Structural and Rheological Properties of Frozen Concentrated Orange Juice (FCOJ) by Multi-Pass High-Pressure Homogenisation (MP-HPH)

ABSTRACT

Multiple passes of high-pressure homogenisation (MP-HPH) was used to promote desirable rheological changes on frozen concentrated orange juice (FCOJ, 66ºBrix), up to 150 MPa and three passes. The final product structure (particle size distribution), rheological properties (steady-state shear and time-dependent flow) and colour were evaluated. By increasing both, homogenisation pressure and number of passes, MP-HPH process showed a decreasing asymptotic effect on the FCOJ consistency. The mean particle size decreased and the particle size distribution changed with the increase in pressure and number of passes. The process showed no impact on product colour. MP-HPH can be an interesting alternative for the production of FCOJ, providing similar effects as a single pass through the HPH equipment, but with lower costs and higher production rates.

KEYWORDS:

• High-pressure homogenisation
• Multiple passages processing
• Rheology

Introduction

In the high-pressure homogenisation (HPH) technology, a fluid is continuously pumped through a narrow gap and according to the mass and energy conservation laws, the speed of the fluid increases drastically, resulting in a high-pressure drop to the atmospheric value. It results in high shear stress, turbulence and cavitation, promoting changes in the product structure and properties^[1,2], mainly on the particles suspended and macromolecules in the product aqueous phase. For example, the fruit cells and its components, like pectin and cellulose are disrupted or deformed due to the high mechanical stress, promoted by this technology, at and or after the gap.^[1] The HPH technology was originally studied with the objective of replacing the thermal food processes. However, recently new interesting applications to HPH have been reported.

HPH can be carried out using consecutive passes, which is also known as Multi-pass high-pressure homogenisation (MP-HPH). It has been shown that MP-HPH processing at lower homogenisation pressures (P_H) was able to produce the same changes as one-pass processing at higher P_H, when considering enzyme activity^[3–5] and microbial inactivation.^[6–8] Martínez-Monteagudo et al.^[2] hypothesised that milder pressure-based technologies facilitate the interactions among food ingredients, allowing formulations with improved sensorial perceptions.

Conducting HPH processes at lower pressures reduces the equipment wear out (consequently, reducing the maintenance costs), allowing higher production rates and resulting in a lower increase of the product temperature. All this subjects are very interesting from an industrial point of view, highlighting the importance of studying the MP-HPH processing.

HPH can alter the rheological properties of several vegetable product, such as cashew apple juice, tomato juice, pineapple juice and mango juice. ^[9–13] Even so, it is improbable to predict the behaviour of each product after the HPH, since every vegetable matrix presents its own composition and structure arrangement, leading to different resistance profile during the high shear promoted by HPH. The release of internal cell content and/or cell fragments due to HPH process is unique to each food matrix, leading to different changes in particle-particle and particle-serum interaction. ^[14]

Orange juice is one of the most popular juices in the world, and its production is often derived from FCOJ. Further than guarantee the desired microbial inactivation^[15], the HPH technology has been shown to be a powerful tool for providing a reduction in the consistency of FCOJ, on both flow and viscoelastic properties.^[16,17] A reduction in the consistency of FCOJ reduces the energy consumption at other downstream unit operations, promoting a more efficient FCOJ processing. However, the effect of MP-HPH on the rheological properties of FCOJ is still unknown, although a lower P_H could result in even smaller processing costs (equipment and operation).

A noticeable and relevant effect of HPH on a product is not a guarantee that consecutive processes of HPH in the same product will show any impact. For example, different studies reported the effect of HPH on several properties of tomato products, including changes in the consistency. ^[10,18–22] On the other hand, the MP-HPH was not effective for further changing the rheological properties of tomato juice^[23], where only the first process showed eligible effects.

Then, the aim of this study was to evaluate the use of multiple passes of high-pressure homogenisation (MP-HPH) on FCOJ below the upper limit, in order to evaluate if a number of processes under mild conditions could be compared with one process at 150 MPa. Then aiming to optimise juice processing by maximising the HPH effect, with lowering costs.

Materials and methods

Frozen concentrated orange juice

The FCOJ (a commercial blend of juices from the Natal and Valencia varieties, 66°Brix, pH 3.8, 3.2% acidity expressed as citric acid, and 11.7% pulp content) was obtained directly from a local producer (Citrosuco, Limeira, São Paulo, Brazil). All samples have come from the same lot, and were divided into 1 L bottles to facilitate handling. The bottles were stored in a conventional freezer at −18°C before processing.

