Food Hydrocolloids 25 (2011) 1319e1326

Contents lists available at ScienceDirect

Food Hydrocolloids
journal homepage: www.elsevier.com/locate/foodhyd

Simplex lattice mixture design approach on the rheological behavior

of glucomannan based salep-honey drink mixtures: An optimization
study based on the sensory properties
Safa Karaman a, Mustafa Tahsin Yilmaz b, Ahmed Kayacier a, *
a
Department of Food Engineering, Erciyes University, 38039 Kayseri, Turkey
b
Safiye Cikrikcioglu Vocational College, Erciyes University, 38039 Kayseri, Turkey

a r t i c l e i n f o a b s t r a c t

Article history: The rheological properties of salep drink sweetened with different honeys were measured using
Received 31 March 2010 a controlled-stress rheometer. Mixture design experiments were used to study the effect of interactions
Accepted 8 December 2010 among pine, flower and highland honeys on the rheological properties of salep-honey drink mixture
(SHDM) samples. In addition, product optimization was carried out using the ridge analysis to determine
Keywords: the optimum mixture proportions based on sensory properties of SHDM samples. Flower honey was the
Salep
component showing the highest effect on the consistency coefficient values of SHDM samples. The
Rheology
preference of panelists was more prominent for the SHDM samples including the higher concentrations
Sensory properties
Mixture design
of highland honey with respect to odour and overall preference parameters. Optimum values of pine,
Optimization flower and highland honeys in the mixture were found to be 0e85%, 0e40% and 15e100%, respectively,
with respect to sensory properties. In addition, ridge analysis results revealed that the SHDM should
include 65% highland honey, 35% pine honey and no flower honey to obtain the maximum overall
preference score (7.24). The consistency coefficient and flow behavior index values of the sample to get
maximum overall preference score (7.24) were predicted to be 3.650 Pa sn and 0.435, respectively.
Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction salep powder and generally served with aromatic spices such as
cinnamon (Telcioglu & Kayacier, 2007). In commercial production
Salep is a hydrocolloid obtained from milling of the dried tubers of as well as in home preparation, table sugar is used to sweeten the
certain wild orchids (Kaya & Tekin, 2001). It is a powder used as an salep drink. There are some speculative reports about the negative
ingredient in various food formulations and has been used in medi- effects of table sugar such as hypoglycaemia and high calorie
cines, drinks and ice cream production (Lange, 1998; Razavi & intake. Diabetes is a worldwide health problem with approximately
Karazhiyan, 2009). Due to its nutritive and demulcent properties, 200 million sufferers. People with diabetes should follow a special
salep powder has been used for many years (Ktistis & Georgakopoulos, diet for their whole life. In this respect, they should consume low
1991; Razavi & Karazhiyan, 2009). It contains glucomannans in high sugar or sugar-free products for reduction of sugar level in blood
levels (16e55%) and these are natural neutral water soluble fibers and weight control (Iop, Silva, & Beleia, 1999).
which help to normalization of blood sugar, relieve stress and prevent Honey which is produced by bees from nectar of flowers and
blood sugar abnormalities like hypoglycaemia (Farhoosh & Riazi, honeydews, is a natural complex food with high nutritional value
2007; Hozumi et al., 1995; Kaya & Tekin, 2001). It has also been (Anklam, 1998; Azeredo, Azeredo, De Souza, & Dutra, 2003;
reported that glucomannans reduce the risk of chronic diseases Lazaridou, Biliaderıs, Bacandritsos, & Sabatini, 2004). It is a concen-
(Vuksan et al., 1999). trated sugar solution and contains about 200 substances having
Salep powder is the main ingredient of a traditional dairy important favorable health effects hence it has been used as
beverage known as salep drink which is commonly consumed a traditional medicine (White, 1979). The major sugars in honey are
during winter in Turkey. The drink is prepared with milk, sugar and monosaccharides, namely glucose and fructose, it also contains
maltose, sucrose and other sugars in small quantities (Kayacier &
Karaman, 2008; Bhandari, D’arcy, & Chow, 1999). It has been
* Corresponding author. Tel.: þ90 352 4374937; fax: þ90 352 4375784. shown that fructose is absorbed more slowly from the gastrointes-
E-mail address: [email protected] (A. Kayacier). tinal tract than glucose. For that reason, fructose is an insulin

0268-005X/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodhyd.2010.12.007
1320 S. Karaman et al. / Food Hydrocolloids 25 (2011) 1319e1326

