Journal of Food Engineering 77 (2006) 550558

www.elsevier.com/locate/jfoodeng

Microwave assisted drying of macadamia nuts

F.A. Silva a,1, A. Marsaioli Jr. a,*, G.J. Maximo a,1
,
M.A.A.P. Silva b,2, L.A.G. Goncalves c,3
a
Food Engineering Department, State University of Campinas, Campinas, Sao Paulo, Brazil
b
Nutrition and Food Department, State University of Campinas, Campinas, Sao Paulo, Brazil
c
Food Technology Department, State University of Campinas, Campinas, Sao Paulo, Brazil

Received 11 April 2005; accepted 28 June 2005

Available online 3 November 2005

Abstract

The macadamia tree (Macadamia integrifolia Maiden & Betche) belongs to the botanical family Proteaceae and is native to the
Australian tropical forests. The most critical stage of macadamia nut processing is the drying of the kernels, as a minimum quality
standard can only be achieved under specic controlled conditions. Otherwise there is deterioration of the nut and devaluation of
kernel quality. This work aimed at verifying the feasibility of producing dried macadamia nuts by applying microwaves to assist the
hot air drying process, thus reducing the drying time and increasing the industrial yield and quality of the kernels as compared
to those from conventional processes. Another specic objective of this study was to analyse and model the drying kinetics of mac-
adamia nut processing using microwaves. The quality of the kernels was controlled for a period of six months after processing by
determining the peroxide values, free fatty acid percentages and sensory acceptance evaluations. The data obtained from these anal-
yses were well below the limit stipulated by the Brazilian legislation and the product was very well accepted in the sensory
evaluation.
 2005 Elsevier Ltd. All rights reserved.

Keywords: Macadamia nuts; Drying; Microwaves

1. Introduction explains why these oils are highly valued by the cosmetic
industry for use in the composition of hydrating lotions,
The macadamia tree (Macadamia integrifolia Maiden and by the pharmaceutical laboratories as a reducer of
& Betche) belongs to the botanical family Proteaceae. cholesterol levels. The oil can also be used for seasoning
Its nuts are rich in mono-unsaturated oils, which salads and for cooking. Roasted, salted macadamia nuts
are widely consumed as an appetizer, and it is believed
that more than half the world production ends up as
*
roasted snacks as an accompaniment for drinks on
Corresponding author. Present address: Faculdade de Engenharia
de Alimentos, UNICAMP, Caixa Postal 6121, CEP13083-862,
social occasions. They are also excellent in the formula-
Campinas, SP, Brazil. Fax: +55 19 37884027. tion of cakes, crackers, chocolates and sweets (Toledo
E-mail addresses: aviocamp@gmail.com (F.A. Silva), tonymars@ Piza, 2000).
fea.unicamp.br (A. Marsaioli Jr.), maximo@fea.unicamp.br (G.J. Similarly to other nuts the macadamia kernel con-
Maximo), mariasi@fea.unicamp.br (M.A.A.P. Silva), lireny@fea. tains a high amount of oil. Mason and Mc Conachie
unicamp.br (L.A.G. Goncalves).
1
Fax: +55 19 37884027.
(1994) reported that the oil content of the Macadamia
2
Fax: +55 37884059. integrifolia nut varies between 66.3% and 81.2%, with
3
Fax: +55 37883983. a product moisture content between 1% and 1.5% d.b.,