High-pressure homogenisation (HPH)

The process was carried out within the range from 0 MPa (control process, with no pressurisation) to 150 MPa (gauge, homogenisation pressures – P_H) in a high-pressure homogeniser (Panda Plus, GEA Niro Soavi, Italy). The juice was thawed at 5°C prior to processing. The inlet temperature was 25ºC to simulate the conditions just after juice concentration during industrial processing, after the concentration and before the freezing step. The FCOJ was introduced into the equipment by suction and quickly cooled using an ice bath after the homogenisation valve. After the first processing and cooling to 25ºC, the samples were subjected again to homogenisation to evaluate the effect of multi-pass processing.

The 0 and 150 MPa samples were processed with 1 pass, while the 50 MPa and 100 MPa samples were processed with 1, 2, and 3 passes. It is worth mentioning that the multi-pass HPH is not a practical processing using only one equipment (industrially equipment can be used sequentially). All processes were carried out, at least, in triplicate.

Structure: particle size distribution (PSD) analysis

The particle size distribution was evaluated by light scattering (Malvern Mastersiser 2000 with Hydro 2000s, Malvern Instruments Ltd, UK). A refractive index of 1.56 was used for this procedure. Droplets of the sample were slowly added into the sample compartment, already filled with distilled water (at room temperature), until obscurity was around ten. The mean diameter was also evaluated based on the particle volume (D[4,3]; Eq. (1)), and the mean diameter based on the particle surface area (D[3,2]; Eq. (2)). This is useful once the particles are not ideal spheres, and the D[3,2] is more influenced by small particles while the D[4,3] is more influenced by the larger ones.^[14,24] The assays were performed immediately after processing, in triplicate.

(1)

(2)

Instrumental colour

The instrumental colour was obtained using a ULTRA PRO colorimeter (Hunter Associates Laboratory, USA), with illuminant D65 and angle of 10°, previously calibrated with a RSIN white reference (L* = 92.03, a* = − 0.88, b* = 0.63). ^[25] The samples were placed in glass cuvettes and three readings were obtained for each replicate.

The CIELab methodology was used for the evaluation, where the values for L* (lightness), a* (redness: green to red) and b* (yellowness: blue to yellow) were first obtained, and then used to express the colour changes. The derived parameters (Hue and Chroma) were also calculated using the following expressions: Hue = arctan (b*/a*) and Chroma = ((a*)² +(b*)²)^1/2.

Rheological properties

Rheological analyses were carried out using a controlled stress (σ) rheometer (AR2000ex, TA Instruments, USA) with the temperature maintained constant at −10°C, simulating the average pumping temperature in industry after the freezing step. The rheometer was equipped with a Peltier system to control the temperature and a cross hatched plate-plate geometry (40 mm of diameter).The samples were analysed immediately after the HPH process. Prior to this, a gap-independency procedure was carried out, as described by Tonon et al.^[26], which consists of varying the distance between the plates until the flow curve reaches a steady state. The gap obtained was 1000 μm.

The rheological evaluation was carried out with not previously analysed samples, which were first placed in the rheometer and maintained at rest for 5 min before shearing. After resting, the samples were sheared at a constant shear rate (300 s⁻¹) for 300 s, while the shear stress was measured for the time-dependent (thixotropic) behaviour evaluation. After the time-dependent shear period, a stepwise linear decreasing protocol (300 s⁻¹ to 0.1 s⁻¹) was used to guarantee steady-state shear conditions for the flow behaviour evaluation.

The FCOJ time-dependent rheological properties were evaluated using the first part of the protocol. The shear stress decay was evaluated using the Figoni and Shoemaker model ^[27] (Eq. (3)). The steady-state shear rheological properties were evaluated using the second part of the protocol and the product flow behaviour was modelled using the power law model ^[28,29] (Eq. (4)).

Both the time-dependent and steady-state shear procedures were carried out at −10ºC to evaluate the rheological properties of the product under freezing and pumping, and during storage and transportation conditions.

