Temperatures, TSS and pH are the important parameters, which affect the rheological properties of Roselle leaves puree.

Roselle leaves were purchased from the local market of Raipur (Chhattisgarh, India), washed thoroughly with running water, and then destemmed. Leaves were blanched at 363 K for 5 min. Efficacy and duration of blanching were determined by peroxidase inactivity test. Chilled leaves after removing excess amount of water were grounded using a mixer grounder (Bajaj, India) at 298 K for 15 min with the addition of distilled water. The grounded puree was filtered through filter paper (Whatman, 125 mm). Uniform consistency of the puree was maintained by passing through a 12-mesh screen. The lab size rotary evaporator (Cole-Parmer, India) prepared the puree concentrates of different TSS at 323 K and 160 mm Hg pressure. Samples of the desired range of TSS (3, 4.35, 5 °Brix) were obtained by diluting with appropriate amounts of distilled water.

Moisture content was determined as per Garcia-Segovia et al. (2010) at 333 K on a wet basis. Ash content was calculated as per the standard method of AOAC (1984). Chlorophyll content was analyzed using the method described by Arnon (1949). The ascorbic acid content of the puree was found out by titration method (Ranganna, 1986). Titratable acidity was calculated in terms of citric acid equivalent as per the process proposed by Wang et al. (1995). TSS (°Brix) and pH was measured using digital refractometer (Atago, Japan) and pH meter (Handy) respectively. The acidic pH (1.3 – 5.3) of purees was maintained by using a buffer for physicochemical and rheological analysis.

The spectra of puree are used for analysis by FTIR instrument (Bruker, Germany) in the frequency ranging from 500 cm^-1 to 4000 cm^-1. The functional groups present in the puree are identified using the spectral data obtained with the reference.

The rheological test was conducted in a modular compact Rheometer MCR 102 (Anton Paar, Germany), mounted with ST22-4V-40 (four-bladed vane geometry). The temperature of the cup was varied from 278 K – 318 K, for various experimental runs. The experiment was performed within shear rate (1 – 100 s^-1). The rheological estimations were investigated utilizing the Rheoplus software package of Anton Paar.

Numerous factors influence the selection of a rheological model to describe the flow behavior of a particular fluid. In this work, the experimental values of shear rate and shear stress were fitted to the various non-Newtonian models as listed in Eq. (1 – 6).

Power Law: (1) τ=K ( γ ) n Herschel-Buckley: (2) τ 0.5 = τ 0 +K ( γ ) n Casson : (3) τ 1/2 = τ 0 1/2 +K γ 1/2 Heinz-Casson:(4) τ n = τ 0 +K γ n Vocadlo: (5) τ 1/n = τ 0 1/n +Kγ Mizrahi- Berk: (6) τ 1/2 = τ 0 1/2 +K γ n

where, τ,γ,K,n, τ 0 is shear Stress (Pa), shear rate (s^-1), consistency index (Pa.s^-1), behavior index (dimensionless), yield stress (Pa), respectively.

The selection of most suitable model among these, for prediction of the flow behavior of puree, was investigated taking account of the statistical parameters, root means square error (RMSE) (Eq. (7)) and coefficient of determination (R²) (Eq. (8)). The fitted data, which showed the lowest values for RMSE and highest values for the R², was considered as the best-fit model. ANOVA and t-test were used to determine the mean differences. The significant difference was defined at p <0.05 with 95% level of confidence.

The average value of all the experimental data was analyzed using statistical parameters including Root mean square error (RMSE) and coefficient of determination (R²) were determined and were defined as,

(7) R M S E = ∑ i = 1 N ( E V − M V ) 2 N (8) R 2 = 1 − ∑ i = 1 N ( M V − E V ) 2 ∑ i = 1 N ( E V − E V a v g ) 2

Where EV is the experimental value, MV is the modeled value EV_avg is the average of the experimental value and N is the number of observations.

Temperature dependency on its viscosity was determined using Arrhenius model (Eq. (9)) (Sengül et al. 2005).

(9) η = η 0 ⋅ exp ( E a R T )

Where, η 0 is the proportionality constant, E_a the activation energy (kJ.mol^-1), R the universal law gas constant (J.mol^-1 K^-1), and T the absolute temperature (K).

Two models, namely, the power law Eq. (10) and exponential model Eq. (11) were used to describe the variation of the activation energy with the TSS, as given below.

(10) E a = a ( C ) b (11) E a = a exp  ( bC )

Where, a (Empirical constant (kJ.mol^-1)), C (TSS (°Brix)) and b (Constant).

