contrast to the
the overrun also increased until maximum value was finding of Brygidyr, Rzepecka, and McConnel (1977) for
obtained at methyl cellulose concentration of 40 g/kg. At tomato puree and Labelle (1966) for orange juice; they
this concentration, the overrun value of 653.2% and foam reported that foams with more additives collapse faster.
density of 0.153 g cm 21 was obtained. However, increasing This conflicting observations may be attributed to the
the methyl cellulose concentration further, beyond this types of foaming agents or stabiliser used; Labelle (1966)
point, seems to produce the opposite effect, i.e. the overrun used glyceryl monostearate as a foaming agent and metho-
decreased. This phenomenon was not reported before in the cel as a stabiliser and Brygidyr et al. (1977) used a mixture
foaming studies using methyl cellulose as the foaming of mono- and diglyceride. In addition, the nature of the raw
agent. It has been suggested (Bikerman, 1973) that high material (i.e. chemical compositions) may affect the stabi-
viscosity liquid would prevent the trapping of air during lity of the foam. Cooke, Breag, Ferber, Best, and Jones
whipping or mechanical mixing. Increasing methyl cellu- (1976) reported that for mango puree, increasing the addi-
lose concentration higher than 40 g/kg would also increase tive concentration resulted in reduced foam density and
the viscosity of the mixture, possibly exceeding the limiting gave a lower loading of foam per surface area of tray,
viscosity at which maximum volume of air can be incorpo- which agrees with our observation. It should be noted that
rated, thus resulting in reduction of the foam overrun. foam drainage is undesirable in foam-mat drying as accu-
Similar observations have been reported on the effect of mulation of the liquid at the conveying surface leads to more
protein concentration on foam expansion and stability. Foam difficult removal of the dried product from the tray of the
overrun have been reported to increase with increasing belt (see Labelle, 1966).
concentration up to a limit, and then decrease (see In foam-mat drying, foam stability is very critical because
Cherry & McWatters, 1981; and Vani & Zayas, 1995). the foam should be able to retain its open structure through-
This was attributed to high viscosity with increasing out the drying process in order to increase the total surface
208 A.A. Karim, C.C. Wai / Food Hydrocolloids 13 (1999) 203–210

Fig. 4. Size distribution of air bubbles in starfruit foam at different methocel concentrations; (a) 0 g/kg (b) 30 g/kg (c) 50 g/kg.

area and capillary effect during drying (see Hart et al., nearly rigid walls, and that the rate of drainage usually falls
1963). In addition, the foam should be ‘‘stiff’’ enough to with time. We also found (data not shown) that the rate of
prevent it from flowing through the wire mesh support drainage was rapid for the first 10 min and then declined or
during drying. The open foam structure is desirable for reached a plateau. According to Mita, Ishida, and Matsu-
rapid drying and ease of removing the dried material from moto (1978), the rate naturally depends on the viscosity of
the tray. If foams break or drain excessively, drying time is the bulk liquid phase and technologically this is raised by
increased, reducing product quality. In contrast, excessively the addition of thickeners, often hydrocolloids.
stable foams retain too many gas inclusions in the dry state. All foams, however, would eventually collapse (see
Upon reconstitution, these produce suds and opacity, and Bikerman, 1973). According to DeVries (1958), foam stabi-
reduce apparent colour intensity (see Hart et al., 1963). lity is influenced by the thickness of the interface, foam size
During drainage, the adsorbed film is subjected to shear distribution, interface permeability and surface tension.
(see Stainsby, 1989). Halling (1981) points out that stability During drying, foam drainage and/or foam collapse may
of foam to drainage requires that the surface films behave as occur. Drainage would result in the separation of the juice
 A.A. Karim, C.C. Wai / Food Hydrocolloids 13 (1999) 203–210 209

Fig. 4. (continued)

(liquid phase) from the interface layer, hence results in thin- the DT value also increased until a maximum value of 381 ^
ning of the interface which eventually leads to foam collap- 25 mm, then decreased again. According to Berry et al.
sing. Foam collapsing during ageing is also caused by (1965), a suitable foam for drying should contain small
gravity force acting on the foam and the increase in foam and uniform size bubbles. He established a relationship
size, as a result small foams combine or merge together (see between the size of air bubbles and the rate of evaporation.
Labelle, 1966; and DeVries, 1958). This statement was supported by Torrey (1974) who
suggested that the optimum size of air bubbles suitable for
3.3. Effect of methyl cellulose concentration on foam drying should be 100 mm or less. The reason is that small air
bubbles bubbles increase the capillary effect, thus facilitate the mass
transfer rate during drying. Our experiment, however,
Fig. 2(a)–(d) shows the photomicrograph of the foam at produced larger bubbles even in puree without any methyl
different methyl cellulose concentration. It can be seen that cellulose addition. At lower concentration of methyl cellu-
the foams formed at different concentrations of methyl lose, relatively small bubbles can be obtained; however,
cellulose were bubbly foam, i.e. each bubble had their only small number of air bubbles can be generated and
own interfacial film. According to DeVries (1958), stability large part of the puree still exist in a liquid phase (Fig. 2(b)).
of round-shaped, small air bubble surrounded with liquid
phase is high. At methyl cellulose concentration of less 3.4. Effect of methyl cellulose on bubble size distribution
than 30 g/kg, the air bubbles were round in shape and well
separated from each other. At higher concentration, Bubble size distribution at different methyl cellulose
however, the bubbles were more packed and touched each concentrations is shown in Fig. 4(a)–(c) (shown for 0, 10,
other at the interfacial wall. Some of the air bubbles could 30, and 50 g/kg methyl cellulose). It is obvious that at low
not retain the round shape and became elongated, possibly concentration, there were more small air bubbles present in
because of the force exerted by the neighbouring bubbles the foam. As the concentration of methyl cellulose
and also perhaps because of excess air trapped in the bubble increased, the formation of large air bubbles also increased
(more than the interfacial film can withstand). Fig. 3 correspondingly. Evidently, at higher methyl cellulose
presents the result in terms of percentage area of air bubbles. concentrations, the size distribution of the air bubbles
This figure clearly shows that increasing the methyl cellu- were more spread and approached a normal distribution.
lose concentration also increased the percentage area which In practice, this phenomenon can be considered a disadvan-
resulted from increase in the number of the air bubbles. tage because if the foam is to be dried, the small and uniform
Fig. 3 also illustrates the effect of methyl cellulose size air bubbles are more desirable. It should be noted,
concentration on average diameter (DT) of the air bubbles. though, that the method we used to estimate bubble size
Without methyl cellulose in the puree, the smallest DT was distribution by pressing down the cover slip may cause
176 ^ 11 mm. As methyl cellulose concentration increased, distortion of large bubbles, thereby giving a misleading
210 A.A. Karim, C.C. Wai / Food Hydrocolloids 13 (1999) 203–210