independent monosaccharide found in honey and it is rapidly taken received the higher sensory score in comparison with other tested
up by the liver when honey is consumed. Thus, blood sugar level concentrations in the preliminary sensory test.
rises very slowly after consuming fructose rich foods such as honey.
Furthermore, honey could be utilized as an appropriate sweetener 2.4. Rheological measurements
alternative for the type II diabetic diet (Bahrami et al., 2009; Crapo,
Kolterman, & Olefsky, 1980). Rheological properties of SHDM samples were determined
Salep drink is commonly consumed in the Mediterranean area. Its using a strain/stress controlled rheometer (Thermo-Haake, Rheo-
rheological and sensory characteristics should be studied since the stress 1, Germany) equipped with a temperature-control unit
quality of such products is based on a variety of properties. Rheo- (Thermo-Haake, Karlsruhe K15 Germany). The rheometer had
logical and sensory properties are of primary importance to food a cone-plate configuration with a cone radius of 35 mm and a gap of
technology engineering applications such as filtration, extraction, 0.140 mm between the cone and plate. Measurements were carried
transport and process calculations such as pump sizing, pump type out in the shear rate range of 1e100 s1 at constant temperature
and power. Also, rheological characteristics of foods affect the (25  C). 0.85 ml salep drink sample was placed between cone and
sensory properties (Marcotte, Hoshahili, & Ramaswamy, 2001). plate and the measurement was started immediately. Total 25 data
Mixture design methodology can be utilized for the investiga- points were recorded at 10 s intervals during the shearing. Each
tion of the functions of the ingredients in processed foods and it measurement was replicated seven times on the same sample with
approves the importance of ingredient interactions. (Arteaga, Li- two repetitions. The apparent viscosity was determined as a func-
Chan, Nakai, Cofrades, & Jiménez-Colmenero, 1993). Using this tion of shear rate. Obtained data were fitted to Power Law model
methodology, the present study was undertaken to investigate the using Rheowin data Manager Pro and consistency coefficient and
effect of interaction between pine, flower and highland honeys on flow behavior index values were calculated according to the
the rheological properties of salep drink. An additional aim was to following model.
find the optimum levels of these honeys to obtain the most
acceptable product with respect to the sensory properties studied. ha ¼ K g_ n1 (1)

2. Material and methods where ha is the apparent viscosity (Pa s), K is the consistency
coefficient (Pa sn), g_ is the shear rate (s1) and n is flow behavior
2.1. Materials index (dimensionless).

All honey samples (flower, pine and highland honeys) were 2.5. Sensory analysis
provided by Ege Bal Co. (Izmir, _ Turkey). Salep powder was
purchased from Özselamog lu Food Co. (Kayseri, Turkey) and UHT The SHDM samples were prepared one day before and stored in
milk was obtained from a local market (Dost Milk, Sakarya, Turkey). refrigerator (4  C). For sensory evaluation, SHDM samples (50 ml)
Commercial grade honeys were preconditioned before the analysis were presented at 75  C to panelists at certain intervals in odour-
since air bubbles and crystals present in the honey may influence less, randomly coded glass beakers of 100 ml capacity covered with
the viscosity. For this purpose, the honey samples were held at glass dishes. Sensory evaluation was performed in a room with
55  C in a water bath for 1 h, and afterwards placed overnight in an appropriate temperature (25  C) in open sitting. The panelists had
oven (Nuve, Ankara Turkey) adjusted to 30  C. access to tap water to cleanse their palates before proceeding to the
next sample. Sensory analyses of the SHDM mixtures were carried
2.2. Physicochemical analysis out by eight selected staff and graduate students of Food Engi-
neering department at Erciyes University. Before the analyses,
Ash contents, pH and soluble dry matter (brix) values of the sensory panel were trained about the sensory evaluation tech-
honey samples were determined as outlined (AOAC, 1995). To niques. The sensory attributes evaluated were taste, odour, mouth
determine ash content, the honey samples were dried at 135  C for coating, consistency and overall preference. All sensory attributes
2 h and incinerated using a muffle oven (Protherm PLF 12015, were evaluated using a sensory analysis form with scale of 1e9
Ankara, Turkey) at 550  C until carbonization of the samples. The points where 1 reflected a very low and 9 a very high score.
pH values were determined using a pH meter (WTW-Inolab Level 3 Panelists evaluated all (15) samples in three sessions (five at each
Terminal, Weilheim, Germany) in a solution of 10% (w/v) honey in
distilled water at 25  C. An automatic refractometer (Reichert AR
Table 1
700, USA) was used for the determination of brix values of the
Compositions of salep drink mixture samples in a simplex lattice mixture design.
samples at 20  C.
Mixtures Ingredient proportions Pine Flower Highland
honey (%) honey (%) honey (%)
2.3. Preparation of salep drink and mixture samples X1 X2 X3
1 1 0 0 100 0 0
To prepare salep drink, 100 ml of milk was heated up to 65  C on 2 0 1 0 0 100 0
a hot plate (Yellowline, Germany) and 0.75 g salep powder (Telcioglu 3 0 0 1 0 0 100
4 0.5 0.5 0 50 50 0
& Kayacier, 2007) was added into milk slowly to prevent agglom- 5 0 0.5 0.5 0 50 50
eration and stirred constantly with a magnetic stirrer up to 75  C for 6 0.5 0 0.5 50 0 50
5 min. Then, salep drink was cooled down to room temperature and 7 0.25 0.75 0 25 75 0
kept at this temperature for 5 min. To prepare salep-honey drink 8 0 0.25 0.75 0 25 75
9 0.75 0 0.25 75 0 25
mixture (SHDM) samples, 10 g of a mixture of the experimental
10 0.75 0.25 0 75 25 0
honeys; pine, flower and highland at different proportions corre- 11 0 0.75 0.25 0 75 25
sponding to each treatment (15 mixtures in Table 1) was added into 12 0.25 0 0.75 25 0 75
90 g of the salep drink and stirred for 1 min to obtain experimental 13 0.5 0.25 0.25 50 25 25
mixture samples. Samples with 10% honey concentration were 14 0.25 0.5 0.25 25 50 25
15 0.25 0.25 0.50 25 25 50
chosen for the study because this concentration in SHDM samples
 S. Karaman et al. / Food Hydrocolloids 25 (2011) 1319e1326 1321