The combined effect of temperature and TSS on apparent viscosity was evaluated by Power law type Eq. (12), firstorder exponential model Eq. (13) and second-order exponential model Eq. (14) (Juszczak and Fortuna 2004; Manjunatha et al. 2012). The estimates and fits of the different parameters were found at 5% significant level (p <0.05). The suitability of the model was choose based on RMSE values and correlation coefficient (r).

(12) η = a ( C ) C E x p ( E a R T ) (13) η = a E x p ( E a R T + c C ) (14) η = a E x p ( E a R T + c C + d C 2 )

Where, η, E a ,R,C,T is apparent viscosity (Pa.s^-1) at shear rates (10, 50 and 100 s^-1), flow activation energy (kJ.mol^-1), gas constant, TSS (°Brix), absolute temperature (K), respectively, a is pre-exponential constant (mPa.s^-1), b=Ea/R, c is constant (°Brix¹) and d is constant (°Brix²).

The effect of temperature on flow behavior index (n) and consistency coefficient (K) was evaluated using a modified Turian approach through a regression analysis (Turian, 1964) (Eqs. (15) and (16)):

(15) log k = log k 0 − A 1 T (16) n = n 0 + A 2 T

Where A₁ and A₂ are the slopes for Turian models. Higher A₁ and A₂ represent more dependency of n and k to the temperature.

Changes in viscosities of puree with TSS (3 °Brix) having different pH value were assessed. The pH range was varied from 1.3 to 5.3 (increment of 2 was adjusted using 0.2 M Na₂HPO₄ and 0.1 M citric acid) at shear rates (10, 50 and 100 s^-1) and fixed temperature of 303 K.

The physicochemical characteristics of purees are shown in Table 1. The purees having 3 °Brix shows the high moisture content, due to the presence of higher concentration of water. Purees having higher TSS showed the highest value of acidity, compared to that of lower TSS value, probably due to the presence of the highest amount of total organic acid (like ascorbic acid) in the leaf (57.2 mg.100g^-1).

The amounts of ascorbic acid reported in the present work were found higher than the ones previously reported in the literature (Nnam and Onyeke, 2003).Table. 2

The differences observed might be due to different varieties, harvest conditions, genetics, environment, and ecology. The different TSS of purees showed different values of chlorophyll content due to the difference in the soluble solid present in the puree of all TSS. Ash content and density of puree increase slightly with an increase in TSS value, as TSS value represent its solid content.

The FTIR spectra confirmed the presence of –OH stretching (have a similarity to 3672 – 29,345 cm^-1), –CH stretching (have a similarity to 2856 cm^-1), bending aromatic compounds overtone (have a close similarity to 1792 and 1734.8), C=N stretching imine/oxime or C=O stretching (have a similarity to 1659.1 cm^-1), conjugated ketone or alkenes (have a similarity to 1635.1 cm^-1), and OH bending phenol (have a similarity to 1368 cm^-1).The frequency of absorbance at 2856 cm^-1 confirmed the presence of –CH stretching due to the formation of Hbond. At 1659.1, broad-shaped bands with medium intensities were observed which indicates the C=O stretching of COOH and Acetyl groups in the sample (Femenia et al. 2003). It was noted that with the increase in TSS, the intensity of the peaks also increased, this might be due to the reason that during the concentration process there is a conversion in functional groups. The presence of the biologically active compound is indicated by the presence of acetyl group, which helps to develop crossmolecular barrier in the cell (Chang et al. 2011). In the case of Aloe Vera powder, Kim et al. (2009) reported similar band formation. The previous studies of strong bands on Aloe Vera depict the presence of monosaccharide units in the branched regions such as glucan and galactose in the range of 1156 – 1230 cm^-1 (Gentilini et al.2014; Ray and Aswatha, 2013). Figure 2

Figure (1-3) exhibits the effect of temperature on steady shear properties of the purees with a TSS of 3, 4.35 and 5 °Brix.

The obtained data were fitted to six non-Newtonian models (Eqs. (1-6)). The best numerical models for characterizing the flow characteristics of purees was chosen based on the coefficient of determination (R²), RMSEs, the presence of yield stress and overall bias factors acquired during the fitting (Table 3 – 8). Although all models fitted well with the test data, at 5 °Brix esteem, Herschel-Buckley, Heinz-Casson, Vocadlo, and Mizrahi- Berk fitted model demonstrated negative values, which are having no physical importance. Hence RMSE values were not calculated for the rest of these unfitted models. Power law and Casson model fit satisfactorily at all temperature ranges examined and the R² value was found to be greater than 0.97 and hence were more appropriate to describe the rheological flow behavior of the puree. Figure 4, Table 4, 5, 6, 7