0260-8774/$ - see front matter  2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2005.06.068
 F.A. Silva et al. / Journal of Food Engineering 77 (2006) 550558
depending on variety, environmental nature, growing
practices, etc.
The fatty acids normally described in macadamia
nuts are: myristic (0.61.8%), palmitic (724%), palmit-
oleic (1530%), stearic (25%), oleic (4065%), linoleic
(1.44.5%) and arachidic (1.24.5%) (Cavaletto, 1983;
Duke, 1989; Kaijser, Dutta, & Savage, 2000).
The macadamia orchards in Brazil are young and still
increasing in production. During the 2004 harvest sea-
son it was estimated that more than 3300 metric tons
of in-shell nuts were collected, ranking Brazil as the sev-
enth largest macadamia producer in the world. The big-
gest macadamia world exporting producers are
Australia, South Africa, Guatemala, Brazil and Costa
Rica, according to the United States Department of
Agriculture (USDA, 2002).
The industrial drying process currently employed for
macadamia nuts has the disadvantage of requiring a
very long total drying cycle, extending for a period of
more than one month. The rst stage starts when the
recently harvested and hulled in-shell nuts, with a mois-
ture content from 20% to 30% (d.b.) in the eld, are sub-
mitted to drying by natural air convection in the shade,
using boxes, trays and aerated silos, a procedure that
takes three to four weeks in order to reduce the product
moisture content to 10% (d.b.). This pre-dried product is
then submitted to the next drying stage, carried out in an
industrial plant located in the production area, loading
the product into a silo-type dryer and applying forced
hot air convection under controlled temperature condi-
tions, starting at 40 C and nishing at 60 C, in order
to produce a nal dried product with a moisture content
of 3.5% (d.b.) within about six days (144 h). The process
occupies enormous areas, with excessive handling of the
in-process material, making it dicult to control the
whole operation and leading to considerable risks of
contamination and loss of quality of the nal dried ker-
nel, apart from the considerable costs involved (Toledo
Piza, 2000).
Mason and Wills (2000) reported that drying is one of
the most important stages in the processing of macad-
amia nuts, it being mandatory that the drying operation
be started soon after harvesting to prevent any hydro-
lytic rancidity or mould development. The same author
states that kernels with a moisture content of about
1.5% (d.b.) can be stored for up to one year without sig-
nicant losses in quality.
Chung and Furutani (1989) tried to dry macadamia
nuts using microwaves, but came to the conclusion that
microwave energy could not be applied to drying these
nuts, since they failed to produce a good quality
product.
Microwave energy has been tested in several drying
studies, the results being highly satisfactory. Sousa,
Pitombo, da Silva, and Marsaioli (2001) studied the
application of microwaves in the drying of bananas
551
and concluded that not only was the drying time
much shorter, but that the sensory acceptance of the
microwave processed samples was much higher than
that of commercial products. Berteli and Marsaioli
(2005) reported that the drying time of short-cut pasta
by a combination of hot air and microwaves was
thirteen times shorter than that of the conventional
drying process. Silva and Marsaioli (2003) also tested
a similar combined process for drying Brazil nuts,
obtaining much shorter drying times in relation to
those of conventional industrial drying, the dried prod-
uct remaining in good condition for more than six
months.
2. Material and methods
2.1. Raw material
The raw material used in this research was hulled
pre-dried in-shell macadamia nuts (M. integrifolia Mai-
den & Betche), with an average moisture content of
10% d.b., from the 2004 harvest season. It was pro-
vided by one of the major processors of macadamia
nuts in Brazil, Queen Nut Macadamia, from Dois
Corregos, SP, located 180 km from Campinas. The fac-
tory also dried a sample from the same batch of nuts by
the conventional process, to a moisture level of 3.5%
(d.b.), providing kernel samples at 1.5% (d.b.). These
samples were delivered to the Food Engineering
Department (DEA), UNICAMP, in Campinas, for a
sensory acceptance evaluation and for analyses of per-
oxide value and free fatty acid percentage. In order to
prevent moisture loss from the samples during trans-
portation, they were packed and sealed into laminated
plastic bags with nitrogen. Immediately after sealing,
the bags were sent by car, arriving at DEA in less than
4 h.
2.2. Equipment
A domestic 38-l microwave oven (Brastemp model
BMV38-A) with a nominal power of 900 W, was
adapted to allow for the control of internal air circula-
tion and removal of the moisture generated by the dry-
ing operation. The oven was also equipped with an
applied microwave power control system as a function
of the product temperature, measured by an infrared
probe (Fig. 1).
2.3. Methodology
The moisture content of the samples was determined
by method number 92540Moisture in Nuts and
Products of Nuts(AOAC, 1997). The methods Cd
8b-90 and Ca 5a-40 (AOCS, 2002) were applied for
552 F.A. Silva et al. / Journal of Food Engineering 77 (2006) 550558

16 17
2 8

7 13
5
3 4 14
1 11
6 10
9 12 15

Fig. 1. Adapted domestic microwave oven: (1) air inlet; (2) pressure meter; (3) valve; (4) fan; (5) temperature controller; (6) electric heater; (7) two
channel temperature meter; (8) two channel temperature recorder; (9) quick clamp connector; (10) air diuser; (11) microwave cavity; (12) sample
baskets; (13) exhausting window; (14); air outlet; (15) cooling air of the microwave generator; (16) infrared sensor signal; (17) set point temperature
on the controller to switch the magnetron on/o.

