Chemical Analysis and Quality Assessment of Honey Obtained from Different Sources

: The purpose of this paper was to evaluate the quality of bee honey from different sources: beekeeper, local market and organic honey. Sensory analysis was performed and the water content, pH, acidity, protein content and total metal content (Cu, Cr, Mn, Co, Ni, Pb, Cd, Fe) were determined. The sensory analysis was carried out by a group of untrained panelists for quality assessment of honey. The metal content was determined by graphite atomic absorption spectrometry (GTAAS). Mineralization was carried out in a microwave digestion system, in a high-pressure polytetraﬂuoroethylene (PTFE) vessel, using a standard acid-digestion protocol. The results regarding the physico-chemical parameters showed that the honey samples were in accordance with the quality regulations for honey as a commercial product. The concentration of metals in the investigated honey samples varied in the order Cu > Cr > Pb > Fe > Ni > Mn > Co > Cd, the values being within the limits established by the EU Commission (No. 1881/2006). The variations observed in the evaluated parameters can be caused by the difference in plant species from which the honey comes, the harvesting period and the level of environmental pollutants. The Pearson correlations between the physico-chemical parameters and the metals indicate that water content (w c ) is strongly negatively correlated with Cd and Ni, while pH is strongly positively correlated with Mn and Fe. Moreover, EC is strongly negatively correlated with Ni and Fe, and the Brix degrees are strongly positively correlated with Cd and Ni. Statistically signiﬁcant positive correlation was found between Brix–Cd, Ni–Cd and Cu–Cr and a statistically signiﬁcant negative correlation was detected between w c and Cd.


Introduction
Bee honey (Apis mellifera) had an important role in history. The first records of beekeeping date back to 7000 BC (cave paintings in Spain), and bee fossils are known from approximately 150 million years ago. Honey was used as an embalming material, to preserve fruits and to obtain alcoholic beverages mixed with wine. Honey is mentioned in the Koran and the Bible [1]. It is produced from the nectar of plants, by the secretion of the living parts of plants or by the secretions of insects subjected to the living parts of plants, by which bees collect the nectar, transform it by combining it with specific substances of their own, deposit it, dehydrate it, store it and leave it in the honeycombs to ripen and mature [2]. It has antioxidant, antimicrobial and anti-inflammatory properties [3]. The antioxidant properties are given by the content of phenolic acids (the main ones being gallic acid, vanillic acid, caffeic acid and flavonoids, such as quercetin, kaempferol and chrysin). Antimicrobial activity in honey is due to acidity, hydrogen peroxide generation and osmolality. For high antibacterial protection, honey must have low glucose-oxidase activity. The anti-inflammatory properties of honey include the healing of wounds, the recovery of wounds at the level of gluteofemoral fistulas, ophthalmic problems due to cataracts and to fighting bacteria. It can inhibit the bacteria Pseudomonas aeruginosa, Staphylococcus hemolticus, Escherichia coli and Staphylococcus aureus [3]. Due to the high sugar content, microorganisms do not develop in honey. The sugars identified in honey are maltose, sucrose, maltulose, isomaltose, turanose, laminaribiose, nigerose, kojibiose, gentiobiose and oligosaccharides. It contains many nutrients, such as glucose (31%) and fructose (38%), as well as amino acids, enzymes, minerals, proteins, vitamins, organic acids and phenolic compounds [4].
The quality of honey depends on the water content, the fructose/glucose ratio, the geographical origin of the flowers, the environmental factors, the season and the eventual treatment by the bees [5][6][7][8][9][10]. Regulation by the European Union has established a common quality standard for bee honey for the characterization of honey or honey mixtures, authenticating origin and detecting falsification. It includes the identification of organoleptic characteristics (consistency, smell, aroma, color, taste), and physico-chemical parameters (water, sugars, mineral contents, acidity, organic acids, vitamins, amino acids, electrical conductivity, proteins, proline, HMF content and enzyme activities). The acidity of bee honey is represented by its organic acid content (acetic, butyric, citric, formic, lactic, oxalic, maleic, succinic and malic acids). The proteins in honey can come from the bees' salivary glands or from pollen. The mineral content (K, Ca, Na, Ca, P, Cu, Fe, Mg, As, Co, I, Mn, Cd, Ba, Li, Ni, Ag, Cr, Se, Zn, As, Pb, Hg and Cd) can be determined by different methods, including ICP-MS, ICP-OES, ETAAS, fluorimetric or electrochemical methods [11][12][13][14][15][16][17], and depends on the botanical origin of the flowers from which the nectar is extracted, the pedoclimatic conditions and the extraction technique. A low humidity means that bee honey does not ferment, the biological activity being reduced, and it can be kept for a long time [18]. Water-soluble vitamins are present in honey (vitamin C and vitamins from the Bcomplex). Turkey, China, Argentina, the USA and Ukraine are the largest global producers of bee honey. Romania is among the top European honey producers after Germany, France and Hungary [19].
The purpose of this paper was to evaluate the quality of bee honey from different sources (beekeeper, local market and organic honey). The organoleptic analysis of the honey samples was carried out and the water content, pH, acidity, soluble solids content (SSC), protein content and total metal content were determined (Cu, Cr, Mn, Co, Ni, Pb, Cd, Fe).