session) consecutively in three days. The manner in which the estimated ridges of maximum and minimum response for
treatment combinations were divided between the sessions and increasing radii from the centre of the original design. The contour
the order in which the samples were presented was randomized to plots generated from the predicted equations for each parameter
minimize the carryover effects (Muir & Hunter, 1991). were superimposed to obtain the optimum region. The correlations
between the parameters were determined using MINITAB for
2.6. Experimental design Windows Release 13Ò (Minitab, 2000). Duncan Multiple Range Test
(MstatC, 1986) was also used to determine the difference between
In the present study, the simplex lattice mixture design (SLMD) mean combination values.
was used to evaluate the effect of pine honey (X1), flower honey (X2)
and highland honey (X3) on the rheological, physical, chemical and 3. Results and discussion
sensory characteristics of the SHDM samples. Estimated ridge
analysis was used to determine optimum combinations to produce 3.1. Physico chemical and rheological properties of the honey
acceptable product, estimate the most acceptable product with the samples
highest sensory score and find the consistency coefficient (K) and
flow behavior index (n) values of such product with the highest The physicochemical and rheological properties of honey
sensory score. Component proportions were expressed as fractions samples used in the formulations of the SHDM mixtures are shown
of the mixture with a sum (X1þX2þX3) of one. These three factors; in Table 2. The ash contents of the honey samples were varied
pine honey, flower honey and highland honey (processing widely, ranging from 0.25% to 0.76%. The maximum ash content
components), levels and experimental design in terms of coded and was observed for the pine honey sample. The observed ash contents
uncoded as 15 combinations are presented in Table 1. The 15 points were similar to those of Indian honey samples (Ahmed, Prabhu,
were 3 single-ingredient treatments, 9 two-ingredient mixtures Raghavan, & Ngadi, 2007). On the other hand, all studied honey
and 3 three-ingredient mixtures (Fig. 1). samples were acidic in nature and the pH values were varied
between 4.76 and 5.21. The acidity of honey was due to presence of
organic acids and our results were consistent with those obtained
2.7. Statistical analysis and modelling of experimental data for Algerian and Indian honeys (Ahmed et al., 2007; Ouchemoukh,
Louaileche, & Schweitzer, 2006). Total soluble solid (brix) values of
The following polynomial equation of function xi was fitted for the samples varied between 75.3% and 79.8%, indicating that flower
each factor assessed at each experimental point. This polynomial honey had the highest solids content. These honey samples were
model differs from full polynomial models because it does not found to contain lower solids compared to some Indian honeys
contain a constant term (intercept equal to zero). The polynomial (Ahmed et al., 2007). The variations in total soluble solids were
model equation used was: reported to be dependent on climate, floral source and some other
factors (Ahmed et al., 2007). The Newtonian viscosity and consis-
Y ¼ b1 X1 þ b2 X2 þ b3 X3 þ b12 X1 X2 þ b13 X1 X3 þ b23 X2 X3 (2)
tency coefficient values of the honey samples were determined to
where Y is the estimated response; b1, b2, b3, b12, b13 and b23 are range from 8.0 to 15.2 Pa s and 8.7e16.6 Pa sn, revealing that the
constant coefficients for each linear and non-linear (interaction) pine honey was the processing component with the most viscous
term produced for the prediction models of processing compo- nature. This result may at first sight appear to be inconsistent with
nents. The analysis was performed using uncoded units. the expected trend reported in the literature where the honey
The computational work, including ternary contour graphical samples with higher brix value were reported to have higher
presentations of the models, was performed using Statistica, 1995 viscosity values. However, discrepancy in these results could be
(Statistica for Windows, computer program manual, Tulsa, OK). attributed to the fact that the viscosity of honeys depends also on
JMP statistical package software (Version 5.0.1.a, SAS Institute. Inc. its origin, thus, their chemical composition, year of production and
Cary, NC) was used to compute the predicted equations. JMP (2009) geographical region. Accordingly, Juszczak and Fortuna (2006)
statistical package software was also used to compute the reported that the viscosity of honey depends also on its origin
and its chemical constituents. This case was also well evidenced by
the study of Sopade et al. (2002) who determined the same results
for Australian honey, which was also different from those obtained
in the previous studies. On the other hand, the flow behavior index
values of the all honey samples were found close to 1, indicating the
Newtonian flow behavior of the honey samples.

3.2. Rheological properties of SHDM samples

For all honey samples tested, the Newtonian flow properties

were observed as stated above. However, they did not behave in
a similar fashion when they were mixed with the salep drink. Fig. 2
shows the flow rheograms of salep drinks sweetened separately
with the processing components; only pine honey, flower honey or
highland honey. The salep drinks sweetened with different honeys
showed a non-Newtonian flow characteristic since the flow
behavior index (n) was found to be less than unity. Shear stress and
shear rate data showed that the relationship of them was non-
linear. As seen in the figure, samples showed different shear stress
values as a function of shear rate. Differences in apparent viscosities
Fig. 1. 15 Points augmented simplex lattice mixture design for the effects of pine of samples at 50 s1, the shear rate in mouth (Bourne, 2002) were
honey (X1), flower honey (X2) and highland honey (X3). found to be significant (P < 0.05) (data not presented).
1322 S. Karaman et al. / Food Hydrocolloids 25 (2011) 1319e1326

Table 2
Physicochemical and rheological properties of the honey samples.