Table 1 average product moisture at time t (min); Me (%, db)

Experimental design levels of the microwave drying process
Independent variables 1 0 1
Air temperature at entrance Tair (C) 58 60 62
Set point temperature Tsp (C) 64 66 68
the determinations of peroxide value and free fatty acid
percentage, respectively. After completing the process-
ing operations, the nal dried kernels were packed into
bags made of nylon and low density polyethylene, and
sealed under vacuum in order to best preserve them dur-
ing the storage period at room temperature (25 3 C),
considering 180 days to be an adequate shelf life for this
product.
In order to study the inuence of some variables and
the behaviour of the macadamia nuts dried by micro-
waves, an experimental design was established using a
complete 22 factorial plan with 3 central points, giving
a total number of seven tests. The variables studied
and their levels are shown in Table 1. The dependent
variables or responses are: drying time (min), peroxide
value (PV) and free fatty acid percentage (% FFA).
The air velocity was xed at 0.3 m/s. The microwave
power was set at 300 W, a value that would provide
a power density (watts divided by initial sample
weight = 0.33 W/g) that would maintain the product
temperature below the prescribed 70 C maximum,
using the temperature controller (17) shown in Fig. 1.
The theory of water migration by diusion, based on
Ficks second law, was considered in the study of the
drying kinetics. Assuming the macadamia nut to be rep-
resented by a sphere, Eq. (1) was used, according to
Crank (1975), who proposed its use for products of this
geometry:
 2  was 4. Panellists were requested to evaluate how much
Mt  Me 6 X1
1 n Def p2
MR 2 exp  t 1
M 0  M e p n1 n2 r2
where n is the number of terms of the Fourier series; MR
is the moisture ratio (dimensionless); Mt (%, db) is the
is the equilibrium product moisture content, calculated
as 1.8 0.1 (%, db), according to Xavier, Toledo Piza,
J. A. N., Toledo Piza, P. L. B., and Valarelli (1996);
M0 (%, db) is the initial product moisture; Def (m2/
min) is the eective diusivity and r (m) is the radius
of the sphere. The eective diusivity was calculated
using Eq. (2)
Def p2
k 2
r2
The adjustment of the models to the experimental
data from the drying kinetics was accomplished by
means of linear and non-linear regression procedures
from Statistica version 5.0 (Microsoft, 1995). The crite-
ria for choosing the best adjustments were the regression
coecient (R2) determined from the observed responses
and the values predicted by the model (Barros Neto,
Scarmino, & Bruns, 1996), and the modulus of the mean
relative deviation (P), whose denition is described by
Eq. (3) (Lomauro, Bakshi, & Labusa, 1985)
100 X
N
jV P  V 0 j
P 3
N i1 V0
where P is the mean relative deviation (%); VP is the val-
ues predicted by the model; V0 is the observed responses
and N is the number of experimental points.
Sensory acceptance tests were carried out at three
points: zero time, 90 and 180 days of storage, using
thirty panellists in individual booths with white illumi-
nation. The samples, equivalent to three kernels or
about 10 g each, were served in a monadic way in dishes
coded with a random three-digit number. The maximum
number of sample evaluations per session per panellist
they liked or disliked the sample using a structured 9-
point mixed hedonic scale. Evaluations were made for
each sample with respect to overall acceptance, appear-
ance and crispness. The data from the acceptance evalu-
ation were submitted to a variance analysis (ANOVA)
 F.A. Silva et al. / Journal of Food Engineering 77 (2006) 550558 553

with two sources of variation (panellist, sample) and a
Tukeys means test (OMahony, 1986).
3. Results and discussion
3.1. Drying kinetic
Table 2 shows the results obtained from the com-
bined hot air and microwave drying process, carried
out in the adapted microwave oven. The values M0
and Mf refer to the initial and nal product moisture
contents (dry basis). The initial nut mass was xed at
0.9 kg for all tests, the microwave power was 300 W
for all conditions and the hot air mass ow rate and
velocity were established at 2.8 m3/min and 1.0 m/s,
respectively. The samples were weighed every 30 minutes
up to the end of the drying period. The drying proce-
dures continued until the macadamia in-shell nut mois-
ture content reached a value close to 3.7% d.b., that is,
the nal kernel moisture was close to 1.5% d.b.
Fig. 2 shows the drying curves, tted according to the
Fick diusion model, for all test cycles, the moisture
terms being dimensionless, as a function of the drying
time. Modelling of the drying curve was accomplished
by taking the rst ten terms of the series found using
the model.
Table 2 and Fig. 2 show that the temperature aects
the drying kinetics, the drying time decreasing with
increasing temperature. However, the set point tempera-
ture showed a greater inuence on the drying process
than the hot air temperature, since the drying time only
changed when the set point temperature was raised. This
can be conrmed by comparing tests MW3 and MW4.
For tests MW5, MW6 and MW7, the corresponding
drying times were equal, these tests representing the cen-
tral point of the experimental design. This was to be
expected, since the conditions of the variables used were
repeated.
Fig. 3 shows the drying rates as a function of the
macadamia nut moisture contents during the combined
hot air and microwave drying processes, for several
combinations of drying air and set point temperature.
It was observed that the drying rates were higher at
the beginning of the drying operation, when the product
moisture contents were higher.