The variation of K value with temperature (278 – 318 K) of puree is given in Table 3 for power law fitted model. K varied from 6.01 to 350.28 Pa.s^-1 in puree with different TSS range. The lower values for K were observed at a higher temperature (318 K). The variation was more noticeable on puree having lower TSS value (3 °Brix). Higher TSS purees were found to have higher K value in magnitude. The increments in K were found with an increase in TSS value at a fixed temperature probably due to increment in the particle-to-particle contact (Chin et al., 2009). A similar result was obtained for another vegetable (Ahmed et al., 2013; Karababa and Develi Isikli, 2005). The impact of factors was discovered significant (p <0.05) for all TSS (Table 3). The dependence of K on the temperature variation was also performed using Turian model (Eq. (15)) for Puree. Highest flow behavior dependency was found for 5 °Brix at varying temperatures (Table 9). Increments in the temperature from 278 K to 318 K significantly affect the consistency index.

The n value of the puree ranged from 0.068 to 0.396 (Table 3) for various TSS studied, signifying gently non- Newtonian pseudoplastic shear thinning behavior of puree (Muller 1974, Steffe, 1996). Turian models (Eq. (16)) were used to monitor the dependence of n on its temperature change and the model parameters are shown in Table 9. The variation of n value with temperature was lower (0.307 – 0.396) in lower TSS value as compared with the higher TSS value (0.068 – 0.193) in all temperature range. This may be because of the increase in soluble solid content. The decrement in n value of puree showed the steady loss of pseudo-elasticity (Sikora et al., 2007). The sample having different temperature demonstrated different flow behavior values for different TSS sample. When the temperature varied from 278 K to 318 K, the n values decrease with increment in temperature. Comparative patterns for the n were accounted by Rao et al. (1981) for tomato juice concentrate. Besides TSS, the impact of every single other variable was found significant (p <0.05) (Table 3).

Effect of temperature on viscosity is important in determining the rates of heat exchange, energy utilization, and flow rates so it is feasible to assure persistent product flow (Nindo, 2005). Karwoski (2013) reported that TSS and temperature influence physical properties, such as viscosity, specific heat, refractive index, density and boiling point, of the fruit product. The plot of the log of viscosity against 1/T (K^-1) at various shear rates shown in Figure 5 and the constants of Eq. (9) were determined. It was observed that viscosity of puree concentrates diminish with increment in temperature over the range of shear rates. Figure 6 shows that viscosity decreases with a shear rate, which may be because of increased versatility of macromolecules, because of temperature rise, causing less resistance to flow. The change of the magnitude of viscosity was more at shear (10 s^-1) contrasted with the little distinction in a shear rate of 50 and 100 s^-1. Interaction impact of temperature and shear rate was additionally assessed and was significant for puree at all Brix value (p <0.05).

Activation energy is minimum energy required for particles to get away from the impact of its neighboring particles during the characterizing the viscous flow. The extent of the vitality of initiation for flow increased with increment in the TSS of the puree, it demonstrates that more energy was required to defeat potential vitality hindrance at higher TSS. Both the power law (Eq. (10)) and exponential model (Eq. (11)) were tried to determine the effect of flow activation energy on TSS of Puree. A nonlinear trend of increment in activation energy (p ≤0.05) of puree with the increase in solids content was observed. The exponential model (r >0.75) was more effective in explaining the impact of the TSS on the flow activation energy of the puree than the power law model (r <0.75) as shown in Table (10). A similar type of results was recorded in the case of pomegranate (Kaya and Sözer, 2005). The activation energy ranged between 3.544 – 8.610 kJ.mol^-1 for puree having different TSS range (3 – 5 °Brix) (Table 11). Higher values of activation energy may be associated with the higher temperature and TSS of the puree (Ahmed et al., 2007). In pseudoplastic vegetable product, the activation energy was proportional to the n i.e., the more pseudoplastic the product, the less the impact of temperature on its apparent viscosity (Sharoba et al., 2005). In this study, it was found that the activation energy was discovered lower for the puree having lower TSS value at a fixed shear rate (Table 11). Similar types of trends were observed by Hernandez et al. (1995) and Vitali and Rao (1984). However, it was also observed that for all TSS Value, the activation energy of puree concentrates increase with an increase in shear rate. Higher activation energy values show a more impact of temperature on its viscosity, i.e. faster change in viscosity with temperature (Sanchez et al., 2009). The outcomes demonstrated that temperature and TSS had significantly affected the thickness of puree concentrates. The viscosity value increases altogether with increment insoluble solid substance. Similar results were obtained by Alpaslan and Hayta (2002); Akbulut et al. (2012); Arslan et al. (2005).

