Journal of Food Engineering 58 (2003) 253–257

www.elsevier.com/locate/jfoodeng

Ultrasonic determination of the composition of a meat-based product

a,*
S. Simal , J. Benedito b, G. Clemente b, A. Femenia a, C. Rossell
o a

a
Department of Chemistry, University of Illes Balears, Ctra. Valldemossa km 7.5, 07071 Palma de Mallorca, Spain
b
Department of Food Technology, University Polit ecnica of Valencia, Cno. Vera s/n, 46071 Valencia, Spain
Received 19 April 2002; accepted 16 September 2002

Abstract
The use of ultrasonics as an analytical technique to estimate the composition of a fermented meat-based product (sobrassada)
was assessed. Moisture, fat and protein contents and ultrasonic velocity at 4, 8, 12 and 25 °C were measured in samples using
different formulae. In this study, it was considered that a meat-based product is formed of three different constituents: fat, water and
protein þ others. Ultrasonic velocity was related, by a semiempirical equation, to the composition and the ultrasonic velocity of the
above three components. The ultrasonic velocity temperature dependence allowed the determination of fat, moisture and
protein þ others contents. The explained variance was 98.0% for protein þ others, 97.6% for fat and 95.6% for moisture. The results
obtained show the feasibility of using ultrasonic velocity measurement to assess, in a rapid and non-destructively way, the com-
position of a meat-based product.
Ó 2003 Elsevier Science Ltd. All rights reserved.

