Due to the low water content of hard candies, Karl-Fischer (KF) titration method was used for determination of water content (Fischer, 1935). The reaction occuring during titration is described by Equations 1, 2 as follows: R − OH + SO 2 + R ′ N   →   R ′ NH SO 3 R + H 2 O + I 2 + 2 R ′ N (1) R ′ NH SO 3 R + H 2 O + I 2 + 2 R ′ N   → 2 R ′ NH I + R ′ NH S O 4 R (2)

By the help of the electrical potential difference occurred in between I₂/I⁻ couple, the water content can be estimated stoichiometrically. As water that is in the titration chamber depletes, electrical potential gradient goes to a constant value. When the rate of change drops under the threshold determined as 50 μS, the reaction is ended, and water amount is calculated. The analysis was made by Karl Fischer Titrator (TitraLab KF1000 Series, HACH, UK) at 25°C with a 2 component system with three replicates.

In addition to water content determination, a_w of the samples were measured by using water activity analyzer (LabStart–a_w, Novasina, Switzerland). The principle of the measurement is based on the equilibrium relative humidity (ERH) method (Subranamiam and Wareing, 2016).

An SR sequence having a relaxation period of 10 s and a delay time between 5–1,500 ms for 16 points was used with 4 scans to measure T₁.

M₂ values were obtained by MSE sequence (see Figure 1) having 10 s repetition delay and 4 scans per each step of the cycling of phases ϕ₁, ϕ₂ and ϕ₃ as described in a previous study (Grunin et al., 2019). The number of points that were used for FID was 512.

Understanding the a_w and TSS properties of hard candy formulations is essential since these properties determine the quality and shelf life of the products (Ergun et al., 2010). As shown in Table 5, the highest a_w belongs to 100_S (0.80) and addition of any doctoring agent significantly decreased (p ≤ 0.05) the a_w of the candy formulations. Additionally, GS–sucrose formulations had higher a_w than the FS–sucrose ones (p ≤ 0.05). These findings are related to the sugar composition and the solubility characteristics of the individual sugar molecules. As the sucrose content was replaced by glucose-fructose coming from either FS or GS, humectant capacity of candies increased and water–sugar interactions enhanced (Hartel et al., 2011). Since fructose is the most water-soluble sugar, high fructose content of FS paved the way for the lower a_w of FS–sucrose formulations (Ergun et al., 2010). In a supportive way, the Pearson correlation coefficient (r) analysis showed a strong negative correlation (r = - 0.89) between the a_w and TSS of FS–sucrose formulations. GS, on the other hand, contains very low levels of fructose but includes high amounts of glucose, maltose and high MW oligosaccharides as shown in Table 4. Generally, the a_w of hard candies is below 0.60 which is not suitable for microbial growth. The a_w of hard candies that contain a doctoring agent in their formulations varied between 0.35 and 0.48. Thus, there is a very low microbial spoilage risk for these samples during storage. In contrast, the 100_S sample has a_w of 0.80 which makes it susceptible to yeast and mold growth (Ergun et al., 2010). Thus, addition of a doctoring agent reduced the microbial risk of the sucrose candies. However, low a_w could be a problem if the RH of the surrounding air is higher than that of the candy (Nowakowski and Hartel, 2002). Therefore, FS or GS containing sucrose hard candies must be stored carefully. The high a_w of 100_S is also a problem in terms of sucrose recrystallization during storage. At such a high a_w level, sucrose molecules will have more mobility to interact with each other and initiate crystallization leading to a grainy texture (Netramai et al., 2018). GS–sucrose hard candies have also a high moisture induced graining risk during storage. The high MW sugars and maltose in GS may form a skin at the candy surface. When this skin absorbs moisture from the environment, moisture is mostly retained at the surface and extensive sucrose crystallization may be observed (Hartel, 2002). On the other hand, FS candies can show more humectancy and sugars interact more intensely with the water molecules leaving less free water to initiate graining. Nevertheless, these samples may show stickiness due to their hygroscopic character (Labuza and Labuza, 2004).