Ash (%) pH Brix (%) Power law parameters

h (Pa s)d K (Pa sn)d nd R2

Pine honey 0.76  0.08a 5.21  0.01a 75.30  0.15c 15.22  0.28a 16.573  0.97a 0.975  0.00a 0.9999
Flower honey 0.34  0.04b 4.76  0.01b 79.80  0.15a 8.09  0.25b 9.323  0.67b 0.974  0.01a 0.9999
Highland honey 0.25  0.01b 4.77  0.01b 76.10  0.20b 7.99  0.04b 8.672  0.40b 0.997  0.00b 1.0000

In a column, means with no common superscripts are significantly different (p < 0.05).
d
Newtonian viscosity (Power-law parameters obtained at 50 s1 & 25  C).

Similar flow behavior properties were also observed in the presented in Table 3. The differences between brix and power law
salep-honey drink mixture (SHDM) samples, exhibiting non-New- parameters were found to be significant (P < 0.05). Table 5 indicates
tonian flow characteristics. Power law model was used to explain the predicted models, significance of the regression coefficients and
the relationship between shear rate and apparent viscosity and all the R2 values obtained for 15 mixtures. These values were impor-
SHDM samples were well fitted to the model with high determi- tant in showing the significance of equations or models developed
nation coefficients. As seen in the Table 3, model coefficients were from the design. As seen in table, the linear terms in all predicted
0.999 for all samples. Flow behavior index values of samples were models were found to be significant (P < 0.05). R2 values of the
determined in the range of 0.373e0.450. Flow behavior values predicted models were higher than 0.74, indicating that they were
showed that rheological characteristics of all samples were shear relatively adequate for the prediction purpose (Henika, 1982). R2
thinning and the apparent viscosity of all samples decreased with values of the power law parameters were close to unity and for that
increasing shear rate. Consistency coefficients (K) of the samples reason, the experimental models for these parameters can predict
were calculated to range between 3.183 and 5.797 Pa sn and the purpose rather well.
differences between the SHDM samples were found to be signifi- Fig. 3 shows ternary contour plots indicating the effects of the
cant (P < 0.05) with respect to the mentioned power law param- interaction between the honey kinds on the brix values and flow
eters. Similar results were reported in the literature in which salep properties of SHDM samples. Fig. 3a indicates that the brix contour
drink fortified with soy milk at different concentrations was indi- values increased toward the flower honey vertex where maximum
cated to exhibit non-Newtonian shear thinning behavior and flow brix contour could be seen; however, the lowest brix values were at
behavior index ranging between 0.703 and 0.606 for control and pine honey-highland honey edge. These results were expected
salep drink samples containing 50% soy milk, respectively (Alpaslan because of the highest brix content of the honey as indicated in the
& Hayta, 2007). Telcioglu and Kayacier (2007) also reported that Table 2. From this point of view, it could be said that the use of
salep drink sweetened with table sugar in a whole milk showed flower honey in salep drink as a sweetener increased the total
shear thinning behavior and consistency coefficient was deter- soluble solid (brix value) of the final product, while use of the
mined to be 243 mPa sn at the shear rate of 1 s1, and decreased to others decreased.
36 mPa sn at 50 s1 and 27 mPa s at 100 s1. In the mixture 2 in which only the flower honey was present and
the mixtures 4 and 7 in which pine and flower honeys were
present, the K values were the highest and n values were the
3.3. Effect of the interaction between the honeys on the brix values lowest. However, the lowest K and the highest n values were
and flow properties observed in the mixture 8 containing 25% flower honey and 75%
highland honey and the mixture 15 containing 25% pine honey, 25%
Brix values and power-law parameters of 15 SHDM samples as flower honey and 75% highland honey (Tables 1 and 3). These
affected by the interaction between the honey varieties are results indicated that flower honey was the component having the
highest increasing effect on the consistency of the mixture samples,
30
Table 3
Mean values for brix and flow properties of the salep-honey drink mixture (SHDM)
25 samples.

Mixture Brix and flow properties

Shear stress (Pa)

20 Brix Ki ni R2
1 18.93  0.19ab 4.577  0.26cd 0.397  0.01e 0.9991
2 19.18  0.33a 5.340  0.64b 0.377  0.01f 0.9994
15 3 19.12  0.11a 3.869  0.51efg 0.424  0.01bc 0.9997
4 19.16  0.36a 5.419  0.16ab 0.375  0.02f 0.9999
salep with pine honey 5 18.66  0.28b 3.719  0.52fg 0.424  0.01bc 0.9996
10 6 18.61  0.21b 3.543  0.56gh 0.438  0.01b 0.9992
salep with flower honey 7 19.16  0.08a 5.797  0.23a 0.373  0.02f 0.9998
8 18.90  0.20 b 3.183  0.17h 0.450  0.01a 0.9999
salep with highland honey
5 9 18.62  0.25b 4.199  0.33de 0.420  0.01cd 0.9999
10 18.89  0.36ab 4.700  0.17c 0.409  0.01d 0.9995
11 18.86  0.10ab 4.273  0.17de 0.417  0.01cd 0.9997
 0.11ab  0.55efg  0.02bc
0 12
13
18.81
19.04  0.22ab
3.936
4.013  0.15ef
0.429
0.417  0.01 cd
0.9992
0.9994
0 20 40 60 80 100 14 19.04  0.43ab 3.927  0.16efg 0.423  0.01bc 0.9992
0.37ab 0.28gh 0.01b
Shear rate (s-1) 15 18.81  3.528  0.435  0.9993