Samples
Six different samples of acacia and polyfloral honey were collected from different sources: beekeeper, local market (Dobrogea area) and organic honey. Beekeeper honey samples were collected from Techirghiol Municipality (Dobrogea area) and organic honey from Botosani County. Market acacia honey was supplied from Blaj town, Alba County and market polyfloral honey was supplied from Ion Roata village, Ialomita County.

Organoleptic Analysis
Organoleptic properties were analyzed according to the national standard SR 784-3:2009 [20]. The samples were homogenized by mixing with a glass rod, filtered and left until completely clear, after which they were subjected to organoleptic analysis (consistency, color, smell and taste). There was no crystallized sample, only a trace of crystals, therefore heating to 45 • C was not necessary. Honey flowing consistency was determined with a glass wand or a wooden spatula, asserting the status: watery, light liquid, viscous liquid or sticky. The color was determined by visual examination (day light, white background) with honey being in one colorless glass ampoule Φ 16 mm. The smell and taste were determined by smelling and tasting the sample. The dominant flavor was written down for multifloral honey and the intensity (pronounced, well highlighted, moderate, discreet) was recorded. The sweet taste intensity (pronounced, well highlighted, moderate) and the random secondary flavors (sour, bitter, astringent, tasteless) were also determined. The sensory test was conducted by a group of untrained panelists (male and female) between 20 and 25 years old who were recruited from the University Ovidius of Constanta (Romania) and the results are presented in Section 3.
This test is important for understanding the product characteristics and its impact on consumer acceptability.

Determination of Water Content (w c )
The water content was determined using an oven (Memmert GmbH + Co. KG, D-911186, Shwabach, Germany) and the AOAC Official Method 969.38 [21]. Water content was determined as the difference between the weight of honey (1-2 g) and the weight recorded after 24 h heating at 105 • C; the results were expressed as a percentage (%).

pH and Free Acidity
The pH of honey was determined for a 10% solution, using a multiparameter (Model C3030; CONSORT, Sepadin, Bucharest, Romania) [22].
Free acidity was given as the total free acids of honey and was expressed in meq/kg honey [23]. Free acidity was carried out volumetrically, by acid-base titration with NaOH 0.1 N.

Electrical Conductivity (EC)
In this work, 20 g of sample was dissolved in 100 mL of distilled water, and the conductivity of the solution, expressed in µS/cm, was measured using a calibrated digital apparatus (pH/Cond 340i SET 1; Weliheim, Germany).

Soluble Solids Content (SSC)
The • Brix value of each sample was measured using a refractometer Atago Abbe NAR-1T Solid (Kobe, Japan) and soluble solids (sugar) content was calculated (SSC, %).

Protein Content Determination
The protein content in honey was determined using the Kjeldahl method, based on the conversion of organic nitrogen present in the analyzed samples, using 6.25 as a conversion factor [21].