Table 2
Results of the drying tests for the macadamia nuts in the microwave oven
Test n RHair (%) Tair (C) Tsp (C) Tproduct (C) M0 nut (%) Mf nut (%) Mf kernel (%) ttest (min)
MW1 58.70 58 0.5 64 5661 10.19 3.79 1.51 330
MW2 59.60 62 0.5 64 5761 9.80 3.72 1.50 330
MW3 57.00 58 0.5 68 6065 9.98 3.70 1.49 270
MW4 58.50 62 0.5 68 6065.5 10.19 3.75 1.52 270
MW5 58.7 60 0.5 66 5863 9.80 3.73 1.52 300
MW6 56.9 60 0.5 66 5963 9.98 3.74 1.54 300
MW7 60.30 60 0.5 66 5863 10.19 3.78 1.55 300

1.0 Tair = 58 C; Tsp = 64 C (MW1)

diffusional model
Tair = 62 C; Tsp = 64 C (MW2)
diffusional model
0.8 Tair = 58 C; Tsp = 68 C (MW3)
MR = (Mt - Me / Mo - Me)

diffusional model
Tair = 62 C; Tsp = 68 C (MW4)
diffusional model
Tair = 60 C; Tsp = 66 C (MW5)
0.6 diffusional model
Tair = 60 C; Tsp = 66 C (MW6)
diffusional model
Tair = 60 C; Tsp = 66 C (MW7)
0.4 diffusional model

0.2

0.0
0 60 120 180 240 300 360
tim e (min)

Fig. 2. Drying curves of macadamia nuts tted according to the Fick diusion model, obtained by combined hot air and microwave drying.
554 F.A. Silva et al. / Journal of Food Engineering 77 (2006) 550558

0.00100
Tair = 58 C; Tsp = 64 C (MW1)
0.00090
Tair = 62 C; Tsp = 64 C (MW2)
Region I
Tair = 58 C; Tsp = 68 C (MW3)

dM/dt (kg water / kg dry matter * min)

0.00080
Tair = 62 C; Tsp = 68 C (MW4)
0.00070
Tair = 60 C; Tsp = 66 C (MW5)
0.00060 Tair = 60 C; Tsp = 66 C (MW6)
Tair = 60 C; Tsp = 66 C (MW7)
0.00050

0.00040
Region II
0.00030

0.00020

0.00010

0.00000
0.00 0.02 0.04 0.06 0.08 0.10 0.12
M (kg water / kg dry matter)

Fig. 3. Drying rate as a function of the macadamia nut moisture content during the combined hot air and microwave drying processes, for various
pairs of air and product temperature values.

During the initial drying period, a rapid decrease of in-shell nuts. The best t was achieved for n = 10 using
drying rate occurred in all tests (Region I). Following the criterion of taking the correlation coecient closest
this initial period of decline and after reaching a specic to 1.
moisture content, the drying rate continued declining, From Table 3 it can be seen that the Fick model
but more slowly and gradually and almost linearly until used for diusion adjusted very well to the experimental
the equilibrium moisture condition was reached (Region data, with very acceptable values for the mean standard
II). In fact, the dierences in drying rate for the dierent deviation (P). Increasing both temperatures employed in
tests were undistinguishable in Region II. It was also the drying tests favoured the mass transfer processes and
observed that in the initial drying period the air and thus the eective diusivity (Def). The values for Def
set point temperatures had a strong inuence on the dry- fully agreed with the drying rate data, because Def is rep-
ing rate, indicating that the higher these temperatures, resentative of the speed with which the water is removed
the higher the drying rate. This was especially true for internally from the material.
tests MW3 and MW4.
The values for eective diusivity (Def), correlation 3.2. Peroxide value
coecient (R2) and mean standard deviation (P) for
each treatment are given in Table 3. The software The linear regression analyses of these values showed
Statistica version 5.0 (Microsoft, 1995) was used to that a signicant (p 6 0.05) linear correlation existed for
determine the eective diusivity. The nal product both types of drying experiment, because the PV values
shape was considered to be a sphere, and a method for changed with the storage time for all tests (Table 4).
non-linear adjustment was used, by taking the rst n From the data shown in Table 4, it was observed that
terms of the exact solution for the Eq. (1) of Ficks these values increased during the storage period, except
model for such geometry. An average diameter of for tests MW2, MW4 and MW7 where the values were
2.39 cm was obtained by measuring 100 macadamia stable for the last three months. The sample dried