In a column, means with no common superscripts are significantly different

Fig. 2. Rheograms obtained by controlled-stress rheometer for the salep drink (p < 0.01).
sweetened separately with pine honey, flower honey or highland honey, at 25  C. i
Power-law parameters obtained at 50 s1 & 25  C.
 S. Karaman et al. / Food Hydrocolloids 25 (2011) 1319e1326 1323

Fig. 3. Ternary contour plots of the effects of processing components on brix and power law parameters of the salep drink samples. (a) % brix, (b) consistency index (K, Pa sn) and (c)
flow behaviour index (n). (Power-law parameters obtained at 50 s1 & 25  C).

followed by pine honey. Generally, the linear terms had a signifi- high total soluble solid value (brix), as a matter of fact that the
cant effect on K and n; however, interactions between the pro- increasing effect of the total soluble solid on consistency was
cessing components had lower coefficients compared to the linear reported by Azoubel, Cipriani, El-Aouar, Anonio, and Murr (2005).
terms (Table 5). The contour plots indicated the highest K, but the They informed that when more solutes such as sugars are dissolved
lowest n values to be at pine honey-flower honey and flower in the solution, the viscosity increases due to an increase in
honey-highland honey edges and flower honey vertex. The lower hydrogen bonding with hydroxyl groups and distortion in the
values of K and higher values of n were observed at highland honey velocity pattern of the liquid by hydrated molecules of the solute.
vertex. Briefly, K values increased with increasing flower honey Accordingly, brix values indicated a significant positive correlation
concentrations (Fig. 3b and c). (R ¼ 0.619; P < 0.05) with K values and negative correlation
It was obvious from these results that flower honey had (R ¼ 0.654; P < 0.01) with n values. However, an inverse trend
a greater effect on the brix and flow parameters, K and n compared appears in the Table 2, indicating that the pine honey had higher K
with the other honey samples, pine honey and highland honeys. value (16.573 Pa sn) than did the flower honey (9.323 Pa sn)
The remarkable effect of flower honey could be attributed to its contrary to the fact that flower honey had the greatest effect to
increase K values of SHDM samples in the salep drink environment
Table 4 as mentioned above. On the other hand, Table 2 also indicates that
Mean values for sensory scores of the salep-honey drink mixture (SHDM) samples. the flower honey was the processing component with lower pH
Mixture Sensory scores value than pine honey. This lower pH value might have caused the
flower honey to have a higher effect on the K values; in other words,
Taste Odour Mouth Consistency Overall
coating preference
another factor contributing the higher effect of flower honey was
thought to be pH. These assumptions explaining the possible
1 6.5  0.96a 6.5  1.71abc 6.8  1.24a 6.5  1.63a 6.5  1.20ab
2 6.2  0.93a 6.1  0.68abc 6.6  1.09a 6.3  1.14a 6.2  0.93ab reason of the inverse effect of the honey samples seen on the K
3 6.8  1.18a 6.6  1.43abc 6.8  1.54a 6.8  1.26a 6.8  1.42ab values of the SHDM samples were consistent with the reports given
4 5.8  1.45a 5.8  1.32c 5.8  1.54a 6.2  1.71a 6.2  1.84ab in the literature. Ayar, Sert, and Akbulut (2009) studied the
5 7.0  0.86a 7.0  1.31ab 7.0  1.09a 6.8  0.89a 6.9  0.72ab _
viscosity of Incir Uyutması dessert prepared with salep, sugar and
6 6.9  2.03a 7.0  1.58ab 6.9  1.38a 7.2  1.46a 7.1  1.58ab
7 6.1  1.32a 6.0  1.21bc 6.2  1.14a 6.1  1.44a 6.2  1.53ab
fig. They found that fig amount decreased the pH values of the
8 7.2  0.73a 7.2  0.96a 7.1  0.83a 7.2  1.02a 7.1  0.89ab desserts and increasing sugar and fig amounts caused significant
9 6.8  1.06a 6.8  1.10abc 7.0  1.05a 7.0  0.88a 6.9  0.70ab increases in viscosity value, meaning that decreasing pH values
10 6.3  1.66a 6.4  1.42abc 6.5  2.09a 6.7  1.32a 6.5  1.31ab increased the viscosity. They also reported that salep, fig and sugar
11 6.2  0.58a 6.5  1.71abc 6.6  2.26a 6.3  1.74a 6.0  1.66b
concentrations had a synergistic effect on the viscosity values. It
12 7.1  1.04a 7.2  1.62a 6.7  1.43a 7.5  1.25a 7.5  1.08a
13 6.8  1.03a 6.8  1.66abc 6.5  1.41a 6.6  1.73a 6.8  1.26ab was also reported that some combinations of polysaccharides and
14 6.5  1.50a 5.9  1.47bc 6.6  1.14a 6.5  1.34a 6.6  1.64ab proteins were found to exhibit synergistic interactions when mixed
15 6.9  0.90a 6.5  1.34abc 6.8  0.96a 6.9  1.13a 6.8  1.41ab with each other (Baeza, Carp, Perez, & Pilosof, 2002). A great
In a column, means with no common superscripts are significantly different increase in viscosity may be anticipated due to strong interaction of
(p < 0.01). salep and fig with dessert components, probably with caseins
1324 S. Karaman et al. / Food Hydrocolloids 25 (2011) 1319e1326

Table 5
Predicted models for the experimental data from salep-honey drink mixture (SHDM) samples.

