The Composition, Rheological and Thermal Properties of Tajonal (Viguiera Dentata) Mexican Honey

Abstract

Samples of Tajonal honey (Viguiera dentata) from Yucatan, Mexico with different moisture contents were studied for their physicochemical, crystallization, rheological, and thermal characteristics. The presence of crystals changed the flow behavior from Newtonian to non-Newtonian. A characteristic glass transition temperature (Tg') was found for each sample, and it was observed to be dependent of the initial moisture content but independent of the storage period. The temperature and latent heat of fusion were not affected by the crystallization phenomenon, while both the moisture content and the glucose/water ratio were useful indexes for crystal growth in honey. Crystallization process can be controlled by harvesting the product with higher moisture contents without exceeding the limits established by international regulations.

Keywords:

• Honey
• Sugars
• Crystallization
• Glass transition temperature
• Fusion temperature

INTRODUCTION

Honey is a product composed primarily of sugars, the main ones being glucose and fructose (in nearly equal proportions), which together represent 85 to 96% of the total. The residual carbohydrates are disaccharides, trisaccharides, and oligosaccharides.[1] Because the sugar concentration is so high, honey sometimes takes a semi-solid state known as crystallized or granulated honey.[2] This is an undesirable process and must be prevented or retarded as much as possible, since it causes the product to become hazy and therefore to lose consumer acceptance. The rate of the crystallization process depends on factors such as the presence of crystallization nuclei, temperature, and concentration of sugars that may crystallize.[3] Honey management with centrifugal extractors and pumps favors the formation of small air bubbles that have a seeding effect.[4] Other factors that influence the tendency of the product to crystallize are the presence of impurities such as small dust particles, pollen, wax, and propolis, as well as temperature, relative humidity and type of container used for storage.[5] Honeys that crystallize within few weeks are considered to be fast-crystallizing, while those which take more than a year before crystallization are classified as slow-crystallizing.[6] Spontaneous crystallization is controlled primarily through proper storage, heating and/or filtering. Some attempts have been made to predict the crystallization tendency of honey, by studying the concentration of water, glucose, fructose[4] and melezitose,[7] as well as the glucose/water[6] fructose/glucose and (glucose-water)/ fructose ratios.[7]

The production of honey is an important economic activity worldwide, but it is very important to Mexico since it is the sixth producer and third exporter in the world.[8] Yucatan is the state that supplies the largest volumes of Mexican honey, producing up to 12,000 tons per year, from which approximately 90% is exported to international markets, mainly. This volume is produced owing to the wide diversity of native species that bloom in different seasons of the year. Among the species of prime importance in apiculture is Tajonal (Viguiera dentata), which produces approximately 40% of the annual volume.[9,10] However, honey derived from Tajonal crystallizes a few weeks after harvest. In terms of consumer appeal, granulated honey is generally regarded as unacceptable. Furthermore, when granulation is incomplete the crystalline layer is overlaid by a layer of liquid with higher moisture content than that of the original honey. This creates a favorable environment for the growth of yeast and may lead to fermentation.[11] On the other hand, this also generates problems at processing plants, since crystals hinder the flow of the product and eventually cause obstructions in the pipes. Crystallization is a complex physical process. In honey, it could be due to factors ranging from origin and composition of the raw material to processing, handling and storage of the final product. Bearing in mind the importance of the crystallization of Tajonal honey from the research and commercial point of view, the objective of this work was to study samples of Tajonal honey (Viguiera dentata) from Yucatan, Mexico with different moisture contents (16.2 to 20.6%) in order to determine its physicochemical, thermal, rheological, and crystallization characteristics. These factors were investigated in an attempt to determine specific precautions that could be taken to delay crystallization in Tajonal honey.

MATERIALS AND METHODS

Materials

Tajonal honey samples were collected from January to February 2001, from a beekeeping farm located in Muna, Yucatan (the place was selected due to its abundance in Tajonal flowering). Samples were collected from new honeycombs with a capped cell percentage of 14 to 100% (20.6 to 16.3% moisture content). Honeycombs were harvested in the laboratory using a stainless-steel centrifuge extractor. The resulting samples were labeled according to their moisture content: Sample 1 (16.3% moisture content); Sample 2 (17.6%); Sample 3 (18.6%); and Sample 4 (20.6%). This selection was done according to the maximum permitted limits established by the European Commission of Honey.[12] Samples of 500 g were stored in labeled plastic containers at 27.8°C (average temperature in the state of Yucatan during 2001). Physicochemical characterization was assessed 24 hours after harvesting; thermal, rheological, and crystallization studies were started 48 hours after the harvest of honey (time zero was considered to be the time when the first readings were made). One sub-sample was used for each analysis. Samples were taken every two weeks up to the 10th or 12th week of storage.