Keywords: Ultrasound; Sobrassada; Meat; Composition

1. Introduction on-line (Abouelkaram et al., 2000; McClements, 1995).

The chemical composition of meat-based products
has a large influence on their nutritional value, func-
tional properties, sensory quality, storage conditions
and commercial value. Thus, its knowledge is of great
importance for quality control, not only for producers
but also for retailers and consumers. Traditional ana-
lytical techniques are of limited use because they are
time-consuming and destructive, and as a consequence
there is a strong emphasis on the development of rapid
and non-destructive techniques (Chanamai & McCle-
ments, 1999).
Low-intensity ultrasound can be used to provide in-
formation about the physicochemical properties of
many foods (Mulet, Benedito, Bon, & Rossell o, 1999). It
has been used to determine the composition, structure
and the physical state of a large variety of food materials
(McClements, 1994). Ultrasound has advantages over
other traditional analytical techniques because mea-
surements are rapid, non-destructive, precise, fully auto-
mated and might be performed either in a laboratory or
*
Corresponding author. Tel.: +34-971-17-2757; fax: +34-971-17-
3426.
E-mail address: susana.simal@uib.es (S. Simal).
0260-8774/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0260-8774(02)00375-8
Most of the studies found in the literature rely on the
measurement of the ultrasonic velocity, probably due to
the fact that it is the simplest and most reliable ultra-
sonic measurement (Nielsen & Martens, 1997).
One of the most widespread and also most promising
ultrasonic application is the utilization of ultrasound for
composition measurement. Recent studies have shown
that the ultrasonic velocity in fish tissue, chicken and
raw meat mixtures can be related to its composition
(solids non-fat, moisture and fat content) using semi-
empirical equations (Benedito, Carcel, Rossell o, &
Mulet, 2001; Chanamai & McClements, 1999; Ghae-
dian, Decker, & McClements, 1997, 1998). Although
these works have shown the feasibility of using ultra-
sounds for the composition assessment of different ideal
mixtures prepared in the laboratory, more effort should
be done to apply these techniques to the complex
products found in the food industry. Ultrasonics are
widely used for measuring meat composition in both live
animals and carcasses including pork, beef and poultry,
and several commercial ultrasonic instruments have
been developed for this purpose (Cross & Belk, 1994;
McClements, 1997; Park, Whittaker, Miller, & Bray,
1994). These instruments rely on the knowledge of the
relationship between the ultrasonic properties and the
254 S. Simal et al. / Journal of Food Engineering 58 (2003) 253–257
composition of meat, although our understanding is still
fairly limited.
Within this context, the objective of this study was to
examine the relationship between the ultrasonic velocity
measured on a meat-based product (sobrassada de
Mallorca) at different temperatures and its chemical
composition and, thus, assess whether ultrasound could
be used as an useful technique to determine the overall
composition of this type of product.
1.1. Ultrasonic properties of materials
The ultrasonic properties of a multicomponent ma-
terial can be described by a first approximation using the
following relationship (Eq. (1)) (Ghaedian et al., 1997)
where Uj and vj are, respectively, the mass percentage
and ultrasonic velocity in component j:
100 X n
Uj
¼ ð1Þ
v 2
j¼1
v2j
According to different authors, this equation has been
shown to allow a good estimation of the ultrasonic velo-
city in food materials where the scattering is not im-
portant, which implies a low air content (Benedito et al.,
2001; Ghaedian et al., 1998; McClements & Povey,
1992).
2. Materials and methods
Meat-based products used in this study (sobrassada)
were manufactured at the island of Mallorca (Spain) by
a local factory following the methodology suggested by
the protected geographical indication (PGI) for the
‘‘Sobrassada de Mallorca’’. Sobrassada consists of a
mixture of lean pork meat, white fat, paprika and salt.
The meat ingredients were kneaded until particle size of
4 mm was achieved and then mixed with the other in-
gredients. Afterwards, the mince was filled into artificial
casings (25  102  102 m long, 7:5  102 m diameter
having a permeability of 2:68  1012 g/Pa s m at 12 °C
and 75% relative humidity) and ripened at 14 °C and
70% relative humidity. In order to obtain samples with
different composition, sobrassadas were made using
different proportions of meat and white fat.
All samples of sobrassada were analyzed in triplicate
for fat, protein and moisture contents according to the
official methods (ISO R-1443, ISO R-937 and ISO R-
1442).
Measurements of ultrasonic velocity at 4, 8, 12 and 25
°C were carried out by using the experimental set-up
described previously by Benedito et al. (2001) which
consisted of a pair of narrow-band ultrasonic trans-
ducers (1 MHz, 0:7500 crystal diameter, A314S-SU
Model, Panametrics, Waltham, MA), a pulser-receiver
(Toneburst Computer Controlled, Model PR5000-HP,
Matec Instruments, Northborough, MA) and a digital
storage oscilloscope (Tektronix TM TDS 420, Tektronix
Inc., Wilsonville, OR) linked to a personal computer
using a GPIB interface. To compute the time of flight
for a sample, five acquisitions were performed and aver-
aged. All analyses were performed in duplicate on two
different samples without removing the casing. Olive oil
was used as matching material. This procedure was used
to ensure that the measurements were non-destructive,
allowing its use as a quality control method for whole
meat-based products.
3. Results and discussion
Results obtained in the chemical composition mea-
surements and ultrasonic velocity at different tempera-
tures (4, 8, 12 and 25 °C) for the 20 samples analyzed are
shown in Table 1. As can be observed in this table, mean
values of composition, ranged from 43.1% to 67.1% (wet
basis) for fat content, from 19.6% to 37.2% (wet basis)
for moisture content and from 5.0% to 13.5% (wet basis)
for protein content.
With regard to the ultrasonic velocity, this measured
property decreased with the temperature of the sample
(Table 1). As an example, the ultrasonic velocity mea-
sured in five samples of different composition has been
plotted against the sample temperature in Fig. 1. As it
can be seen in this figure, ultrasonic velocity decreased
with temperature according to a straight-line (r2 P 0:99
in all cases), increasing its slope with increase in fat
content and decrease in moisture and protein contents
of the samples. This varied from 1.93 m/s °C for sample
1, with a 43.1% fat content (% wet basis) to 4.21 m/s °C
for sample 20, with a 67.1% of fat content (% wet basis).
Meat-based products are manufactured using diffe-
rent raw materials. In this study, it is considered that
the sobrassada is formed by three different constitu-
ents: fat, water and protein þ others. Other authors
have considered two phases, fat and aqueous phase
(water þ protein þ others) when estimating the compo-
sition of different meat and fish products (Chanamai &
McClements, 1999; Ghaedian et al., 1998). Based on
Eq. (1), the ultrasonic velocity of a meat-based product
can be calculated from the following equation (Eq. (2)):
100 Uf Uw Upþo
¼ 2þ 2 þ 2 ð2Þ
v2 vf vw vpþo
where v is the ultrasonic velocity and vf , vw and vpþo are
the ultrasonic velocities in fat, water and protein þ
others, respectively. Moreover, Uf , Uw and Upþo are the
percentages of fat, water and protein þ others (wet
basis), respectively.
The relationship between water and temperature
(Eq. (3)) was taken from the literature (Kinsler, Frey,
 S. Simal et al. / Journal of Food Engineering 58 (2003) 253–257 255