Solubility characteristics of the sugars in the mixtures also affected the TSS of the hard candies (Zumbé et al., 2001). According to Table 5, addition of a doctoring agent increased the TSS of the candy formulations (p ≤ 0.05). The increase in TSS was more evident with GS addition. Except for the 75:25_GS formulation, GS–sucrose candies maintained a higher TSS profile than FS–sucrose candies (p ≤ 0.05). The main reason was the high total glucose-fructose content of FS–sucrose candies. Solubility of sucrose decreased when the total glucose-fructose content was increased in the system (Bund and Hartel, 2010). As the sucrose content was replaced by corn syrup, sugar molecules competed with each other to interact with water and the supersaturation driving force for sucrose decreased (Hartel et al., 2011). Clearly, glucose and fructose preferentially became soluble in FS–sucrose formulations and sucrose solubility decreased which ended up with lower TSS for FS containing candies. Since GS containing hard candies have much lower total glucose-fructose content, sucrose readily dissolved in these formulations and contributed to the TSS. However, there was an unexpected steep decrease in TSS of 75:25_GS samples. Increasing the GS concentration up to 75% (w/w) in GS–sucrose formulations resulted in the lowest TSS among all formulations (p ≤ 0.05). The reasons behind this behavior at this formulation could be the increase in both high MW oligosaccharide and glucose-fructose composition of GS–sucrose candies with respect to the other candy formulations having lower GS content (Šmídová et al., 2003). Despite its lower glucose-fructose content with respect to its FS counterparts, GS candies reached the highest total glucose-fructose proportion ( ∼ 21% w/w) at this specific formulation (75:25_GS). Consequently, sucrose solubility may have been restricted for the first time in GS–sucrose formulations. In addition to the lower levels of dissolved sucrose, total concentration of the high MW oligosaccharides also reached its top level in this formulation. In the presence of higher glucose-fructose concentration that shows high solubility, long saccharide chains may have also experienced a lower solubility. Therefore, it is believed that the combined effect of decreased sucrose and high MW sugar solubilities resulted in lower TSS for 75:25_GS candies (Levenson and Hartel, 2005).

Another factor that needs to be considered is the T_g values of the hard candies. 100_S candies showed the highest T_g as 53.32°C whereas addition of FS or GS produced lower T_g values as demonstrated in Table 5. Without any doctoring agent, sucrose molecules were able to interact with each other to form a crystal lattice (McFetridge et al., 2004). As a result, 100_S candies reached the glassy state during cooling at a higher temperature with respect to other samples. On the other hand, presence of doctoring agents, diminished the degree of interactions between sucrose molecules in the rubbery state in which molecular mobility is sufficient enough to form crystal structures (Hartel et al., 2011). Moreover, the T_g values of GS–sucrose candies were higher than that of FS–sucrose candies mainly due to the presence of higher amounts of maltose and high MW sugars (Ergun et al., 2010). These sugars have high T_g values individually contributing to the increased T_g level of GS containing samples. Contrarily, FS candies have more fructose and glucose concentration having much lower individual T_g values (5°C and 31°C) compared with the T_g of maltose and high MW oligosaccharides (87 and 112°C–180°C) (Liang et al., 2007) (Table 2). The plasticizing effect of water on T_g was apparently observed on FS–sucrose samples since increasing FS concentration (from 25:75_FS to 75:25_FS) increased the water content from 5.83% to 7.08% (w/w) and decreased T_g from 25.61°C to 4.03°C. Although, GS addition to sucrose candies also decreased T_g, there is not a clear trend between GS concentration and T_g values. The reason could be the distortion of the ideal volume mixing behavior of the GS–sucrose systems (Reinheimer et al., 2010). Generally, when a low T_g component is mixed with a high T_g component, the T_g of the system attains a level between the individual T_g values of the mixed components. However, mixing of two high T_g components may not show an ideal mixing behavior (Šmídová et al., 2003). Since T_g values of GS components are higher than those of FS components, improper mixing may have occurred in GS containing formulations. Moreover, addition of high MW compounds to another component induces a broader glass transition range that would lead to a high variability in calculated T_g values (Kawai et al., 2019). This could be the case in GS–sucrose hard candies that revealed variable T_g values at different GS concentrations.

The T_g results showed that the doctoring agents (FS and GS) clearly decreased sucrose crystallization taking place in the rubbery state during the cooling period. It should also be noted that, T_g values of the FS–sucrose candies remained below the room temperature except for the 25:75_FS sample (Table 5). This is quite risky since such T_g values may lead to temperature induced internal graining during storage (Liang et al., 2007). GS samples, on the other hand, possess higher levels of T_g reducing the risk of temperature induced sucrose crystallization (Levenson and Hartel, 2005). However, GS - sucrose candies are more prone to moisture induced surface graining than FS - sucrose candies as previously discussed when explaining a_w results. Although GS candies would absorb less moisture with respect to FS candies at high RH storage conditions due to their higher a_w, the free water population would be higher in GS candies in the absence of a sufficient humectant content (Hartel et al., 2011). FS candies would suffer less from the moisture induced sucrose crystallization but they may become stickier than GS samples due their increased level of moisture absorption.

T₁ analysis of hard candies showed a monoexponential relaxation behavior except for 100_S candies showing biexponential relaxation (Table 6). The monophasic relaxation of FS or GS added candies indicated that components coming from doctoring agents had enhanced affinity with water and provided a more homogeneous sample (Okada et al., 2019). Additionally, doctoring agents are also able to place themselves between the sucrose molecules and this may have also contributed to the homogeneity of the continuous phase in the hard candies. The presence of solely sucrose probably induced more interactions between the individual sucrose molecules creating distinct proton populations within the samples (Mariette, 2009). The relaxation peaks of 100_S candies exerted a very long T₁ (1,627 ms) and a much shorter ( ∼ 78 ms) components supporting the previous claim. Probably, intense interactions between the sucrose molecules created a non-homogeneous water and sucrose distribution within the samples and locally gathered ordered crystalline sucrose fractions contributed to the long T₁ component (Aso et al., 2007). Besides, T₁ values of hard candies containing doctoring agents showed some differences since spin-lattice relaxation behavior is sensitive to the alterations in the quantity and order of crystalline state (Adam-Berret et al., 2008). Generally, longer T₁ values are observed with compact and highly ordered crystalline state arrangements at high solid concentrations (Adam-Berret et al., 2009). 100_S samples having the highest T_g also attained the longest T₁ which was compatible with the previous claim that related the long T₁ with the higher order of the crystalline state. The addition of doctoring agents extremely lowered the T₁ values in consistence with their sucrose crystallization retarding effect.