METHODS

Physical and Chemical Parameters

Moisture content, ash, pH, free acidity, electric conductivity, hydroxymethyl furfural, diastase, and invertase activities, as well as sugar composition were analyzed according to the methods proposed by the European Commission of Honey,[12,13] in order to characterize the samples and determine their compliance with the standards of the European Commission of Honey. Moisture was determined by measuring the refractive index with an ATAGO NAR-3T refractometer (Atago Co. Ltd., Tokyo, Japan) coupled to an ultra-thermostatic bath Grant W28 (Grant Instruments Cambridge Ltd. Barrington, Cambridge, U.K.). Free acidity was measured and acidity calculated as meq/kg. Conductivity was expressed as miliSiemens/cm (mS/cm). Hydroxymethyl furfural was calculated and expressed as mg/kg. Diastase activity was expressed in Schade units, which are defined as the amount of enzyme needed to hydrolize 0.01 g of starch at 40°C in one hour. Invertase results were reported as invertase units, defined as the g of sucrose hydrolyzed in one hour by the amount of enzyme present in 100 g of honey.

HPLC Analysis of Carbohydrates in Honey

The fructose, glucose, sucrose, and maltose content were assessed in Tajonal honey samples 1, 2, 3, and 4 according to the method proposed by the European Commission of Honey[13] with some modifications. The samples were analyzed on a Perkin Elmer 200 Series HPLC chromatographer with a RI detector. Separation of the carbohydrates was carried out on a GROM-SIL 120 Amino-2 PA column (4.6 × 250 mm). The flow rate was 1.3 mL/min using a mobile phase of acetonitrile:water (80:20 v/v), and a column temperature of 30°C. Mixed sugar standards were used in the analysis, containing 2 g fructose, 1.50 g glucose, 0.25 g sucrose, and 0.15 g maltose, which were weighed and dissolved in 40 mL HPLC water. The solution was transferred to a 100 mL beaker, which contained 25 mL methanol HPLC grade. Then the volume was completed with HPLC water up to 100 mL. Honey solutions were prepared with 50 g of honey and then followed the same procedure described for standards. Before injection onto the column, all solutions were filtered through a 0.20 μm membrane, and then 20 μL were injected into the chromatographer. All samples were analyzed in triplicate. Since previous researchers have attempted to predict crystallization tendency using ratios involving the composition of honey as compared to the content of glucose and fructose,[6,7] in the present study the ratios glucose/water, fructose/glucose, and (glucose minus water)/ fructose ratios were obtained, in order to relate the sugar composition of Tajonal honey with the crystallization phenomenon.

Crystallization Process

Shape, number and size of crystals

Nucleation kinetics, crystal growth, and shape must be known in order to control the formation of a correct number, size, and polymorphism of crystals.[14] An Axiophot Zeiss microscope equipped with software for image recording and analysis, and a KS400 vrs 3.01 Zeiss digital camera were used to calculate areas and the shortest length of crystals. A drop of honey was placed on a microscope slide and the shape of the crystals was observed, using the polarized light technique. The crystals were counted and the data recorded as the approximate number of crystals per kg of honey (number of crystals/kg). The area of the crystals was measured and expressed as μm² (crystal area).[14]

Viscosity

The measurement of viscosity was done according to the method reported by Mora-Escobedo et. al.[15] A Haake Fisions PK 100 Rheometer (Haake Buchler Instruments, Inc., Saddle Brook, NJ) with a plate and cone geometry was used, with a Pk 5 cone at a distance of 10 μm and a temperature of 25°C. 2 mL of honey were placed into the instrument. Triplicate analysis was done for each sample. Samples were equilibrated for not less than 25 minutes, and the rheometer was calibrated using standard liquids. Fluid consistency index (k) and flow behavior index (n) were calculated from the rheological model, also known as the Ostwald-DeWaele power law model, which relates shear stress (τ) and shear rate (γ) for non-Newtonian food materials:

The logarithmic transformation of Eq. (1) yields:

After obtaining τ and γ, a linear regression of ln (τ) and ln (γ) was computed. The intercept and slope values estimated ln (k) and n, respectively. The data fitted the power law model at shear rates until 300 s ⁻¹. Fluid consistency index and flow behavior index were then calculated in the range of 3 to 300 s⁻¹. The aim of this part of the study was not to compare the viscosity directly but rather to examine the deviation from Newtonian behavior. This deviation could then possibly be related to the extent of crystallization.

Phase transitions

The most common method used to determine phase transitions is the differential scanning calorimetry (DSC), which detects the change in heat capacity occurring over a transition temperature range.[16] Calorimetric heat flow was used to obtain qualitative and quantitative data regarding the net heat changes produced mainly by the carbohydrates in honey when heated. A TA Instruments Thermal Analyst 2000 System (New Castle, Deleware) DSC was used to follow the thermal behavior of the samples. The calorimeter's furnace was cooled to approximately −60°C with liquid nitrogen. The heating range was from −60 to 200°C to obtain the complete thermal behavior of pure honeys from low temperatures to high temperatures. The temperature changing rate was of 10°C/ min.[17] The instrument was calibrated according to the procedure used by Netto et al,[18] using the fusion point of indium (156.6°C) and double distilled water (0°C). At least three measurements were taken for each sample, using approximately 5 mg, in order to determine the Tg', defined as transition B (midpoint of the step) that it is due to the backbone of a large polymer or less mobile component[16] as well as the fusion behavior of the samples at higher temperatures, and to determine the effect of moisture on phase transitions.