Table 1
Mean and standard deviation of fat, moisture and protein contents (% wm) and ultrasonic velocity (m/s) at 4, 8, 12 and 25 °C
Sample Fat % wm Moisture % wm Protein % wm v (4 °C) (m/s) v (8 °C) (m/s) v (12 °C) (m/s) v (25 °C) (m/s)
1 43.1  0.6 35.8  0.5 11.0  0.1 1635.0  0.7 1625.5  0.9 1617.0  2.6 1593.8  1.1
2 43.9  0.2 37.2  0.7 9.3  0.3 1624.5  1.2 1616.9  0.6 1607.5  2.0 1583.4  1.1
3 44.3  0.4 33.8  0.1 13.5  0.2 1646.2  1.6 1638.0  2.5 1628.4  0.3 1597.6  1.0
4 45.6  0.3 33.5  0.6 10.5  0.2 1642.3  4.1 1629.8  2.2 1619.1  3.0 1591.4  2.6
5 46.3  0.6 31.7  0.3 10.2  0.1 1649.0  1.1 1638.8  1.5 1625.3  2.2 1597.5  1.5
6 48.3  0.3 32.5  0.3 10.5  0.2 1630.3  1.3 1621.8  2.2 1609.3  3.6 1581.5  1.5
7 48.7  0.7 34.2  0.2 9.7  0.1 1619.6  1.0 1607.7  2.0 1599.1  0.7 1570.8  1.4
8 49.8  0.3 31.9  0.3 9.9  0.1 1627.4  2.1 1615.0  1.7 1605.9  1.4 1571.6  1.1
9 50.8  0.2 30.2  0.2 9.5  0.3 1636.8  1.9 1623.3  1.6 1609.6  2.2 1578.1  1.8
10 52.8  0.4 29.4  0.2 9.3  0.2 1630.4  1.7 1618.7  1.8 1603.8  1.4 1571.0  1.0
11 53.0  0.5 27.6  0.1 10.1  0.3 1641.5  2.2 1629.5  1.7 1615.7  2.5 1578.1  0.9
12 54.5  0.6 28.0  0.3 9.0  0.2 1630.4  3.3 1617.2  3.1 1603.8  1.4 1571.0  0.6
13 57.6  0.5 26.5  0.2 8.4  0.3 1626.6  1.2 1607.8  2.2 1594.7  2.0 1552.7  1.3
14 58.6  0.7 24.9  0.2 8.0  0.3 1629.3  0.8 1613.8  2.1 1597.9  1.9 1555.5  1.8
15 61.5  0.8 22.6  0.2 7.7  0.4 1628.4  3.3 1613.8  1.7 1597.9  1.9 1554.0  0.8
16 62.4  0.6 23.8  0.7 6.6  0.2 1623.4  0.5 1602.2  1.4 1593.1  1.1 1545.5  0.8
17 61.5  0.4 25.1  0.2 5.9  0.4 1613.9  2.0 1602.1  1.3 1584.8  1.3 1543.9  2.1
18 61.8  0.6 22.5  0.4 6.5  0.3 1624.1  1.9 1603.7  1.5 1592.2  1.4 1548.2  0.4
19 63.2  0.5 20.4  0.4 8.0  0.3 1631.7  1.5 1612.7  1.4 1595.4  1.3 1551.4  0.1
20 67.1  0.9 19.6  0.8 5.0  0.5 1625.5  0.9 1607.7  1.9 1588.5  3.3 1536.7  1.4

against temperature for 5 samples with different chemical

composition in Fig. 2. As can be inferred from this figure,
the calculated vpþo decreased with temperature from 4 to
25 °C in all samples. Furthermore, chemical composition
influence was also observed in decreasing the calculated
figures for vpþo when fat content increased and moisture
and protein content decreased. Similar results were
observed for the remaining samples.
From these results, it was concluded that, probably
due to the fact that the protein proportion varies in
‘‘protein þ others’’ between samples with different
chemical composition, the ultrasonic velocity in pro-
tein þ others not only depends on temperature but also
on chemical composition. Thus, an empirical equation
Fig. 1. Influence of sample temperature on ultrasound velocity. Sam-
was proposed to relate the vpþo with temperature and fat
ples of sobrassada with different fat contents (% wet basis).
content (Eq. (5)).