(3)

(4)

Statistical evaluation

The parameters for each model (Eq. (3) and Eq. (4)) were obtained by linear or nonlinear regression using the CurveExpert Professional software (v.1.6.3, http://www.curveexpert.net/, USA) with a significant probability level of 95%. Moreover, the effects of the homogenisation pressure (P_H) and of the number of passes on the properties of the FCOJ were evaluated using the analysis of variance (ANOVA) and the Tukey’s test at a 95% confidence level. The STATISTICA 7.0 (StatSoft, Inc., USA) software was used for this purpose.

Results and discussion

FCOJ structure: particle size distribution (PSD)

Fig. 1 shows the effect of HPH process (homogenisation pressure and number of passes) on FCOJ PSD. This result is similar to that observed by other authors for other vegetable matrices considering only one pass. ^[10,11,16] Regarding the number of processes, two and three processes at 50 MPa showed a distribution close to that obtained with one single process at 100 MPa. Furthermore, two and three processes at 100 MPa showed a larger amount of smaller particles than the single 150 MPa process.

Figure 1. Effect of MP-HPH on the particle size distribution (PSD) of FCOJ.

Analysing the shapes of the PSD curves, an increase in P_H resulted in narrower distributions, as expected, with an asymptotic behaviour. The number of passes also presented an asymptotic effect, i.e., the subsequent processes have a smaller impact on PSD when compared with the previous process using the same homogenisation pressure.

When both P_H and the number of passes increased, there was an accumulation of small particles, evidenced by the formation of a plateau, probably due to the disruption of suspended particles. In fact, the distribution has changed from a monomodal to a bimodal distribution, as can be seen in Fig. 1. There is a minimum particle size that the equipment can produce, due to the equipment dimensions, which is related to the peak of small particles (~1–2 μm. The peak position practically does not change with P_H or the number of passes. However, the amount of larger particles decreased, leading to a decrease in the mean particle size, as evidenced by the change in position of the peak of larger particles.

Fig. 2 shows the mean diameter of FCOJ as a function of the homogenisation pressure (P_H) and number of passes. The value of D[4,3] was almost tenfold higher than the value observed for D[3,2] for all samples. D[4,3] is more influenced by larger particles and no statistical difference (p < 0.05) was observed among the processing conditions at 50 MPa/2 passes and the others. On the other hand, D[3,2] is more influenced by smaller particles and presented significant differences (p < 0.05) with multiple passes. This corroborates with the PSD results (Fig. 1), since the multiple HPH passes increased the number of small particles.

Figure 2. Effect of MP-HPH on the mean particle diameters (D[4,3] and D[3,2]) for FCOJ (Vertical lines are standard deviation; different letters represent significantly different values (p < 0.05).).

The results indicated that the MP-HPH has higher effect on larger particles than on smaller ones, as could be seen in Fig. 1. The number of smaller particles increased (i.e. the area of the peak increased) without changing the position of its peak (i.e., the minimum diameter), which can be explained by the physical limitation regarding the gap dimension. Therefore, the proportions of small and large particles have changed with the increase in both the homogenisation pressure (P_H) and the number of passes, suggesting changes on product rheology.

Instrumental colour

The effects of both homogenisation pressure (P_H) and the number of passes on the instrumental colour parameters of FCOJ (CieLAB system; L*, a*, b*) are presented in Fig. 3. None of the primary parameters of instrumental colour was influenced by the HPH processing, as well the hue and chroma (data not shown), neither by single nor by multiple processing (p < 0.05). This result is compatible with those obtained by Leite et al.^[16] with single pass HPH processed FCOJ. FCOJ showed major changes in its structure during its production (for example, during extraction, pulping, thermal processing, concentration, and freezing). Consequently, the pigments were already well dispersed in the fluid medium before the HPH process, which did not promote additional pigment liberation. On the other hand, on products that still contain whole cells, for example tomato juice, the HPH processing was able to rupture the cells containing pigments, which can be oxidised during storage, leading to colour changes.^[30] It is worth mentioning that this result is highly desirable since the consumers will not perceive the use of HPH in the production of FCOJ.^[31]

Figure 3. FCOJ primary parameters of instrumental colour (CieLAB system) as a function of the MP-HPH process (vertical lines are the standard deviation, the values for each parameter show no statistical difference in relation to the homogenization pressure (p < 0.05)).

Rheological behaviour

Fig. 4 shows the effect of multiple passes on the flow behaviour of FCOJ. The product showed shear thinning behaviour without yield stress, being therefore, described by the Power Law model (Eq. (4); R² > 0.99). Table 1 shows the flow behaviour equations for the Power Law model obtained for each sample (homogenisation pressure and number of passes).