Statistical Analysis

A one-way variance analysis (Software STATGRAPHICS PLUS version 4.0) was used to evaluate significant statistical differences among the studied parameters and storage periods. The comparison of means was performed using Duncan's method.[19] Significance was defined as p ≤ 0.05

RESULTS AND DISCUSSION

Physicochemical Parameters of Tajonal Honey

Table 1 shows the physicochemical parameters of Tajonal honey samples with different moisture contents. All are within the quality standards established by the European Commission of Honey (ECH).[20] The lowest moisture content (16.3%) was found in a sample extracted from a comb with 100% capped cells, while the highest content (20.6%) corresponded to the sample with a capped cells percentage of 15%. This is a very important finding from the commercial point of view, since it is possible to obtain ripe honeys with a lower percentage of capped cells. Therefore, Tajonal honey might be harvested earlier than others. The sample with 17.6% moisture presented the highest hydroxymethyl furfural and ash contents, and as consequence of these values, the highest electric conductivity. Even though these values were the highest among the samples, they were still far away from the limits established by the ECH.[20] The pH of the honey ranged among 3.61 to 3.79. This is a rather small variation (not a statistically significant difference p ≥ 0.05).

Table 1 Physicochemical parameters, sugar content and rations describing crystallization of Tajonal honey ^a/ .

Parameter	Sample 1	Sample 2	Sample 3	Sample 4	ECH ^1
Moisture (%)	16.3 ^a ± 0.07	17.6 ^b ± 0.05	18.6 ^c ± 0.05	20.6 ^d ± 0.06	21*
Ash (%)	0.15 ^a ± 0.04	0.26 ^b ± 0.06	0.15 ^a ± 0.04	0.15 ^a ± 0.03	0.6*
pH	3.64 ± 0.15	3.79 ± 0.09	3.61 ± 0.11	3.67 ± 0.10
Free acid (meq/kg)	19.6 ^a ± 0.7	24.8 ^b ± 1.2	18.8 ^a ± 0.8	19.2 ^a ± 0.9	40*
Conductivity (mS/cm)	0.22 ^a ± 0.05	0.36 ^b ± 0.03	0.19 ^a ± 0.03	0.22 ^a ± 0.04	0.7*
Hydroxymethylfurfural (mg/kg)	3.7 ^a ± 0.6	6.8 ^c ± 0.7	5.2 ^b ± 0.4	2.8 ^a ± 0.6	40*
Diastase activity (U.Schade)	31.7 ^c ± 1.0	16.2 ^b ± 0.8	9.0 ^a ± 0.7	12.3 ^ab ± 0.7	8**
Invertase activity (NI)	22.5 ^c ± 0.8	17.4 ^b ± 0.5	15.9 ^b ± 0.5	12.2 ^a ± 0.3	10**
Glucose + Fructose (%)	78.2 ^b ± 1.2	76.5 ^a ± 0.8	76.1 ^a ± 0.9	76.9 ^a ± 1.0
Glucose (%)	36.8 ^b ± 0.9	35.2 ^a ± 0.7	35.3 ^a ± 0.5	36.5 ^b ± 0.8
Fructose (%)	41.4 ^b ± 1.0	41.3 ^b ± 0.8	40.8 ^ab ± 0.5	40.4 ^a ± 0.6
Sucrose (%)	0.04 ^a ± 0.01	0.25 ^b ± 0.05	0.03 ^a ± 0.01	0.04 ^a ± 0.01
Maltose (%)	1.17 ^a ± 0.2	1.82 ^b ± 0.3	1.13 ^a ± 0.2	1.11 ^a ± 0.2
Glucose/water	2.26	2.00	1.90	1.77
Fructose/glucose	1.13	1.17	1.16	1.11
(Glucose-water)/fructosa	0.50	0.43	0.41	0.39

The enzymatic activities of diastase and invertase are shown in the same Table. The activity of both enzymes decreased as moisture increased. There was a strong correlation between moisture content and invertase activity (r = 0.9459), since sucrose that comes from the nectar is hydrolyzed in glucose and fructose by the action of invertase.[21] The presence of invertase in high concentrations shows that honey has not yet reached a suitable maturation point. On the other hand, diastase activity showed a weaker correlation with moisture content (r = 0.770).

Tajonal honey was mainly composed by glucose and fructose (76.1 to 78.2%) (Table 1). The sample with the highest content of these two sugars had also the lowest moisture content, showing a significant difference when compared to the rest of the samples (p ≤ 0.05). Since sample 1 had the highest sugar concentration, and the lowest moisture content, it can be expected that this sample will also have a higher tendency towards crystallization. Glucose content varied from 35.2 (sample 2) to 36.8% (sample 1) and it was higher than the glucose content found in a sample of honey from Alberta, Canada.[4] Tajonal honey was supersaturated with glucose, and therefore has the potential to cause spontaneous crystallization at room temperature in the form of glucose monohydrate. Bogdanov[20] reported that honey samples with glucose contents higher than 30% crystallized faster than those with lower contents.