Coppens, & Sanders, 1982). Further, the relationship

between the ultrasonic velocity and temperature in fat
was obtained by Benedito et al. (2001) from experiments
carried out using vacuum dried pork fat (Eq. (4)).
vw ¼ 1403 þ 5T  0:06T 2 þ 0:0003T 3 ð3Þ
vf ¼ 5:6076T þ 1651:7 ð4Þ
As can be observed in Eqs. (3) and (4), the ultrasonic
velocity in water and fat depend on temperature. Thus, it
could be expected that vpþo will also be directly related to
temperature. With the aim of establishing an equation
relating vpþo with temperature, Eqs. (2)–(4) and the ex-
perimental results obtained in the measurements of ul-
trasonic velocity and the experimental percentages of fat
and water contents (Table 1) were used to calculate the Fig. 2. Influence of temperature on ultrasound velocity of
values of vpþo at different temperatures (4, 8, 12 and 25 protein þ others calculated for samples of sobrassada with different fat
°C). As an example, the calculated vpþo has been plotted contents (% wet basis).
256 S. Simal et al. / Journal of Food Engineering 58 (2003) 253–257
vpþo ¼ a1 þ a2 Uf þ a3 T þ a4 T 2
In order to determine the parameters a1 , a2 , a3 and a4
of Eq. (5), calculated values of vpþo at 4, 8, 12 and 25 °C
through Eqs. (2)–(4) for five samples with different
chemical composition (samples 1, 6, 12, 15 and 20) were
used. SOLVER, an optimization tool (GRG2 method)
included in the Microsoft Excel 7.0e spreadsheet was
used, obtaining Eq. (6), which provided the lowest sum
of the square differences between the values calculated
using Eqs. (2)–(4) and those estimated using the pro-
posed Eq. (6).
vpþo ¼ 3250:03  14:49Uf  40:95T þ 0:9190T 2
The percentage of explained variance obtained
through the comparison of the calculated vpþo through
Eqs. (2)–(4) and estimated vpþo by using Eq. (6) was
98.1%. As an example, estimated figures have also been
plotted against temperature in Fig. 2 (continous lines
show for five different samples). Benedito et al. (2001)
found that vpþo was only temperature dependent, which
could be due to the fact that in the present work the
amount of spices added for the sobrassada manufac-
turing greatly affected the component protein þ others.
3.1. Ultrasonic composition assessment
Using Eq. (2) at two different temperatures (T1 and
T2 ) and taking into account that the sum of the com-
ponents of the product is 1, the following three-equation
system was obtained:
100 Uf Uw Upþo
¼ 2 þ 2 þ 2
v2T1 vf T vw T vpþoT
1 1 2 determination. Similar results were found in raw meat
100 Uf Uw Upþo
¼ 2 þ 2 þ 2
v2T2 vf T vw T vpþoT
2 2 2
Uf þ Uw þ Upþo ¼ 100
To improve the accuracy in the computation of Uf ,
Uw and Upþo from Eqs. (7)–(9), it is advisable to consider
two temperatures different enough in order to obtain
significant differences between the coefficients in Eqs. (7)
and (8). The temperatures chosen in this study were 4
and 25 °C.
With the aim of estimating the chemical composition
of the sobrassada samples, fat, moisture and protein þ
others contents, equations representative of vw , vf and
vpþo (Eqs. (3),(4) and (6)) and the results obtained in the
measurements of ultrasonic velocity at 4 and 25 °C for
samples 1–20 were considered. Thus, Uw , Uf and Upþo
were estimated for each sobrassada sample. Fig. 3 shows
the comparison between the estimated composition (fat,
moisture and protein þ others contents) and the experi-
mental values. As can be observed in this figure, there
was a close fit between both estimated and experimental
data.
ð5Þ
ð6Þ
Fig. 3. Estimated vs experimental composition of sobrassada samples:
fat, moisture and protein þ others contents (% wet basis).
In order to mathematically evaluate the accuracy of
the simulation obtained using the proposed method to
determine the chemical composition of a meat-based
product, the percentage of explained variance was
computed (Simal, Sanchez, Bon, Femenia, & Rossell o,
2001). The explained variances obtained by comparison