Parameter Predicted models R2

Brix Y ¼ 9.42X*1 þ 9.57X*2 þ 9.55X*3 þ 0.19X1X2  0.28X1X**
3  0.33X2X**3 0.7701
K Y ¼ 2.27X*1 þ 2.79X*2 þ 1.96X*3 þ 0.20X1X2  0.50X1X3  1.31X2X*3 0.9235
n Y ¼ 0.20X*1 þ 0.19X*2 þ 0.21X*3  0.00X1X2 þ 0.01X1X3 þ 0.04X2X3** 0.8964
Taste Y ¼ 3.25X*1 þ 3.02X*2 þ 3.46X*3  0.24X1X2 þ 0.29X1X3 þ 0.41X2X3 0.8531
Odour Y ¼ 3.31X*1 þ 3.03X*2 þ 3.42X*3  0.46X1X2 þ 0.25X1X3 þ 0.38X2X3 0.7385
Mouth coating Y ¼ 3.46X*1 þ 3.26X*2 þ 3.48X*3  0.66X1X*2  0.05X1X3 þ 0.24X2X3 0.8435
Consistency Y ¼ 3.30X*1 þ 3.08X*2 þ 3.53X*3  0.14X1X2 þ 0.41X1X3 þ 0.13X2X3 0.8201
Overall preference Y ¼ 3.25X*1 þ 3.02X*2 þ 3.52X*3 þ 0.01X1X2 þ 0.44X1X3 þ 0.15X2X3 0.8086

*p < 0.01.
**p < 0.05.
K, consistency coefficient; n, flow behaviour index.

(Camacho, Martinez-Navarrete, & Chiralt, 2005; Kök, 2007). In the the interaction of the honey samples (a source of saccharides) with
present study, there was a similar salep environment where the salep as a hydrocolloid. Accordingly, Nishinari, Zhang, and Ikeda
sucrose was replaced with the honey (as a sweetening agent) (2000) reported that hydrocolloids bind certain chemicals
whereby the pH fall was observed because of their acidic nature. (mostly polysaccharides and proteins) that are colloidally dispers-
Therefore, differences in the results that appear to be discrepant at ible in water. Thus, becoming effectively ‘‘soluble’’, they change the
the first sight could be ascribed to the pH values of the honeys rheology by raising the viscosity and inducing gelation. To
which were thought to change the magnitude order of their effects summarize, the combination effect of low pH and the highest brix
on apparent viscosity values in the salep drink environment. In this values of the flower honey was thought to lead the flower honey to
respect, the apparent viscosity value of salep drink prepared with have the highest effect on the K values.
pine honey was determined to be 0.39 Pa s at 50 s1 while that of
salep drink prepared with flower and that of highland honey were 3.4. Sensory properties of the SHDM samples
0.47 and 0.37 Pa s, respectively. These results revealed that the
flower honey was the honey type giving the highest apparent Higher odour and overall preference scores were given to the
viscosity values to the salep drink samples because of its the mixtures 8 and 12 in which the highland honey was present at
highest brix value (Table 2). In addition, the honeys did not behave higher concentration and the flower honey was either absent or
in a similar fashion when they were mixed with the salep drink; present at lower concentration. On the other hand, the mixtures 4
namely, their flow properties changed from Newtonian to non- and 11 in which the flower honey was present at higher concen-
Newtonian flow (Fig. 2) as mentioned before even though each trations received a lower sensory score (Tables 1 and 4). Taste,
honey sample exhibited the Newtonian flow behavior (Table 2), not mouth-coating and consistency scores were not remarkably affected
being in the salep drink environment. The possible reason could be from the processing components (Table 4). Sensory properties were

Fig. 4. Ternary contour plots of the effects of processing components on sensory properties of salep-honey drink mixture (SHDM) samples. (a) taste, (b) odour, (c) overall
preference.
 S. Karaman et al. / Food Hydrocolloids 25 (2011) 1319e1326 1325

honey vertex and inclined toward the pine honey and flower honey
vertexes (Fig. 4b). Overall preference score was higher at the points
closest to highland honey vertex and pine honey-highland honey
edge, indicating that the SHDM samples were more preferred by the
panelists when the pine honey was mixed with highland honey in
the salep drink (Fig. 4c), probably because of the synergistic effects
of the honeys upon the sensory parameter of the SHDM samples. On
the other hand, an increase in flower honey concentration reduced
the preference scores of the mixture samples which was seen at the
flower honey vertex in the figure (Fig. 4c). As a result, the addition of
highland honey as a sweetener generally increased the sensory
scores of samples.