The four samples analyzed in the present study had fructose concentrations between 40.4% and 41.4%, which were similar to the concentrations reported by Assil et al.[4] for fast-crystallizing honeys (35.4–41.7% fructose). The highest fructose concentrations were found in samples 1 and 2, followed by 3 and 4. Given those results, it is highly probable that the four samples under study will crystallize after short time.

The amounts of sucrose and maltose were not statistically different among the samples except for sample 2, which had contents of 8.3 or 1.6 times higher than the rest. The glucose/water relation, used as an indicator of crystallization, ranged between 1.77 and 2.26 (sample with the highest moisture content to the sample with the lowest moisture content). White[6] reported that values higher than 2.10 indicate fast crystallization, while values lower than 1.70 correspond to honey samples that remain fluid for a long time. Therefore, it is expected that sample 1 will crystallize the fastest, followed by sample 2, 3, and 4.

Honey samples with an unusual high percentage of fructose remain liquid for a long time, thus the fructose/ glucose and (glucose-water)/ fructose ratios were determined in order to consider the effect of fructose composition.[4] The values for the relation fructose/glucose were 1.13, 1.17, 1.16, and 1.11 (samples 1, 2, 3, and 4, respectively). A fructose/glucose ratio lower than 1.14 indicates a fast crystallization[6] and therefore samples 1 and 4 are expected to crystallize rapidly, while samples 2 and 3 should crystallize after a longer time, although in a similar manner.

Tajonal honeys presented a ratio (glucose-water)/ fructose of 0.50, 0.43, 0.41, and 0.39 (samples 1, 2, 3 and 4, respectively). Those samples with a ratio (glucose-water)/ fructose higher than 0.50 crystallize in a short time; with a ratio of 0.20 remain fluid for years, and the ones with intermediate values show an irregular behavior.[7] Bearing in mind the concentration of sugars and the ratios glucose/water; fructose/glucose; (glucose-water)/ fructose of the studied samples, it is possible to predict the crystallization of all of them. However, the speed at which this phenomenon will take place could be different, as it is also influenced by moisture content. Sample 1 would present a fast crystallization, while the rest would show irregular behavior.

Evaluation of the Crystallization Process

At the beginning of the experiment (time zero), crystallization nuclei was observed in all the samples under study (Fig. 1). This finding can be supported by the sugar over-saturation that was observed in those samples. The use of a centrifugal method of extraction for the honey may also play a relevant role, according to some empirical evidence that supports the bubble theory, which states that some crystals can be seen next to bubbles.[4] There has been reported that the presence of numerous crystallization nuclei with diameters smaller than 30 μm causes a rapid crystallization of honey,[22] which agrees with the observations of the four samples of this study. At zero time, the four samples averaged 2.67 × 10⁵ nuclei/kg honey, with areas of 35.7 μm²and a shortest length of 4.9 μm. The size distribution was within the interval of 1–50 μm² for 98% of the crystals. It is noteworthy that the nucleation kinetics, crystal growing, and formation must be known in order to control the formation of the number of crystals, size, shape and polymorphism.[14]

Figure 1 Crystallization “nuclei” from sample 1 (1000X).

Growth morphology of honey crystals

Fig. 2 shows the evolution of the crystal habit in sample 1 (16.3% moisture) during storage. The samples were examined under different magnifications (1000X [a, b], 400X[c], 200X[d] four, six, eight, ten weeks of storage) due to the large size of the crystals. The photographs depict size, shape and morphology of the crystals along storage time. The honey crystals showed a well defined lattice (Fig. 2) as pentagonal and hexagonal crystals. This crystallization pattern results from the formation of monohydrate glucose crystals,[4] which vary in the number of shapes, dimension and quality according to the composition of a particular sample of honey (Table 1). The ratio length/height (b/a) was 2.0, 2.4, and 3.9 (6, 8, and 10 weeks respectively), this shows the elongation of the habits towards the b-axis. In the first two to four weeks, a growth in the negative direction of the positive end of the polar b-axis of the crystals was observed. The growth rate was initially higher than during the later growth phase, changing the lattice from a hexagonal to a pentagonal form (Figs. 2c, d). These effects can be influenced by the presence of other components of honey such as fructose, sucrose, and maltose as reported by Sgualdino et al.[23] and Wang et al.[24] Samples 2 and 3 showed a crystal pattern similar to sample 1, however in this case no important changes were found in the b/a ratio between 6 and 10 weeks of storage (1.6 and 1.9 for samples 2 and 3, respectively). This means that the sizes of the crystals were smaller than sample 1. Sample 4 showed a different crystallization pattern, being b/a ratio 2.0, 2.4, 2.2, 2.3, and 2.9 (2, 4, 6, 8, and 10 weeks of storage). In spite of it, it was possible to detect crystals earlier than in samples 1, 2, and 3, the elongation of the habits toward the b-axis was not significant, and the crystal size was smaller than the other samples. The slow crystal growth rate of glucose, which depends on the degree of super saturation, causes that glucose as a part of sucrose can be incorporated into the lattice of the growing crystal. Then, an equilibrium state establishes between the adsorbed molecules and those going back into solution, which apparently controls the rate of growth.[24] The different behavior among the samples under study can be explained by the systematic effect of the other components in the bulk structure of the crystal,[23,24] but strongly relates to water content: the lower the water and the higher the glucose content of honey, the faster the crystallization.