Determination of Cu, Cr, Mn, Co, Cd, Ni, Pb, Fe by Graphite Atomic Absorption Spectrometry
Mineralization was carried out in a microwave digestion system (Berghof Speedwave ® ; Thuringia, Germany) in a high-pressure polytetrafluoroethylene (PTFE) vessel using a standard acid digestion protocol [24].
A sample amount of 0.3 g was accurately weighed into a PTFE vessel and 10 mL of HNO 3 (65%) and 2 mL H 2 O 2 (35%) were added. The vessel was closed and placed inside the microwave digestion system for 35 min at 200 • C (5 min ramp to 135 • C, 5 min at 150 • C and maintained at 190 • C for 15 min, 10 min at 70 • C). At the end of digestion, the samples were removed from the digestion oven, cooled at room temperature and diluted to 50 mL final volume with deionized water. Finally, the solutions were filtered using a 0.45 µm pore size filter. The metal content was determined using graphite atomic absorption spectrometry ((GTAAS) model: Contra 800; Analytica Jena Instruments, Thuringia, Germany).
A multi-element standard (ICP multi-element standard solution IV; Merck, Germany) was used for the preparation of intermediate solutions to obtain calibration curves. Deionized water (Direct Q UV, Millipore, approximately 18.0 MΩ; Analytica Jena Instruments, Thuringia, Germany) was used in the preparation of all solutions.
The analytical methods were verified for quality control of performance parameters: Limit of Detection (LOD), Limit of Quantification (LOQ), linearity of calibration curves.
The metal content was determined using GTAAS and the performance parameters of the analytical method are presented in Table 1

Statistical Analysis
The R package "corrplot" was used to compute the pair-wise Pearson correlation between the analyzed samples. The correlogram of each pair of variables is displayed on the lower triangle in Figure 1. The strength of the linear dependence between variables is visualized in Figure 1, upper triangle: a correlation coefficient between 0.5 and 1 (or −0.5 and −1) indicates a strong positive (negative) linear dependence, while a moderate positive (negative) dependence corresponds to the values of the correlation coefficient between 0.3 and 0.5 (or −0.3 to −0.5). The statistical significance of each correlation coefficient was tested using Student's t-test with n − 2 degrees of freedom, where n is the size of the sample. The null hypothesis of the test states that the correlation coefficient between two pairs of variables is closer to zero, while the alternative hypothesis affirms that the linear dependence is statistically different from zero. If the p-value associated with the Student's t-test is lower than the significance level 0.05, we conclude that the correlation is statistically significant [25].

Results
The results are presented in Tables 2-6. The organoleptic parameters of the analyzed samples are presented in Table 2.

Results
The results are presented in Tables 2-6. The organoleptic parameters of the analyzed samples are presented in Table 2. The water content (w c ), pH, acidity, conductivity and SSC of the analyzed honey samples are presented in Table 3. The results for protein content are presented in Table 4. Metal concentrations obtained for studied samples are presented in Table 5. The results obtained for analyzed honey samples were compared with those of other scientists (Table 6). ---0.03-0.13 0.14-1.53 Italy, mg/kg [28] 0.86-0.99 -0.0-0.24 0.05-0.14 -Romania, mg/kg [29] 0.76-3.41 0.05-3.81 --2.00-33.01 The correlation coefficient for the studied physico-chemical parameters and metal levels in honey samples are presented in Figure 1.