Table 3
Values for eective diusivity (Def), correlation coecient (R2) and mean standard deviation (P) for the diusion model
Test n Tair (C) Tsp (C) k (1/min) Def 108 (m2/min) R2 P (%)
MW1 58 0.5 64 0.00252 3.65 0.979 6.39
MW2 62 0.5 64 0.00249 3.61 0.983 5.32
MW3 58 0.5 68 0.00348 5.03 0.991 3.61
MW4 62 0.5 68 0.00325 4.70 0.988 4.48
MW5 60 0.5 66 0.00297 4.30 0.987 3.73
MW6 60 0.5 66 0.00300 4.34 0.986 4.31
MW7 60 0.5 66 0.00275 4.03 0.983 5.57
 F.A. Silva et al. / Journal of Food Engineering 77 (2006) 550558 555

Table 4
Peroxide values in the lipid fraction of the macadamia kernels dried by microwave and hot air
PVperoxide value (meq O2 kg1 oil)
Storage time (days) MW1 MW2 MW3 MW4 MW5 MW6 MW7 CVL C*
c b b b c c b c
0 0.80 0.77 0.84 0.80 1.10 1.13 1.06 1.52 0.54a
90 1.73b 2.14a 1.68ab 2.17a 2.12b 1.90b 2.30a 2.83b 0.55a
180 2.57a 2.32a 2.55a 2.56a 2.49a 2.47a 2.45a 3.29a 0.54a
Samples in the same line with the same letter are not signicantly dierent (5%/Tukey).
*
Ccontrol sample, since this was the fresh raw material, it was kept at minus 18 C during the storage time.

conventionally (CVL), presented the highest peroxide
value at the end of the storing period, suggesting that
the combined hot air and microwave drying process con-
tributed less to the oxidative reactions of the kernels. In
spite of the higher values however, it was noticed that
the peroxide values did not reach the tolerated maximum
limit of 10 meq O2 kg1, as stipulated by the Brazilian
legislation for raw oils (ANVISA, 1999). An increase in
the peroxide value is an indication that oxidative reac-
tions have already started in the stored samples. The var-
iation in peroxide values that occurred during storage
may have been related to the dierent conditions used
for drying, since the control sample (C) maintained prac-
tically the same value throughout the storage period.
From the Tukey test, the rst signicant dierence in
the peroxide value occurred on the 90th day of storage
for all the tests, whereas for the control sample (C), the
peroxide value maintained its initial value the whole time.
3.3. Free fatty acids (% FFA)
On applying a linear regression analysis to the acidity
values as a function of storage time, for the lipid fraction
of the macadamia nut kernels dried by the combined hot
air and microwave drying processes (Table 5), no signif-
icant linear correlation (p 6 0.05) was observed for any
of the tests.
The acidity shown by the lipid fraction of the macad-
amia nut kernels (Table 5) was quite low and practically
unchanged during the storage period, which leads to the
conclusion that the product did not show any hydrolytic
reaction. Inactivation of the enzyme usually associated
with hydrolytic reactions during the drying of macad-
amia nuts may have been responsible for this behaviour.
From the Tukey test, no signicant dierence was
observed in the values for % FFA during storage for
any of the tests, including those of the conventionally
dried sample (CVL) and the control sample (C). It can
be observed that the values for % FFA were well below
the maximum tolerated limit of 1.5% FFA for raw oils
as stipulated by the Brazilian legislation (ANVISA,
1999).
3.4. Sensory acceptance
Table 6 presents the average results for the sensory
acceptance tests of the macadamia kernel samples dried
by microwaves and hot air, as compared to the sample
obtained industrially by the conventional hot air process
(Test no. CVL). No signicant dierences were found
(p 6 0.05) amongst the samples for any of the attributes
evaluated (appearance, crispness and overall accep-
tance) at zero storage time. The same result was
obtained for the sample evaluations carried out after
90 days storage. After 180 days storage tests MW2
and MW5 showed signicant dierences for crispness
(p 6 0.05). Concerning the other attributes, the samples
were similar (p 6 0.05). At zero storage time, the aver-
age values for the attribute appearance ranged from
6.17 to 6.83, for the attribute crispness from 7.37 to
7.93, and for the attribute overall acceptance from
7.03 to 7.33. After 90 days storage, the range of values
for appearance had changed to from 6.63 to 7.30, for
crispness from 7.10 to 7.80 and for overall acceptance
from 6.77 to 7.47, respectively. After 180 days storage,
the range of values for appearance had changed to from
6.03 to 7.03, for crispness from 6.77 to 7.87 and for over-
all acceptance from 6.60 to 7.23, respectively.
At zero storage time there was a minimum of 53.3%
acceptance for appearance, 76.7% for crispness and
66.7% for overall acceptance, with scores above level 7
(liked moderately). After 90 days storage, there was a