Figure 2 Evolution of the crystal habit in sample 1 (16.3% moisture) during storage. a) 4 weeks (1000X); b) 6 weeks (1000X); c) 8 weeks (400X); and d) 10 weeks (200X).

Number and shape of crystals

A significant increase in the number of crystals (p ≤ 0.05) was observed during the first four weeks of storage in all samples (4.9 × 10⁶, 2.6 × 10⁶, 3.0 × 10,⁵ and 4.3 × 10⁵ crystals/kg of honey for samples 1, 2, 3, and 4 respectively). No important changes in sizes in sample 2 and 3 were found. However, in sample 1 it was possible to see the crystal structure increase threefold as compared to time zero. A change in the size distribution of the crystals occurred, being 86, 95, 93, and 90% within the range of 1–50 μm² for samples 1, 2, 3, and 4 respectively (Fig. 3a). At the sixth storage week, it was observed that sample 1 followed with a higher increase in crystal size, until reaching 2.5 times the crystal size obtained in the fourth week. Samples 2, 3, and 4 showed a slighter growth. With regards to size distribution, 61, 60, 90, and 53% of the crystals had sizes smaller than 50 μm² (1, 2, 3, and 4 respectively). As seen in Fig. 3b, the crystals of every sample had begun to increase their sizes at this point. Sample 4 had a higher percentage (34.5%) of crystals with sizes larger than 300 μm² than the ones found in sample 1 (5.5%). The largest crystals in sample 4 had an area of 879 μm², while in sample 2 the largest was of 14,845 μm². Different crystal sized were evident for each sample. Similar findings were reported by Goltz[25] in Australian samples of honey. The author reported a number of very large crystals suspended in a liquid matrix.

Figure 3 Tajonal honey crystal size distribution a) 4 weeks; b) 6 weeks; c) 8 weeks; d) 10 weeks of storage.

After eight weeks of storage, sample 1 was the only one that significantly increased the number of crystals (5.7 × 10⁶ crystals/kg of honey). Non statistical differences (p ≥ 0.05) were found in samples 2, 3, and 4, since they maintained similar values as compared to the former period: 3.9 × 10⁶, 3.3 × 10⁵, and 9.5 × 10⁵ respectively. Along this storage week, sample 2 had a fast crystal growth, increasing the rate 10 times as compared to the former period, even though it did not show an increase in the number of crystals. Meanwhile, the crystals in sample 1 increased their sizes fourfold, in sample 3, crystal size increased threefold, and sample 4 showed no significant increase. As it can be observed in Fig. 3c, during this period the crystals grew in size, except for sample 4 that still showed a high percentage (84%) within the 1–50 μm² interval. Samples 1 and 2 also had a large percentage of crystals (89% and 86%) in the interval of 50–100 μm². Sample 3 had a more heterogeneous size distribution, being 44% of the crystals within the 1–50 μm² interval, 24% within the 50–100 μm² interval, 6% within the 200–300 μm² interval, and 26% had a size larger than 300 μm².

At the tenth week of storage (Fig. 3d), samples 1 and 2 did not show an increase in the number of crystals (6.0 × 10⁶ and 3.1 × 10⁶ crystals/kg respectively). However, the number of crystals in samples 3 and 4 did increase significantly as compared to the former period. Crystals in sample 1 increased their sized almost twofold, 88% being larger than 300 μm² (Fig. 3e), with areas up to 79,000 μm². On the other hand, the growth of crystals in sample 2 placed 92% of them within the 150–200 μm² interval. Sample 3 had 65% of its crystals in the range of 1–50 μm², and 34% had sizes larger than 300 μm². Crystals in sample 4 increased their sizes twofold as compared to the 8^th week, being 85% within the 1–50 μm² interval, and 13% having sizes larger than 300 μm², with a maximum size of 2540 μm².

It has been reported that lactose crystals with sizes between 4244 and 25,462 μm² were sensory inadequate.[14] Similarly, the large sizes and high numbers of crystals found in sample 1 at the tenth week were responsible for a poor sensory quality of the product. This effect was not observed for samples 2, 3, and 4 since the size of their crystals were much smaller than those in sample 1. Both the sizes and the number of crystals in sample 1 at this point agreed with the expectations made upon its sugar composition and relations glucose/water, fructose/glucose and (glucose-water)/ fructose, all of which pointed to a fast crystallization rate.