3.5. Optimization of the processing components based on sensory

score

Figs. 5 and 6 show the superimposed binary and ternary contour

profile plots giving the optimum values of pine honey, flower honey
and highland honey in the mixture, respectively. As seen in the
contour profiles, the optimum amounts were 0e85%, 0e40% and
15e100% for pine honey, flower honey and highland honey,
respectively. In general, flower honey did not improve the sensory
properties of salep drink. For that reason, flower honey should be
used at maximum 40% concentration for a preferred salep drink.
Pine honey improved the sensory properties and increased the
preference of the samples together with highland honey. But it can
be added up to a maximum of 85%. Highland honey should be
added into the salep drink at least at 15% concentration. It could be
said that these optimum levels of the honeys could be used in the
formulation of dairy drinks to sweeten them with desired sensory
properties.
In addition, finding maximum levels of sensory parameters
should be the final goal to obtain an acceptable product in terms of
sensory properties. In this respect, it is necessary to find the levels
of the processing components that maximize and minimize the
sensory score values. Ridge analysis, which computes the estimated
ridge of optimum response for increasing radii from the centre of
the original design (Chwen-Jen, Koehler, & Akoh, 1996) indicated
that maximum taste score (7.05) could be obtained from the
mixtures containing 32% pine and 68% highland honeys. Maximum
mouth-coating (7.02) and odour score (6.99) could be obtained

Fig. 5. Superimposed binary contour profiles of the effects of mixture components

(pine honey, flower honey and highland honey) on sensory properties of salep-honey
drink mixture (SHDM) samples.

significantly (P < 0.05) affected from the linear terms of the pre-
dicted models as indicated by the adjusted regression models
developed for all the sensory evaluation parameters despite the fact
that the sensory properties were not significantly influenced from
two component interactions. Contour plots showed that taste and
odour parameters seemed to behave symmetrically when highland Fig. 6. Superimposed ternary contour profiles of the effects of mixture components
honey concentration was equal to those of pine and flower honeys (pine honey, flower honey and highland honey) on sensory properties of salep drink
(Fig. 4a and b). The odour score increased toward the highland samples.
1326 S. Karaman et al. / Food Hydrocolloids 25 (2011) 1319e1326

