Analysis of sugar crystal size in honey

Honey consists typically of more than 80% sugars, predominantly fructose and glucose. Glucose-rich honey crystallizes more rapidly than honey with a high fructose content. However, the size of the sugar crystals is crucial for the mouth feel of crystallised honey. Honeys containing small crystals have a creamy consistency, which is preferred by most consumers. In contrast, large crystals cause a coarse mouth feel. Factors affecting crystal size are of vital interest for the production of high-quality honey and thus analysis of sugar crystal size in honey is crucial. Here we present a simple and efficient method for measuring the size of sugar crystals in honey. A honey drop is placed on a coverslip, which is centrifuged using a converted smoothie maker. This spreads the drop over the coverslip and separates the sugar crystals from each other. Subsequently, the size of the crystals can be conveniently measured by microscopy. Compared to squeezing the honey drop between slide and coverslip, this approach avoids the risk of breaking the crystals. Moreover, the method is highly reproducible as indicated by intra-day and inter-day standard deviations of 7 to 14% for crystal sizes. Simple method for preparation of honey for crystal size analysis by microscopy. Use of cheap, easily accessible equipment. High intra and inter-day reproducibility.


Specifications table
Subject area: Agricultural and Biological Sciences More specific subject area: Crystal size analysis Method name: Analysis of sugar crystal size in honey Name and reference of original method: Costa et al. [2] Brazilian Journal of Food Technology 18, 155-161. DOI: 10.1590/1981-6723. 7314 Dettori et al. [5] LWT -Food Science and Technology 95, 333-338. DOI: 10

Method details
Background Honey consists typically of more than 80% (w/w) sugars and less than 20% (w/w) water. Other compounds are crucial for colour and taste but are present in total at less than 1% (w/w). Amongst the sugars, the monosaccharides fructose and glucose are present in high quantities of typically more than 30% (w/w) while other sugars including the oligosaccharides maltose, isomaltose, turanose, trehalose and raffinose are only present in small quantities [6] . The composition of the sugars, particularly the ratio of fructose and glucose, depends mainly on the forage flowers of bees. Consequently, the fructose to glucose ratio can vary in a wide range from approximately 0.4 to 2.5 [3 , 7 , 8] . Fructose-rich honey remains usually liquid for several months or years. In contrast, crystallisation of glucose-rich honey often begins shortly after harvesting, as glucose is significantly less soluble in water than fructose; solubility of glucose: 47 g/100 g water [14] ; solubility of fructose: 385 g/100 g water [15] . Importantly, the crystallisation conditions determine the size of the sugar crystals [4 , 11 , 13] , which can reach a size of more than 200 μm in length [11] . Most consumers prefer honey with small crystals since it has a creamy mouth feel [12] while honey with large crystals is sensed as coarse grained. Measures to influence the creaminess of honey are therefore intensively investigated [2 , 5 , 9 , 11 , 12] . This, however, requires simple and reliable methods for crystal size determination in honey. Currently, the crystal size is mainly measured by microscopic analysis [2 , 5 , 10] and, in a few cases, by laser diffraction [2] . In principle, microscopic analysis of crystal size is straightforward and simple. However, due to its high sugar content honey is highly viscous, which complicates uniform spreading of the crystals on a microscopic slide. Simple spread out, although frequently used [5 , 10] , is only successful with honey samples containing a few crystals per gram. In order to disperse more crystallised samples strong forces would be required, which would break particularly large crystals and thereby prevent accurate crystal size analysis.
Here we describe a simple method for determining the crystal size in honey by microscopy. A drop of honey is placed on a coverslip, covered with polyethylene glycol 200 and centrifuged using a converted smoothie maker. This spreads the honey sample over of the coverslip and removes the polyethylene glycol 200. Subsequently, the coverslip is placed upside up on a microscopic slide and the crystal size is determined by microscopic analysis. This method has the advantage that the crystals are well distributed over the slide without application of high forces and the thickness of the honey film is thin enough preventing overlapping of the crystals.

Instrumentation
A smoothie-maker (AEG MiniMixer SB 2700, Electrolux, Stockholm, Sweden, 23,0 0 0 rpm) was converted to a coverslip centrifuge by cutting four 10 mm deep threads into an aluminium cylinder of 48 mm in diameter and 23 mm width to accept M3 metal screws ( Fig. 1 ). A whole with a diameter of 25 mm and a depth of 8 mm depth was milled on the underside of the cylinder to accommodate the wheel of the smoothie maker. The aluminium cylinder was fixed to the wheel of the smoothie maker using two-component adhesive ( Fig. 2 ). Using the screws, a coverslip can be clamped to the cylinder for centrifugation.
For microscopic analysis a microscope with at least 250x magnification and equipped with a camera should be used. In this study a VHX-70 0 0N digital microscope (Keyence, Osaka, Japan) with 3D modus was used. Note 1: Sample preparation is depictured in Supplementary Fig. 1 and a video Fig. 2a ). In contrast, excessive centrifugation leads to crystal conglomeration ( Supplementary Fig. 2c and 2d ).

Method validation
The method described above allows simple and reproducible spreading of honey samples over the coverslip ( Fig. 3 ).
In contrast to spreading samples between slides and coverslips, centrifugation is a gentle procedure for distributing crystals over the coverslip. However, centrifugation may pose the risk of sample loss due to centrifugal forces. To investigate potential sample loss, the smoothie maker was thoroughly cleaned to prevent presence of any traces of sugars inside of the centrifugation chamber from previous experiments. Subsequent the interior surface was rubbed off with a wetted cotton swab, which was subsequently placed in 3 ml distilled water for 10 min to dissolve any sugar crystals that might have been collected by the cotton swab. Analysis by ion chromatography with pulsed amperometric detection ( Fig. 4 A and 4 B) confirmed absence of any fructose. Next, a spring flower honey sample was centrifuged as described above. The spring flower honey had a fructose content of 37.9 ±0.4% (w/w) and a glucose content of 36.3 ±0.7% (w/w) as determined by HPLC using a Nucleosil 100-5 NH2 250 × 4 mm column and 80% acetonitrile as eluent [1] . After centrifugation, the interior of the smoothie maker was again rubbed off with a wetted cotton swab, which was placed in 3 ml distilled water for 10 min prior analysis. As shown in Fig. 4 D, no fructose indicative for sample loss could be detected. This experiment was repeated six times and in all measurements fructose was below the limit of detection. The ion chromatography system used had a sensitivity of 0.05 mg/l (limit of detection) for fructose, which would even allow detection of a single fructose particle with a length of 250 μm and a width of 20 μm dissolved in 3 ml water. In contrast to fructose, glucose could not be measured using this system because compounds leached from the cotton swab interfered with the glucose peak ( Fig. 4 A and 4 C). Nevertheless, this result clearly demonstrates that there is no detectable sample loss during centrifugation.
Having established that the method is suitable for uniform spreading of sugar crystals present in honey samples on a coverslip, we investigated the intra-and inter-day reproducibility of the methods. For that purpose, the same honey sample as used above was analysed in quadruplicate (each time with three technical repeats) on four consecutive days ( Table 1 ). These data confirm that the developed method is highly reproducible with inter-day and intra-day relative standard deviations (RSD) in the range of approximately 7-14%.

Conclusion
Preparation of honey for microscopic analysis can be challenging due to its high viscosity. Here we show that centrifugation using an adapted smoothie maker is a simple, rapid and gentle method for spreading honey over a cover slip. Subsequently, crystal size is determined by microscopic analysis. The method shows high reproducibility with an intra-day RSD of 7-14% and an inter-day RSD of 12%.