tail to the bubble-size distribution. Rapid ageing of foam Brygidyr, A. M., Rzepecka, M. A., & McConnel, M. B. (1977). Can. Inst.
and difficulty in placing it in known quantity on the test Food Sci. Technol. J., 10 (4), 313–319.
Cherry, J. P., & McWatters, K. H. (1981). In: Cherry, J. P. (Ed.), Protein
slide make it an awkward aspect of the study. functionality in foods (pp. 149–149). ACS Symposium Series, 147.
Cooke, R. D., Breag, G. R., Ferber, C. E. M., Best, P. R., & Jones, J. (1976).
J. Food Technol., 11 (5), 463–473.
4. Conclusions DeVries, A. J. (1958). Rubber Chemistry and Technology, 31, 1142–1205.
Durian, D. J., & Weitz, D. A. (1994). In M. H. Grant (Ed.), Encyclo-
Evidently, a foam could not be formed from the starfruit pedia of chemical technology, (pp. 783). New York: Wiley-Inter-
puree without the addition of a suitable foaming agent. Air Science Publ.
bubbles in the puree were relatively small and large part of German, J. B., O’Neill, T. E., & Kinsella, J. E. (1985). J. Am. Oil Chem.
Soc., 62 (9), 1358–1366.
the puree still exists in the liquid phase. In contrast, addition
Ginnette, L. F., Graham, R. P., Miers, J. C., & Morgan Jr., A. I. (1963).
of foaming agent, methocel 65HG was found to increase the Food Technol., 17 (6), 133–135.
foaming capacity of the puree significantly (p , 0.05). Greminger Jr., K. G., & Krumel, K. L. (1980). In R. L. Davidson (Ed.),
Increasing the methyl cellulose concentration also increased Handbook of water-soluble gums and resins, (pp. 1). New York:
the size of the air bubbles which may be undesirable McGraw-Hill Book Company.
Grover, J. A. (1984). In M. Glicksman (Ed.), (pp. 121). Food hydrocolloids,
because it will affect the duration of drying in foam-mat
3. Florida: CRC Press.
drying process. The foaming process of starfruit puree Halling, P. J. (1981). Crit. Rev. Food Sci. Nutr. 15 (2), 155–203.
with methyl cellulose has a great potential to be developed Hart, M. R., Graham, R. P., Ginnette, L. F., & Morgan Jr., A. I. (1963).
further possibly by using other combinations of surface- Food Technol., 17 (10), 90–92.
active agent and foam stabiliser in order to get improved Labelle, R. L. (1966). Food Technol., 20, 1065–1070.
Mita, T., Ishida, E., & Matsumoto, H. (1978). J. Colloid Interface Sci., 64,
foam characteristics suitable for foam-mat drying process.
143–147.
Morgan, A. I. Jr, & Ginnette, J. M. (1960). US Patent 2955943.
Prin, A. (1988). In E. Dickinson & G. Stainsby (Eds.), Advances in food
References emulsions and foam, (pp. 91). London: Elsevier Applied Science.
Sauter, E. A., & Montoure, J. E. (1972). J. Food Sci., 37 (6), 918–920.
AOAC (1984). Association of Official Analytical Chemists (13th ed.). Stainsby, G. (1989). In J. R. Mitchell & D. A. Ledward (Eds.), Functional
Washington, DC: AOAC. properties of food macromolecules, (pp. 315). London: Elsevier
Bates, R. P. (1964). Food Technol., 18, 93–96. Applied Science.
Berry, R. E., Bisset, O. W., Wagner Jr., C. J., & Veldhuis, M. K. (1965). Stanley, D. W., Goff, H. D., & Smith, A. K. (1996). Food Res. Int., 29 (1),
Food Technol., 21 (3), 126–128. 1–13.
Bikerman, J. J. (1973). Foams. New York: Springer-Verlag. Torrey, M. (1974). Dehydration of fruit and vegetables. London: Noyes
Bisset, O. W., Tatum, J. H., Wagner, C. J., & Veldhuis, M. K. (1963). Food Data Corporation.
Technol., 17 (2), 92–95. Vani, B., & Zayas, J. F. (1995). J. Food Sci., 60 (5), 1025–1028.