Table 5
Free fatty acids of the lipid fraction of the macadamia kernels dried by microwave and hot air
FFAfree fatty acids (%)
Storage time (days) MW1 MW2 MW3 MW4 MW5 MW6 MW7 CVL C*
a a a a a a a a
0 0.095 0.097 0.096 0.096 0.095 0.094 0.094 0.123 0.055a
90 0.110a 0.112a 0.109a 0.109a 0.094a 0.111a 0.110a 0.136a 0.054a
180 0.113a 0.122a 0.124a 0.109a 0.129a 0.123a 0.113a 0.148a 0.055a
Samples in the same line with the same letter are not signicantly dierent (5%/Tukey).
*
Ccontrol sample, since this was the fresh raw material, it was kept at minus 18 C during the storage time.
556 F.A. Silva et al. / Journal of Food Engineering 77 (2006) 550558

Table 6
Scores (09) for the sensory acceptance tests of macadamia kernels dried by microwave and hot air
Test no. Appearance Crispness Overall acceptance
0 day 90 day 180 day 0 day 90 day 180 day 0 day 90 day 180 day
CVL 6.67a 7.30a 7.03a 7.93a 7.30a 7.73ab 7.27a 7.27a 7.00a
MW1 6.67a 6.63a 6.13a 7.40a 7.10a 7.37ab 7.23a 7.00a 7.10a
MW2 6.17a 6.70a 6.57a 7.37a 7.57a 7.87a 7.27a 6.77a 7.23a
MW3 6.43a 6.83a 6.70a 7.93a 7.33a 6.90ab 7.13a 7.23a 6.90a
MW4 6.83a 6.43a 6.33a 7.50a 7.13a 7.37ab 7.33a 7.27a 7.07a
MW5 6.60a 6.77a 6.03a 7.53a 7.80a 6.77b 7.03a 7.43a 6.60a
MW6 6.73a 7.10a 6.83a 7.53a 7.60a 7.30ab 7.13a 7.47a 7.00a
MW7 6.43a 7.07a 6.33a 7.67a 7.37a 7.47ab 7.17a 7.13a 7.13a

minimum of 63.3% acceptance for appearance, 66.7%
for crispness and 63.3% for overall acceptance, with
scores above level 7 (liked moderately). After 180 days
storage, there was a minimum of 46.7% acceptance for
appearance, 60.0% for crispness and 63.3% for overall
acceptance, with scores above level 7 (liked moderately).
Thus, according to the scale used, the panellists evalu-
ated the samples from liked moderately to liked very
much for all attributes on the three occasions, with no
drop occurring in the sensory acceptance scores up to
180 days of storage.
3.5. Experimental design analysis for the macadamia
kernels dried by microwaves and hot air
The behaviour of the macadamia kernels submitted
to the microwave and hot air drying was analysed by
means of a statistical treatment of the data using the
software Statistica version 5.0 (Microsoft, 1995). The
operational conditions and respective responses with
respect to drying time (min.), peroxide value (meq
O2 kg1 oil) and free fatty acids (% FFA) are shown
in Table 7. Values for the peroxide value and % free
fatty acids were obtained for the experimental design
at the end of the storage period.
Table 8 shows the eects of the independent variables
(air temperature and set point temperature) as well as
their interactions with the dependent variables (drying
time, peroxide values and free fatty acids) for the mac-
adamia kernels dried by microwaves and hot air. The
interpretation of the results from Table 8 should be
based on the magnitude and signs of the regression coef-
cients. The higher the value of the coecient, the
greater the inuence on the response studied. The sign
indicates the direction of the proportionality coecient
in relation to the response, that is, a plus sign implies
that the relation between the independent variable and
the response is directly proportional, denoting that a
crossing from a lower to a higher level of the indepen-
dent variable results in an increased response value.
For a minus sign, the relation is inversely proportional.
The value of p is concerned with the level of signi-
cance of the independent variable with respect to the
response studied. The range of condence usually cho-
sen is 95%. Thus it can be said that for values of p
smaller than 5%, the variable can be considered as statis-
tically signicant, and to the contrary, not signicant.
Looking at the eects of the variable drying time pre-
sented in Table 8, it can be observed that the variables
air temperature and set point temperature were not sta-
tistically signicant, although the latter was the most
important, because, according to Table 7, the drying
time was much shorter when the temperature was
68 C, independent of the air temperature, that is, the
drying time was the same for temperatures of 58 and
68 C. The value for the standard deviation was zero,
as there were no dierent responses for the replicates
at the central point, therefore the values for t and p
could not be calculated. The value for the correlation
coecient was 0.992.