The 92% of the crystals formed in sample 2, as previously mentioned, were among the 150–200 μm² range. This confirms the results given by the glucose/water relation, which indicated that sample 2 should crystallize after sample 1. At the tenth week of storage, sample 4 had a high percentage of small-sized crystals. However, they were present in large numbers as compared to sample 3.

It was observed that the formation of the largest crystals was directly related to the moisture content. Significant differences (P ≤ 0.05) were found among the samples from the fourth week of storage, with the exception of sample 2 that showed a slower crystal growth rate. This may be due to the much higher content of sucrose and maltose of this sample, which in this case act as impurities retarding crystal growth.[23,24]

The low moisture content of sample 1 and its high concentration of glucose and fructose resulted in a higher over saturation level, leading to a better crystallization. Both moisture content and sugar composition have an influence over the crystallization process of honey, in such way that at lower moisture contents, higher crystallization levels,[8] and at higher solute saturation, the percentage of large crystal sizes increases.[14]

The different indexes previously mentioned have characterized Tajonal honey as fast-crystallizing. This means that it shall crystallize within few weeks, which was evident after four weeks of storage. However, these indexes are not enough to describe the crystallization sequence of the samples, as well as the maximum size reached by the crystals. The relation glucose/water was the one that better predicted the size reached by the crystals at the tenth week of storage, with the exception of sample 2. The predictions made by the relation (glucose-water)/ fructose approximated to the results observed in sample 1.

With regards to crystal shape, Fig. 2 shows the state of the crystallization nuclei after four weeks of storage. They presented an oval-shaped external appearance (habit), as opposed to those crystals observed after eight weeks of storage, which were hexagonal-shaped. Since the crystallization process is a multi-factorial phenomenon, controlled by variables such as over saturation, temperature, pH, presence or absence of contaminants, that affect both the crystal growth rate and their shape, it may vary once variables have been modified.

Viscosity

Flow behavior of honey samples along the crystallization process is described in this section and it was studied until the 12^th week of storage. The viscosity showed a Newtonian behavior in samples 2, 3, and 4 up to the 12^th storage week, since the flow behavior indexes (n) were equal to 1. Same behavior was reported by Sopade et al.[25] in Australian honeys. Sample 1 showed a Newtonian flow behavior up to the 10^th week of storage. However, at the 12^th week it presented a strong tendency to behaving as a non-Newtonian fluid. Generally, honey is a Newtonian liquid but non-Newtonian behavior has been reported,[26] and then with the obtained data, the power-law model was applied. The characteristic flow parameters k and n were estimated from the shear force and shear rate (Table 2). In a Newtonian fluid, such as water, flow behavior index (n) is the unit, and the fluid consistency index (k) corresponds to viscosity. In Table 2 also shows the r² value, which represents how well this equation fits the data, values 1 and −1 represents a good fit of the data.

Table 2 Flow behavior index and consistency index for Tajonal honey (16.3% moisture content) along storage.

Storage weeks	Consistency index (k) (Pa.s)*	Flow behavior index (n) (Dimensionless)	(r^2)
0	3.5	1.00	0.989
2	3.6	1.01	0.988
4	3.7	1.03	0.975
6	3.5	1.01	0.991
8	4.5	1.06	0.983
10	4.5	1.04	0.992
12	17.0	0.69	0.998

It can be seen that the values of k were all similar up to the sixth week of storage. Along that period, sample 1 increased its number of crystals. However, their sizes did not increase significantly since 60% of them were still in the size range of 1–50 μm². Important changes were found only at the eighth and tenth weeks of storage when crystals increased their size significantly, which in turn increased the consistency index. This resulted in a viscosity increase due to the presence of large-sized crystals. Up to this moment, the flow behavior index was of approximately 1, indicating a Newtonian fluid behavior.

At the 12^th week of storage, the flow behavior index was smaller than the unit (n < 1) and the consistency index increased abruptly, changing the Newtonian to a non Newtonian fluid behavior. Goltz[26] reported similar results in Australian crystallized samples, however it is noteworthy that the change in flow behavior may be due to the type of test, since the decrease in viscosity (that leads to the change in flow behavior) may have occurred as a consequence of the rupture of crystals when the sample was under shear force. This hypothesis was corroborated by taking a small amount of honey from the central part of the sample and the periphery of the cone, before and after the viscosity test. The analysis was assessed three times as viscosity determination. These samples were analyzed in a Axiophot Zeiss microscope and then the images of crystals before and after the test were compared (Fig. 4). It was observed that the crystals from the central part had smaller sizes after the test(Fig. 4b). On the other hand, peripheral crystals showed a large size Figure 4c) and therefore it is possible to say that the change in flow behavior for the sample at the 12^th week of storage was due to the migration of larger particles away from the center of the cone.

Figure 4 Crystal images before and after viscosity test. a) Image before test; b) Crystals from central part; c) Peripheral crystals.