Discussion
The bee honey analyzed had the right color, smell, taste and consistency according to the requirements of the specialized literature [20].
The water content (w c ) of honey should not exceed 20%, and superior honey should not have a content higher than 17-18% [30]. The water content of the investigated honey samples ranged from 6.43 to 10.36%. Similar results were obtained in previous studies in Brazil [19]. Moisture level in honey is a parameter that determines the capability of honey to remain stable. If the water content is above the allowed limit, there is a risk of mold and fungus proliferation and honey fermentation [31].
The pH and free acidity are of particular importance during the extraction and storage steps, as they can influence the texture and stability of honey [3]. The pH values of investigated honey samples varied from 4.10 to 5.0; these values being within the acceptable range for honey [32]. The organic honey showed slightly lower pH values. The variations in the pH of the investigated honey are due to the variation in the content of different acids in the honey [33]. Floral differences can also contribute to the variability of pH values. The pH is an important characteristic of bee honey because it can inhibit and limit the growth of microorganisms, and it contributes to the stability of the honey [3].
The free acidity values of the honey samples were between 18.8 and 40.2 meq/kg, being within the permitted range and below the limits established by EU legislation (50 meq/kg) [30].
It has been reported that the acidity of honey may be due to the presence of acids such as aromatic, aliphatic acids and especially gluconic acid, which is considered the main organic acid responsible for the acidity of honey, especially during ripening season. [34]. It can be concluded that the season and the species of plant can influence the acidic aspect of the honey [35].
In our work, honey from beekeepers had the highest acidity, especially acacia honey of 40.2 meq/kg. This source of honey probably contains larger quantities of the compounds responsible for the acidity of honey. The results of this study were comparable to the acidity values reported by Gebru et al. [36] and Adisu and Malede [37], indicating the freshness of the honey samples.
Electrical conductivity is considered as one of the parameters used to differentiate honey of different floral origins [38]. In honey, the electrical conductivity can be influenced by the botanical origin of the nectar, as well as by other inorganic or organic molecules that can act as electrolytes [39]. Directive 2001/110/EC of the European Union established the maximum level of electrical conductivity allowed for nectar honey at 0.8 mS cm −1 [30].
Yadata et al. [40] mentioned that high conductivity values indicate high mineral content, which is an important nutritional property of honey as a source of minerals.
In this study, the electrical conductivity varied between 296 in acacia honey and 655 µS cm −1 in polyfloral honey. The lower average value of 324 µS cm −1 was found in the supermarket honey. Variability in the values of this parameter was observed within the same type of honey: in acacia, the values varied from 296 to 387 µS cm −1 , while in polyfloral honey, the values varied between 278 and 655 µS cm −1 . The differences in EC values of the investigated samples may be due to the differences in the quantity of minerals and organic acids, as well as the variability of the floral origin of the honey. Similar values were found by Albu et al. [41] and Scripca et al. [42] in acacia and polyfloral honey samples.
The soluble solids' content is represented by total soluble sugars, which are the most abundant soluble solid substances in honey solutions [43] and expressed in Brix degrees ( • Brix). It has been reported that the • Brix value is a measure of the sugar content of honey samples [44]. In the present work, the • Brix values indicated that the samples from the beekeepers have a soluble solids content (SSC) between 76.5% and 81%, in accordance with the values that are prescribed by the EU Council [30]. Lower SSC values were obtained for the supermarket and organic honey (69-70%), which can indicate that these honey samples have been subjected to slight degree of further processing. It was presented that invertase converts sucrose from pure honey into fructose and glucose [45]. If the honey is processed, the heat can destroy the invertase in the honey, reducing the fructose in the honey [46].
The protein content of analyzed bee honey ranged from 0.0437 to 0.2627%, higher values being observed in the honey from beekeepers. The differences observed between the values can be associated with the presence of proteins derived from flower nectar or enzymes introduced by bees in the products. Habib et al. (2014) noticed that the protein content of honey depends on the floral source [33].
The concentration of metals in investigated honey samples varied in the order of Cu > Cr > Pb > Fe > Ni > Mn > Co > Cd ( Table 5). The values obtained from the analysis of bee honey were within the limits established by EU Commission [47], indicating their quality.
Cadmium was detected in honey samples, ranging between 0.019-0.52 µg/Kg, the highest being in the honey from beekeepers and the lowest in the supermarket honey.
The level of Cd in the studied samples, did not exceed the limit established for this metal in honey [47]. Cd contamination usually has an anthropogenic source such as placing hives near high traffic areas, using fertilizers or using processing equipment with a higher Cd content [48].