Table 7
22 Factorial design with triplicates at the central point for the macadamia kernels dried by microwaves and hot air
Test no. Independent variables Dependent variables
Tair (C) Tsp (C) ttest (min) PV (meq O2 kg1 oil) FFA (%)
MW1 58 (1) 64 (1) 330 2.57 0.113
MW2 62 (+1) 64 (1) 330 2.32 0.122
MW3 58 (1) 68 (+1) 270 2.55 0.124
MW4 62 (+1) 68 (+1) 270 2.56 0.109
MW5 60 (0) 66 (0) 300 2.49 0.129
MW6 60 (0) 66 (0) 300 2.47 0.123
MW7 60 (0) 66 (0) 300 2.45 0.113
 F.A. Silva et al. / Journal of Food Engineering 77 (2006) 550558 557

Table 8
Analyses of the eects on the dependent variables for the macadamia kernels dried by microwaves and hot air
Dependent variables Factor Regression coecients Standard deviation t (2) p
Time Average 300.00 0.00
(1) Tair(L) 0.00 0.00
(2) Tsp(L) 60.00 0.00
1L 2L 0.00 0.00
Peroxide value Average* 2.49 0.01 319.86 0.00
(1) Tair(L)* 0.12 0.02 5.91 0.03
(2) Tsp(L)* 0.11 0.02 5.21 0.04
1L 2L* 0.13 0.02 6.45 0.02
% Free fatty acids Average* 0.12 0.00 38.59 0.00
(1) Tair(L) 0.00 0.01 0.37 0.74
(2) Tsp(L) 0.00 0.01 0.21 0.85
1L 2L 0.01 0.01 1.45 0.28
*
p < 0.05 (statistically signicant to 95% condence).

With respect to the peroxide values, it can be
observed from Table 8 that the two variables, air tem-
perature and set point temperature, were statistically sig-
nicant, showing a positive value. Table 7 shows that
the peroxide values were much lower than the value of
10 meq O2 kg1 stipulated by the Brazilian legislation
(ANVISA, 1999). The value for the standard deviation
was small, which denotes that the replicates at the cen-
tral point were satisfactory. The correlation coecient
for the dependent variable peroxide value was 0.977.
As for acidity, it can be seen in Table 8 that the vari-
ables air temperature and set point temperature were
not statistically signicant, although the values obtained
were much smaller as compared to the value of 1.5%
FFA stipulated by the Brazilian legislation (ANVISA,
1999). The correlation coecient for the dependent var-
iable acidity was 0.687. The low correlation values
revealed a particular behaviour of the product, dierent
with respect to the linear model used in the statistical
treatment (Barros Neto et al., 1996). Thus it was con-
rmed that the only purpose of the design developed
here was to know and to study the inuence of some
of the variables on the process and the behaviour of
the kernel under the treatment applied, rather than to
obtain a mathematical model to describe such behav-
iour. This was not possible by way of the linear
treatment.
amia nuts as compared to the quality obtained using
the conventional drying process.
The measurements obtained from the analyses of
peroxide values and percent free fatty acids were well
below the values permitted by the Brazilian legislation,
indicating that even using temperatures a little bit higher
than those used in the conventional drying process, the
product did not lose its quality with respect to oxidative
reactions occurring during the six month storage period.
The results obtained from the sensory acceptance
tests lead to the conclusion that the macadamia kernels
dried by microwaves and hot air did not dier signi-
cantly (p 6 0.05) from the conventionally dried macad-
amia kernels with respect to the attributes evaluated
(appearance, crispness and overall acceptance). Of the
macadamia kernels dried by microwaves and hot air,
only test MW5 showed a slight drop in acceptance.
For all remaining microwave tests the macadamia ker-
nels were very well accepted for all attributes, in general
not presenting a signicant drop in acceptance during
the storage period studied.
It was possible to model the drying kinetics of the
macadamia nuts for all microwave experiments accord-
ing to Cranks model for products of spherical geometry.
Acknowledgement