Phase Transitions

In all analyzed honey samples in a range of −60 to 200°C, two thermal phenomena were observed (Fig. 5). A Tg' (−55.5°C to −35.4°C) is seen, often linked with a relaxation effect the amplitude of which depends on the thermal history of the samples, or the temperature at which polymeric materials change from amorphous solid (glass) to an amorphous rubber and it is specific for each amorphous material.[18] A very wide and intense endothermic reaction due to the fusion temperature of the sugars present in the samples varied from 119.5°C to 133.6°C, with an average of 126.3 ± 3.7°C. The peak observed in Tajonal honey samples begun at 93–123°C and ended between 140 and 175°C, depending on the particular sample. This peak was similar to the second one found by the same author, which corresponds to the fusion of sugars (mono, di-, tri-, and oligosaccharides).[18] The first peak described elsewhere was not found in Tajonal honeys. This peak is attributed to the fusion of the water/sugars/starch complex, and it is an important finding indicating a particular composition of Tajonal honey. Since this honey does not contain high levels of starch or a sugar other than glucose, fructose and small amounts of sucrose and maltose, the thermal parameters extracted from this curves can be successful to characterize Tajonal honey, as Cordella et al.[27] reported in their study of the adulteration effect on honey thermal behavior. DSC analysis in Tajonal samples did not show the presence of a typical endothermic peak in the range −20 to 200°C linked to the free water content of the samples, because the water could be included in the sugar network.[17,27]

Figure 5 DSC thermogram of sample 1 showing: A) Tg; B) Tg; and C) Sugar Fusion Temperature.

DSC thermograms show transition 2 (marked as B in Fig. 5) or Tg' in a temperature range of −55.5°C to −35.4°C, and significant differences were found among the four samples (P ≤ 0.05) (Table 3). The lowest transition temperature was in sample 4 (−53.3°C average), followed by samples 2 and 3, which had intermediate temperatures (−42.7°C and −42.0°C respectively), and sample 1 with the highest temperature (−37.7°C). These differences can be attributed to the different moisture contents of the samples. The Tg' decreased with an increase in moisture and the effect could be more pronounced as the estimated average crystal size decreased, as was reported by Schaller-Povolny, et al.[28] for inulin. Roos and Karel[29] in a model sugar/water systems, also reported that at higher moisture contents, Tg' tended to present lower values. Other authors reported that the glass transition of honey samples with different moisture contents was the lowest for those samples with high moisture levels.[17] This is due to the role that moisture plays as a plasticizing agent in the thermal behavior of a food, causing an increase in molecular mobility and reaction rate, as well as a decrease in viscosity.[29] Tg' values for Tajonal honeys were similar to the ones reported for pure honey (−42.5°C and −45°C to −39°C respectively).[17,27] No significant differences (p > 0.05) in the Tg' of the samples were found for the storage periods.

Table 3 Glass transition temperatures (Tg') for Tajonal honey at different moisture contents ^a/ .

		Tg ^1 (°C)
Simple	Moisture (%)	Tgi	Tgm	Tgf
1	16.3%	−40.6 ^a ± 1.9	−37.9 ^a ± 0.6	−36.9 ^a ± 2.0
2	17.6%	−43.6 ^a ± 0.6	−43.3 ^b ± 0.6	−39.7 ^b ± 1.6
3	18.6%	−43.4 ^a ± 4.8	−41.6 ^b ± 1.2	−38.1 ^a ± 2.1
4	20.6%	−55.5 ^b ± 2.1	−47.6 ^b ± 0.1	−48.0 ^b ± 0.1