of the estimated composition and the experimental one
were satisfactory, being 97.7% for the fat content, 96.3%
for the moisture content and 97.0% for the
protein þ others content. In the case of fish composition
assessment, Ghaedian et al. (1998) also found a close fit
between the actual and predicted fat and aqueous phase
ð7Þ content when considering Eq. (2) for the composition
mixtures by Benedito et al. (2001).
ð8Þ
ð9Þ
4. Conclusions
The ultrasonic properties of a meat-based product
depend on composition and temperature. Furthermore,
the semiempirical equations proposed in this study ac-
curately predicted fat, water and protein þ others con-
tents of a meat-based product (sobrassada) from
ultrasonic velocity measurements at two different tem-
peratures.
Although this study was carried out by using a typical
meat product from Spain, the obtained results demon-
strate that a non-destructive technique of ultrasounds
could be used to accurately estimate the chemical com-
position of any meat-based product. As long as the
measurements are carried out on the surface of the
whole sobrassada, ultrasound might offer a rapid and
non-destructive method for analyzing the overall com-
position of a meat-based product for quality control
purposes. Hand-held ultrasonic devices could be
 S. Simal et al. / Journal of Food Engineering 58 (2003) 253–257
developed and used on an industrial basis to determine
meat-based products composition.
Acknowledgement
The authors would like to acknowledge the financial
support of CICYT (FD97-1246-C03).
References
Abouelkaram, S., Suchorski, K., Buquet, B., Berge, P., Culioli, J.,
Delachartre, P., & Basset, O. (2000). Effects of muscle texture on
ultrasonic measurements. Food Chemistry, 69, 447–455.
Benedito, J., Carcel, J., Rossell
o, C., & Mulet, A. (2001). Composition
assessment of raw meat mixtures using ultrasonics. Meat Science,
57(4), 365–370.
Chanamai, R., & McClements, D. J. (1999). Ultrasonic determination
of chicken composition. Journal of Agriculture and Food Chemistry,
47, 4686–4692.
Cross, H. R., & Belk, K. E. (1994). Objective measurements of carcass
and meat quality. Meat Science, 36, 191–202.
Ghaedian, R., Coupland, J. N., Decker, E. A., & McClements, D. J.
(1998). Ultrasonic determination of fish composition. Journal of
Food Engineering, 35, 323–337.
257
Ghaedian, R., Decker, E. A., & McClements, D. J. (1997). Use of
ultrasound to determine cod fillet composition. Journal of Food
Science, 62, 500–504.
Kinsler, L. E., Frey, A. R., Coppens, A. B., & Sanders, J. V. (1982).
Fundamentals of acoustics. New York: John Wiley and Sons.
McClements, D. J. (1995). Advances in the application of ultrasound
and food analysis and processing. Trend in Food Science and
Technology (9), 293–299.
McClements, D. J. (1997). Ultrasonic characterization of food and
drinks: principles, methods and applications. Critical Reviews in
Food Science and Nutrition, 37(1), 1–46.
McClements, D. J., & Povey, M. J. W. (1992). Ultrasonic analysis of
edible fats and oils. Ultrasonics, 30, 383–388.
McClements, D. J. (1994). Ultrasonic NDT of foods and drinks. In
W. J. McGonnagle (Ed.), International Advances in Nondestructive
Testing (vol. 17). Switzerland: Gordon and Breach Science.
Mulet, A., Benedito, J., Bon, J., & Rossell o, C. (1999). Ultrasonic
velocity in cheddar cheese as affected by temperature. Journal of
Food Science, 64(6), 1038–1041.
Nielsen, M., & Martens, H. J. (1997). Low frequency ultrasonics for
texture measurements in cooked carrots. Transactions of the ASAE,
40(6), 1741–1748.
Park, B., Whittaker, A. D., Miller, R. K., & Bray, D. E. (1994).
Measuring intramuscular fat in beef with ultrasonic frequency
analysis. Journal of Animal Science, 72, 117–125.
Simal, S., Sanchez, E. S., Bon, J., Femenia, A., & Rossell
o, C. (2001).
Water and salt diffusion during cheese ripening: effect of the
external and internal resistances to mass transfer. Journal of Food
Engineering, 48, 269–275.