The authors are indebted to the Coordenacao de

4. Conclusions
It was concluded that it was perfectly possible to dry
macadamia nuts by applying microwave energy, the dry-
ing times (4.55.5 h) being much shorter than those
required in the conventional hot air drying process
(about 144 h). Also, the proposed method of applying
microwaves during the drying process was shown to be
ecient in preserving the natural quality of the macad-
Aperfeicoamento de Pessoal de Nvel SuperiorCapes,
Braslia, DF, Brasil, for the scholarship associated with
this work.
References
ANVISAAgencia Nacional de Vigilancia Sanitaria. Brasil, Resolu-
cao no. 482, de 23 de setembro de 1999. BrasliaDF, no. 3029,
republicada em 20/06/2000.
558 F.A. Silva et al. / Journal of Food Engineering 77 (2006) 550558
AOACOcial Methods of Analysis of Association of Ocial
Analytical Chemists (1997). K. Helrich (Ed.), Arlington.
AOCS (American Oil Chemists Society) (2002). Ocial methods and
recommended practices of the American Oil Chemists Society (5th
ed.). Champaign: AOCS.
Berteli, M. N., & Marsaioli, A. Jr., (2005). Evaluation of short cut
pasta air dehydration assisted by microwaves as compared to the
conventional drying process. Journal of Food Engineering, 68,
175183.
Barros Neto, B., Scarmino, I. S., & Bruns, R. E. (1996). Planejamento
e otimizacao de experimentos. Campinas: Unicamp.
Cavaletto, C. G. (1983). Macadamia nut. In H. Chan, Jr. (Ed.),
Handbook of tropical foods (pp. 361397). New York: Marcel
Dekker, Inc..
Chung, M. T. C., & Furutani, H. (1989). Microwave drying of
macadamia nuts. Transactions of the ASAE, 5(4), 565567.
Crank, J. (1975). The mathematics of diusion (2nd ed.). Oxford:
Claredon Press (414p).
Duke, J. A. (1989). Handbook of nuts. New York: CRC Press, Inc.,
p. 343.
Kaijser, A., Dutta, P., & Savage, G. (2000). Oxidative stability and
lipid composition of macadamia nuts grown in New Zealand. Food
Chemistry, 71, 6770.
Lomauro, C. J., Bakshi, A. S., & Labusa, T. P. (1985). Evaluation of
food moisture sorption isotherm equations. Part I: fruit, vegetable
and meat products. Lebensmittel-Wissenschaft und Technologie, 18,
112122.
Mason, R. L., & Mc Conachie, I. (1994). A hard met to crak. A review
of the Australian macadamia nut industry. Food Australia, 46(10),
466471.
Mason, R. L., & Wills, R. B. H. (2000). Macadamia nut quality
research: The processing challenge. Food Australia, 52(9), 416419.
Microsoft (1995). Statistica v.5.0 (Copyright19842000 by StatSoft,
Inc.).
OMahony, M. (1986). Sensory evaluation of foodsstatistical methods
and procedures. New York: Marcel Dekker Inc., 478 p.
Silva, F. A., & Marsaioli, Jr., A. (2003). Drying Brazil nuts using hot
air assisted by microwaves, compared to conventionally dried
samples. In: J. Binner, (Ed.), Proceedings of the 9th international
conference on microwave and high frequency heating, Loughbor-
ough, UK, pp. 341344.
Sousa, W. A., Pitombo, R. N. M., da Silva, M. A. A. P., & Marsaioli,
A. Jr., (2001). Sensory evaluation of dried bananas obtained from
air dehydration assisted by microwaves. In M. Willert-Porada
(Ed.), Advances in microwave & radio frequency processingreport
from the 8th international conference on microwave and high
frequency heating (pp. 289302). Berlin: Springer-Verlag.
Toledo Piza, P. L. B. (2000). Segunda etapa de secagem da noz
macadamia. Botucatu, 93p. Tese (Mestrado em Agronomia/ Energia
na Agricultura)Faculdade de Ciencias Agronomicas, UNESP.
USDAUnited States Department of Agriculture (2002). Situation
and outlook for macadamia nuts. Available from www.fas.
usda.gov/htp/Hort_Circular/2002/02-04/Stats/MAC.pdf. Accessed
29.01.2003.
Xavier, J. A., Toledo Piza, J. A. N., Toledo Piza, P. L. B., & Valarelli,
I. D. (1996). Estudo comparativo entre duas metodologias para
determinacao do teor de umidade de equilbrio da noz macadamia
em casca. In Congresso Brasileiro de Engenharia Agrcola, 25,
Bauru. Anais do 25 Congresso Brasileiro de Engenharia Agrcola,
Bauru, Sao Paulo, (CD-ROM).