The fusion temperature of the sugars present in the samples varied from 119.5°C to 133.6°C. Since honey is a mix of sugars with approximately 95% of glucose and fructose, its fusion temperature is given by the combination of glucose and fructose with fusion temperatures in a range of 146°C and 103°C respectively.[30] When multiple solutes present in a mixture like food show a single fusion point it means that all solutes interact and act as a single solute.[16] On the other hand, Cordella et al.[27] found in un-adulterated honey a similar very wide an intense endothermic peak at 100–120 to 180–220°C corresponding to the fusion of sugars (mono-, di-, tri-, and oligosaccharides). Hurtta et al.,[31] reported that sugars do not have sharp melting temperatures and their melting proceeds over a temperature range. Even though there was no change in fusion temperatures, sample 1 showed a large gap between the beginning and the end of the transition process, as the crystals were appearing. At the beginning this study, sample 1 had a Ti of 110°C and a Tf of 149°C, with a 39°C difference. At the 12^th week of storage, Ti was 93°C and Tf 156°C, the difference being 63°C. This behavior was not observed in samples 2, 3, and 4 (Fig. 6). This phenomenon can be explained by the longer time that glucose requires to go from liquid (fluid honey) to solid (crystallized honey), which results in a wide temperature range. With regards to the latent heat of fusion of Tajonal honeys, it was found to be statistically the same (P ≥ 0.05) among storage periods, which indicated that an important change in heat was not required while the size and number of crystals were growing. However, statistically significant differences (P ≤ 0.05) were found among the samples (Table 4). Enthalpies of fusion appear slightly higher than those reported for Robinia and Lavandula honeys.[27] In order to melt, samples 1 and 4 required higher energy levels than samples 2 and 3, with averages of 355.1 ± 23.2 J/g and 361.8 ± 18.6 J/g (samples 1 and 4), and 280.4 ± 9.2 J/g and 294.6 ± 15.9 J/g (samples 2 and 3). The similar behavior observed for samples 1 and 4 is not explained by the moisture content or the crystallization degree of the samples, but by the content of glucose and the sum glucose+fructose, which were statistically the same for both samples. In this case, a high correlation (r = 0.9743) was found between the latent heat of fusion and the glucose content, but there was a lower correlation between the latent heat of fusion and the sum glucose+fructose (r = 0.7256), indicating that glucose concentration was the factor of highest importance that determined the amount of heat required in this thermal transition. Results reported,[27] indicated that the sugars are responsible for the amount of heat required for Tajonal honey to melt. In the same work, fusion enthalpies of −256 J/g for Lavandula honeys, and −213 J/g for Robinea were reported, which were similar to the ones found for samples 2 and 3, but lower than the ones for samples 1 and 4. This corresponds to a higher concentration of glucose, since these authors found higher fusion enthalpies in honey samples added with glucose syrup, reporting values of −560.8 J/g to −466.3 J/g.

Table 4 Heat fusion (J/g) Tajonal honey during storage ^a/ .

Weeks of storage	Sample 1	Sample 2	Sample 3	Sample 4
0	325 ± 13.1 ^b	296 ± 5.3 ^a	291 ± 6.0 ^a	336 ± 12.6 ^b
2	333 ± 11.4 ^b	283 ± 7.5 ^a	283 ± 6.3 ^a	364 ± 9.3 ^b
4	347 ± 9.2 ^b	280 ± 8.2 ^a	301 ± 10.1^a	361 ± 4.7 ^b
6	394 ± 10.2 ^b	268 ± 12.7 ^a	280 ± 9.8 ^a	342 ± 7.1 ^b
8	368 ± 7.9 ^b	278 ± 5.5 ^a	327 ± 3.0 ^a	387 ± 10.3 ^b
10	360 ± 8.7 ^b	286 ± 4.2 ^a	286 ± 7.1 ^a	381 ± 11.1 ^b
12	345 ± 15.1 ^b	272 ± 9.7 ^a	294 ± 8.3 ^a	362 ± 13.0 ^b

Figure 6 Differences between initial and final fusion temperatures along storage.

CONCLUSIONS

From the commercial point of view an important finding in this study was that the Tajonal honey might be harvested earlier than others, since it is possible to obtain ripe honeys with a lower percentage of capped cells. Tajonal honey will crystallize invariably, due to the particular characteristics given by its chemical composition. This composition is given by the nectar of the floral source, which has a high sugar concentration (mainly glucose). In honey samples with low moisture percentages, crystallization took place within a short time (12 weeks). However, this time was longer as the moisture percentage of the samples increased. The Tajonal honey sample with an intermediate moisture content (18.6%) presented a lower crystallization rate than the sample with the lowest moisture content (16.3%). After ten weeks of storage, it remained fluid without a high number of large-sized crystals. The rheological measurements indicated an increase in the viscosity of the samples as the number and size of glucose crystals increased. Nevertheless, the change of Newtonian fluid to pseudoplastic should be taken with reserves, since it was due to the nature of the test, which caused rupture of the crystals. The thermal measurements were not adequate to describe the crystallization of the samples. However, they allowed a differentiation of them regarding moisture content and sugar composition. The crystallization of Tajonal honey was unavoidable; however it may be delayed by increasing the moisture content of the product, or by treating it thermally in order to dissolve the crystallization nuclei. Nevertheless, this last option also diminishes the physicochemical quality of honey and does not avoid crystallization in the future. Another solution is to influence the crystallization phenomenon in order to obtain crystallized products with acceptable sensory characteristics.

ACKNOWLEDGMENTS

This study was supported by fellowships from COFAA– IPN and a grant from CONACYT. One of the authors (Y. Moguel-Ordoñez) wish to thank CONACyT for PhD scholarship.

Notes

¹European Commission of Honey;

^a/ Average of three experiments ± Standard deviation;

^a,b,c,d Different letters in the same row indicate significantly difference (P ≤ 0.05) according Duncan's test;

*Maximun **Minimum.

*1 Pa.s = 1000 centipoise (cP).

^a/ Average of three experiments ± Standard deviation;

^a,b Different letters in the same row indicate significative difference (P ≤ 0.05) according Duncan test;

¹Average value for Tg' from all data during the storage.

^a/ Average of three experiments ± Standard deviation;

^a,b Different letters in the same row indicate significantly difference (P ≤ 0.05) according Duncan test.