From the food process designing perspective, it is necessary to get a single equation, which represents the effect of both temperature and TSS dependency on the viscosity of puree. Few authors have utilized different types of equation to depict the joined impact of temperature and TSS on viscosity of the liquids (Juszczak et al., 2010; Kaya and Sozer, 2005; Ibarz et al., 2009; Nindo et al., 2005; Nindo et al., 2007; Altan and Maskan, 2005; Juszczak and Fortuna, 2004).

Table 12 demonstrates the parameters for the various models for depicting the combined impact of temperature and TSS on the viscosity of puree. The correlation coefficients values of 0.992, 0.998 and 0.999 and % RMSE values of 10.259, 10.324 and 6.887 were obtained for the first order, exponential first order, and exponential second-order models, respectively. The second order exponential condition was ideal, because of high correlation coefficient and low percent RMSE values. The Estimated parameter values of Eqs. (12), (13) and (14) are given in Table 12 and model equation may be represented as (Eqs 17– 19).

(17) ln η = − 8.13685 + ( 602.275 T ) + 6.045833 (18) ln η = − 6.1791975 + ( 602.275 T ) + 1.56987034 C ) (19) ln η = − 4.900 + ( 602.275 T ) + 0.89 C + 0.085 C 2 )

The viscosity of puree was significantly (p <0.05) influenced by temperature and TSS of puree. The surface plot for the joined impact of temperature and TSS on the consistency of puree at various shear rates is shown in Figure 7 (a – c).

At lower temperatures the magnitude of viscosity increases quickly with a fixed TSS of Puree and increases insignificantly at higher temperatures, this was because of increment in thermal energy of the particles, which increase its intermolecular spacing. A comparable sort of results was accounted for other liquid food (Nindo et al., 2005; Ibarz et al., 1989). Size, shape, nature of solute and its condition of hydration and the type of solute and solvent influences the viscosity determination, thus might account for deviation in various cases. In the case of Roselle leaves, the solids present in the puree may be mainly fiber, diverse sugars (sucrose, fructose), vitamins, minerals, amino acids, proteins, hormones etc. The condition of hydration was distinctive for various TSS, the magnitude of viscosity relies on upon the sort of solids parts present in the puree (Nindo et al., 2005; Telis et al., 2007).

The influence of pH on the apparent viscosity of puree is shown in Figure 8. As the initial puree prepared from leaves is having pH of 1.3 ±0.26, so the puree is an anionic polysaccharide containing a great number of negative charge-bearing groups such as carboxyl groups, which may undergo different degrees of ionization with changing pH value, leading to changes in viscosity and rheological properties (Lapasin and Pricl, 1995). At shear rates less than 50 s^-1, the highest viscosity was obtained over the pH range of 1.3 – 5.3.

There is an increment in viscosity is due to the ionization of carboxyl groups. As per Feng et al. (2007), the viscosity will be at a most extreme when its atomic chains present in puree are entangling with each other. Past research work by the researcher on the impact of pH on viscosity of sweet potato puree and quince puree demonstrated that the consistency of puree increase with pH value (Ice et al., 1980). On the other hand, in the more basic region, the solution consistency dropped at a fixed TSS. As the pH is raised, the functional group present induces electrostatic repulsion that tends to keep the atoms in a developed shape, in this manner producing an exceedingly thick solution (Onweluzo et al., 1994; Launay et al., 1986). Results likewise demonstrated that at higher shear rates (more than 50 s^-1); changes in pH had no extensive impact on the viscosity.

In addition, the pH-dependence of viscosity may be due to change in puree confirmation. At low pH, polysaccharide chains tend to appear in coil state with acid groups in free acid form. With increasing pH, acid groups of coils are gradually ionized and the coils are expanded due to increase in electrostatic repulsion between functional groups, leading to more intermolecular interactions among the coils and consequent higher viscosity of the puree (Feng et al., 2007).

The maximum viscosity was obtained around the pH 5.3 for Puree, where the shape of chains may be close to rod conformational state (Achi and Okolo, 2004). This condition usually appears at where acid groups are ionized and electrostatic repulsion reaches a maximum and consequently, tends to keep the molecules in an extended form, leading to a high viscous solution and higher K values (Medina-Torres, 2000; Coupland, 2013). The decrease of viscosity from pH 1.3 to 5.3 may be explained by the neutralization effect of added alkali on the negative charges of the puree, which reduces the hydrodynamic volume of the puree and consequent viscosity (Chen et al., 2001; Porto et al., 2015) and puree depolymerization under alkali condition, proposed by Achi and Okolo, (2004).