In addition to T₁, M₂ is also a valuable parameter to understand crystalline state molecular dynamics with more precision as proton mobility within a crystal lattice determines the M₂ of a system. A higher M₂ value is associated with a lower proton mobility in a solid state (Van Duynhoven et al., 2002). The reason behind this behavior is the enhanced dipolar interactions between the protons in lower molecular mobility conditions (Wang et al., 2002). Therefore, higher M₂ values are expected at higher degrees of crystallinity (Grunin et al., 2019). The data in Table 6 indicates this trend as the 100_S candies had the highest M₂ (10.96) indicating that 100_S samples possessed the highest crystallinity level (p ≤ 0.05). Similar to the T₁ results, incorporation of doctoring agents also reduced the M₂, thus the degree of crystallinity of the sucrose hard candies (p ≤ 0.05). Addition of 25% (w/w) doctoring agent (GS or FS) decreased the degree of crystallinity and attained similar M₂ levels for both samples (25:75_FS and 25:75_GS). However, when the added syrup content increased to 50% (w/w), GS–sucrose candies demonstrated a higher crystallinity (M₂) than FS–sucrose candies (p ≤ 0.05). This means 50:50_GS candies experienced more sucrose crystallization than 50:50_FS candies. However, 50:50_GS candies have a higher T_g than the T_g of 50:50_FS candies (p ≤ 0.05). Normally, a higher T_g corresponds to an earlier entrance to the glassy state in which molecular mobility is severely restricted so that almost no sucrose crystallization can take place. A high T_g value is also an indicator of shorter time spent in the rubbery phase where sucrose crystallization is possible during cooling (Roos, 2002). Nevertheless, 50:50_FS candies with much lower T_g showed less crystallinity. Here, the high total glucose-fructose content ( ∼ 35% w/w) of FS predominated the system and retarded the interactions between the sucrose molecules even at the rubbery state during candy production. Glucose and fructose molecules are able to interact intensively with sucrose molecules via interactions such as hydrogen bonding (Spanemberg et al., 2022). Therefore, sucrose nucleation rate was diminished, and crystal lattice formation was restricted in 50:50_FS samples. On the other hand, the low total glucose-fructose concentration ( ∼ 13.5% w/w) of 50:50_GS samples could not retard the sucrose crystallization taking place in the rubbery state before entering the glassy state, as efficiently as the F50:50_FS samples. In this way, even the time spent in the rubbery state was short, the amount of sucrose crystals formed in this state was high. When the added syrup concentration increased to 75% (w/w), this trend has changed. 75:25_GS candies showed a very low degree of crystallinity whereas 75:25_FS candies attained higher crystallinity levels (Table 6). Although the total glucose-fructose content of GS–sucrose candies was still much lower than the FS–sucrose candies at this concentration, the increased level of high MW sugars clearly predominated the GS candy system. One of the effects of high concentrations of high MW oligosaccharides is the increased system viscosity (Raudonus et al., 2000). The increased viscosity may have compensated the effect of low glucose-fructose content by decreasing the sucrose mobility effectively in the rubbery state (Hartel, 2002). Thus, formation of a crystal lattice by sucrose molecules may have been retarded. Based on these results, it is possible to claim that the type and concentration of syrups added to sucrose candies have distinct effects on the sucrose crystallization properties of hard candy formulations.

In fact, M₂ analysis provided a very important information on the crystallization pattern in hard candies. T_g of the 100_S and 75:25_GS candies are high and close to each other as 53.32°C and 45.23°C, respectively. Therefore, one can assume that both candy formulations may have similar crystallinity properties since both samples reached the metastable glassy state at similar temperature regions during production. However, the crystallization processes taking place in these samples were completely opposite to each other. While 100_S candies experienced the highest sucrose crystallization level, 75:25_GS candies showed much less crystal lattice formation. Consequently, M₂ analysis provided a detailed understanding of the glass transition and crystallization properties of hard candies. Finally, correlations between the NMR and other measured parameters could be mentioned. Strong positive correlations between T₁ - T_g (r = 0.89), M₂ - T_g (r = 0.90) and T₁ - a_w (r = 0.99) have been detected for the FS candies. The only correlation that has been observed for GS candies belonged to the T₁ - a_w with an r of 0.99.