Copper is known to be an anti-infective, dynamic, antiviral and anti-inflammatory mineral. Being an indispensable element for life, the body has the ability to immobilize copper in case of microbial aggression. Cu is incorporated into enzymes that help regulate iron metabolism. From the results of Table 5, it can be seen that the acacia honey from the supermarket has the highest copper content (24.67 µg/Kg-a value comparable to the concentration of copper in the honey obtained from clover [14]), and the lowest content is observed in organic acacia honey (below LOD).
Chromium concentrations were detected from 2.94 µg/Kg (in organic acacia honey) to 12.75 µg/Kg (in acacia honey from the supermarket). Contamination of honey with Cr can occur during harvesting (stainless steel utensils) or during the preparation of honey for the market, due to the corrosive effect of honey acidity [49].
Manganese can intervene in the composition of honey through the soil, being a natural component of it. As manganese is found in agricultural fertilizers, these can be considered the most important source produced by humans [16]. The Mn values obtained in the investigated honey samples are higher than those reported in Argentinian honey samples [5] and those reported in Turkish honey [17].
Cobalt was found in the analyzed honey in the range of 0.05-0.854 µg/Kg, even though it is considered a very important mineral element for human health. Nickel concentrations were detected in all studied samples in the concentration range of 0.95-3.66 µg/Kg; the highest value was found in acacia honey from the supermarket.
Following the analysis of lead in the honey samples, concentrations varied from 1.7 to 5.28 µg/Kg; lower Pb values were observed in organic honey. Contamination of honey with this metal can be related to environmental pollution, such as heavy traffic, metallurgical activities and the use of fertilizers for crop production.
In this study, Fe was not detected in organic polyfloral honey, while in the rest of the samples it varied from 1.71 to 4.38 µg/Kg. The variation observed in the investigated samples can be influenced by the availability of iron in botanical sources of floral plants.
The results obtained in this study were compared with those reported by other authors (Table 6) [26,27]. Metal concentrations investigated in honey samples were generally lower than those reported by some authors [28]. Several factors may cause these differences such as: difference in honey type, soil composition, amount of precipitation, air temperature, plant species from which honey originates, harvesting period and degree of pollution in the regions of origin [29]. It has been reported that honey can have both the environment where bees collect honey and the techniques applied by beekeepers as sources of metal contamination [50]. The bee collection area can include various environments, plants and food sources. When collecting nectar and pollen from an area of interest, bees come into contact with plants, air, water and soil. It has been established that high concentrations of metals in honey come from the soil; they are transported through the root system to the honeydew plants, enter the nectar and then the honey produced by the bees [51]. Figure 1 shows the pair-wise Pearson's correlation coefficient between w c , pH, EC, • Brix, Pb, Cd, Cr, Ni, Mn, Co, Cu and Fe. Based on the samples, w c is weakly positively correlated with Pb, Cr and Co and strongly negatively correlated with Cd and Ni. Moreover, moderate positive dependence was found between pH and Pb and Ni, respectively, while stronger positive dependence was detected between pH and Mn and Fe, respectively. Furthermore, EC is strongly negatively correlated with Ni and Fe, while EC is weakly correlated with the other metals. These correlations show that in the investigated honey samples, organic acids predominate in a larger amount compared to mineral acids. Lastly, • Brix is strongly positively correlated with Cd and Ni, and has a moderate negative correlation with Cu. The relative weak correlation between the physico-chemical parameters and some of the metals, i.e., Mn, Cu and Fe, could be caused by the fact that these metals were not detected in all the samples. To decide which of the Pearson's correlation coefficients in the table are statistically significant, we used Student's t-test with 4 degrees of freedom. The null hypothesis states that there is no significant linear correlation between two elements, while the alternative indicates the existence of a significant correlation between them. The reported p-values indicate that a statistically significant positive correlation was only detected between • Brix-Cd, Ni-Cd and Cu-Cr, respectively. The Pearson correlation coefficients for these pairs are above the significant level 0.5, which suggests strong linear dependences between the pairs of elements. Moreover, a statistically significant negative correlation was obtained between w c and Cd.

Conclusions
Bee honey is considered a very valuable product. The physical-chemical indicators of the studied bee honey samples collected from different sources (beekeeper, local market and organic honey) fall within the values allowed in European quality regulations.
The concentration of metals in the investigated honey samples varied in the order Cu > Cr > Pb > Fe > Ni > Mn > Co > Cd, the values being generally lower than those reported by other researchers and within the limits established by the EU Commission. The differences in metal concentrations depends on different factors such as: geographical factors, soil composition, plant species from which honey originates, harvesting period and degree of pollution in regions.
The Pearson correlations were computed between physical-chemical parameters and the metals. The analysis indicates that strong positive and statistically significant linear dependences were found between Brix-Cd, Ni-Cd and Cu-Cr, using the Pearson correlation implemented in the R package "corrplot". Moreover, a strong negative and statistically significant linear dependence was obtained between w c and Cd.