from the samples containing 28% flower and 72% highland honeys Camacho, M. M., Martinez-Navarrete, N., & Chiralt, A. (2005). Rheological charac-
terization of experimental dairy creams formulated with locust bean gum (LBG)
and 38% pine and 62% highland honeys, respectively. To obtain
and l-carrageenan combinations. International Dairy Journal, 15, 243e248.
maximum preference score (7.24), 35% pine and 65% highland Chwen-Jen, S., Koehler, P. E., & Akoh, C. C. (1996). Optimization of sucrose polymer
honeys should be used in the salep drinks. The consistency and flow synthesis using response surface methodology. Journal of Food Engineering, 61,
behavior index values of a sample to get maximum overall prefer- 97e100.
Crapo, P. A., Kolterman, O. G., & Olefsky, J. M. (1980). Effects of oral fructose in
ence score (7.24) were predicted to be 3.650 and 0.435 Pa sn, normal diabetic and impaired glucose tolerant subjects. Diabetic Care, 3,
respectively. 575e582.
Farhoosh, R., & Riazi, A. (2007). A compositional study on two current types of salep
in Iran and their rheological properties as a function of concentration and
temperature. Food Hydrocolloids, 21, 660e666.
4. Conclusion Henika, R. G. (1982). Use of response surface methodology in sensory evaluation.
Food Technology-Chic, 36, 96e101.
The results of present study showed that if salep drink is Hozumi, T., Yoshida, M., Ishida, Y., Mimoto, H., Sawa, J., Doi, K., et al. (1995). Long-
term effects of dietary fiber supplementation on serum glucose and lipoprotein
sweetened by some honeys as a replacement of sugar; their rheo-
levels in diabetic rats fed a high cholesterol diet. Endocrine Journal, 42, 187e192.
logical and sensory properties can be improved. Salep drink is Iop, S. C. F., Silva, R. S. F., & Beleia, A. P. (1999). Formulation and evaluation of dry
produced by using sugar, but there are numerous reports about dessert mix containing sweetener combination using mixture response meth-
odology. Food Chemistry, 66, 167e171.
adverse health effects of sugar like diabetes. In this respect, such
Jmp Sas. (2009). JMP statistics and graphics guide, version 8 (trial). Cary, North Car-
modifications would be of great importance for a healthy diet. olina: SAS Institute, Inc.
Flower honey could be used to increase the viscosity of the salep Juszczak, L., & Fortuna, T. (2006). Rheology of selected Polish honeys. Journal of Food
drink; however, the sample with flower honey did not receive Engineering, 75, 43e49.
Kaya, S., & Tekin, A. R. (2001). The effect of salep content on the rheological char-
a high sensory score. Therefore, it should be used as a mixture with acteristics of a typical ice-cream mix. Journal Food Engineering, 47, 59e62.
highland honey to improve sensory properties when optimum Kayacier, A., & Karaman, S. (2008). Rheological and some physicochemical charac-
combination levels of these honeys are taken into account. teristics of selected Turkish honeys. Journal of Texture Studies, 39, 17e27.
Kök, M. S. (2007). A comparative study on the compositions of crude and refined
locust bean gum: inrelation to rheological properties. Carbohydrate Polymers,
70, 68e76.
References Ktistis, G., & Georgakopoulos, P. P. (1991). Rheology of salep mucilages. Pharmazie,
46, 55e56.
Ahmed, J., Prabhu, S. T., Raghavan, G. S. V., & Ngadi, M. (2007). Physico-chemical, Lange, D. (1998). Europe’s medicinal and aromatic plants: Their use, trade and
rheological, calorimetric and dielectric behavior of selected Indian honey. conservation. Cambridge: Traffic Europe/International.
Journal of Food Engineering, 79, 1207e1213. Lazaridou, A., Biliaderıs, C. G., Bacandritsos, N., & Sabatini, A. G. (2004). Composi-
Alpaslan, M., & Hayta, M. (2007). Effect of soymilk substitution on the rheological tion, thermal and rheological behavior of selected Greek honeys. Journal of Food
and sensory properties of salep (traditional Turkish milk beverage). Interna- Engineering, 64, 9e21.
tional Journal of Food Properties, 10(3), 413e420. Marcotte, M., Hoshahili, T., & Ramaswamy, A. R. (2001). Rheological properties of
Anklam, E. (1998). A review of the analytical methods to determine the selected hydrocolloids as a function of concentration and temperature. Food
geographical and botanical origin of honey. Food Chemistry, 63, 549e562. Research International, 34(8), 695e704.
AOAC. (1995). Official methods of analysis (16th ed.). Washington, DC: AOAC Minitab. (2000). Computer program, MINITAB release 13.0 for Windows. USA: Minitab Inc.
International. MstatC. (1986). Version 4.00. East Lansing, MI: Michigan State Uni.
Arteaga, G. E., Li-Chan, E., Nakai, S., Cofrades, S., & Jiménez-Colmenero, F. (1993). Muir, D. D., & Hunter, E. A. (1991). Sensory evaluation of Cheddar cheese: order of
Ingredient interaction effects on protein functionality. Journal of Food Science, testing and carryover effects. Food Quality Preference, 3, 141e145.
58, 656e662. Nishinari, K., Zhang, H., & Ikeda, S. (2000). Hydrocolloid gels of polysaccharides and
Ayar, A., Sert, D., & Akbulut, M. (2009). Effect of salep as a hydrocolloid on storage protein. Current Opinion in Colloid & Interface Science, 5, 195e201.
stability of ‘Incir Uyutması’ dessert. Food Hydrocolloids, 23, 62e71. Ouchemoukh, S., Louaileche, H., & Schweitzer, P. (2006). Physicochemical charac-
Azeredo, L. C. C., Azeredo, M. A. A., De Souza, S. R., & Dutra, V. M. L. (2003). Protein teristics and pollen spectrum of some Algerian honeys. Food Control, 18, 52e58.
contents and physicochemical properties in honey samples of Apis mellifera of Razavi, S. M. A., & Karazhiyan, H. (2009). Flow properties and thixotrophy of
different floral origins. Food Chemistry, 80, 249e254. selected hydrocolloids: experimental and modelling studies. Food Hydrocolloids,
Azoubel, P. M., Cipriani, D. C., El-Aouar, A. A., Anonio, G. C., & Murr, F. E. X. (2005). 23, 908e912.
Effect of concentration on physical properties of cashew juice. Journal of Food Sopade, P. A., Halley, P., Bhandari, B., D’Arcy, B., Doebler, C., & Caffin, N. (2002).
Engineering, 66, 413e417. Application of the WilliamseLandeleFerry model to the viscosity-temperature
Baeza, R. I., Carp, D. J., Perez, O. E., & Pilosof, A. M. R. (2002). k-Carregeenaneprotein relationship of Australian honeys. Journal of Food Engineering, 56, 67e75.
interactions: effect of proteins on polysaccharide gelling and textural proper- Statistica. (1995). Statsoft statistica for windows, Release 5.0. Tulsa, OK: Statsoft Inc.
ties. Food Science and Technology, 35, 741e747. Telcioglu, A., & Kayacier, A. (2007). The effect of sweeteners and milk type on the
Bahrami, M., Ataie-Jafari, A., Hosseini, S., Foruzanfar, M. H., Rahmani, M., & rheological properties of reduced calorie salep drink. African Journal of
Pajouhi, M. (2009). Effects of natural honey consumption in diabetic patients: Biotechnology, 6(4), 465e469.
an 8-week randomized clinical trial. International Journal of Food Sciences and Vuksan, V., Jenkins, D. J., Spadafora, P., Sievenpiper, J. L., Owen, R., & Vidgen, E.
Nutrition, 60(7), 618e626. (1999). Konjac-mannan (glucomannan) improves glycemia and other associ-
Bhandari, B., D’arcy, B., & Chow, S. (1999). A research note: rheology of selected ated risk factors for coronary heart disease in type 2 diabetes. A randomized
Australian honeys. Journal of Food Engineering, 41, 65e68. controlled metabolic trial. Diabetes Care, 22, 913e919.
Bourne, M. C. (2002). Physics and texture. Chapter 3 Food texture and viscosity.558 concept White, J. W. (1979). Composition of honey. In E. Crane (Ed.), Honey: A comprehensive
and measurement (pp. 59e106). Harcourt Place, London, UK: Academic Press. survey (pp. 157e158). London: Heinemann.