Table 1. Effect of single and multiple HPH processing on flow properties of FCOJ: Ostwald–de-Waele model (Eq. (4)) at −10°C.

PH (MPa)	Passes
	1	2	3
0		–	–
50
100
150		–	–

Figure 4. Flow behaviour of FCOJ at −10°C: Effect of MP-HPH.

As expected the HPH reduced the juice consistency, as described by the consistency index (k). The HPH disrupted the suspended particles, resulting in smaller particles and a different ratio between the smaller and larger ones. Smaller particles showed less resistance to flow, leading to a lower consistency index (k). The smaller particles can occupy the spaces between the larger particles, creating a lubricant effect, leading to less friction between particles.^[32] Further, the reduction on FCOJ consistency could also be due to changes in the serum phase, as the HPH process can promote reduction in the molecular size of polysaccharides.^[16] Therefore, both phenomena are related to the effect of HPH on the FCOJ rheology.

An asymptotic behaviour was also noticeable for both the homogenisation pressure (P_H) and the number of passes, in relation to the consistency index (k). The effect of processing with multiple passes was higher at lower homogenisation pressures. The juice processed at 50 MPa showed larger changes when compared with the samples processed at 100 MPa. Two passes at 50 MPa had almost the same effect as one pass at 100 MPa. Three passes at 50 MPa showed a reduction in consistency similar those obtained with a single process at 100 MPa. Finally, the results were similar between one pass at 150 MPa, and two passes at 100 MPa. Three passes at 100 MPa showed a larger reduction than one pass at 150 MPa.

It was thus demonstrated the potential of using multiple passes at lower pressures, in spite of single one at higher pressures, in order to reduce the FCOJ consistency. Nevertheless, it must be highlighted that MP-HPH is not a practical processing (in an industrial perspective); thus, it would be more interesting if a small number of passes were effective. Therefore, similarly to Picart et al.^[33], only three passes were evaluated.

The reduction in PSD also affects the rheological properties of FCOJ due to the particle alignment factor, which can be measured by the flow behaviour index (n). Smaller particles easily align to the flow direction, when compared with larger ones. Newtonian fluids have no alignment factor, showing a flow behaviour index equal to 1.0. The “n” value for FCOJ increased towards 1.0 with the increase in both homogenisation pressure and the number or passes. This indicates that the disrupted suspended particles have a lower alignment factor, which also contributes to a less complex flow and less friction during flowing – both interesting results from an industrial perspective.

In fact, the flow behaviour index (n) increased with both the number of passes and homogenisation pressure. Two passes at 50 MPa had almost the same effect than one pass at 100 MPa. Three passes at 50 MPa showed similar flow behaviour index to that obtained with one single process at 100 MPa, and close to that obtained with two passes at 100 MPa. Three passes at 100 MPa showed a large increase in the flow behaviour index when compared with the 150 MPa sample.

The asymptotic behaviour with the number of passes has been shown in other studies considering the microbial inactivation. ^[6] However, increasing the number of passes was not effective for additional changes in the rheological properties of tomato juice. ^[23] Therefore, the effect of MP-HPH was not an additive effect, since it was related to a physical limitation of the equipment (the gap dimension and pressure), which limits the maximum shear stress and mechanical energy delivered to the product. Larger particles are more affected by HPH, being broken into small particles, whose resistance to disruption is greater. Consequently, the smaller particles pass through the gap with little or no change. This was confirmed by the increase in the amount of small particles, with concomitant decrease in the larger ones, as shown in Fig. 1, but only until a specific dimension. Thus, the multi-passage processing stands out when using few passes at small homogenisation pressures.

Fluid consistency plays a major role in the process of FCOJ, since it has a great impact on energy consumption during processing. A reduction in the consistency of FCOJ is thus highly desirable for various unit operations, especially with respect to the amount of FCOJ produced. The reduction in consistency at 150 MPa represents approximately 58% when compared to the original value of the apparent viscosity at 300 s⁻¹ of the FCOJ (Table 1). However, the use of such high pressures (up to 150 MPa) increases the equipment costs resulting in high energy consumption, smaller processing capacity and higher erosion. For example, to obtain an equivalent consistency reduction of a single pass at 100 MPa, two passes at 50 MPa could be used. Equipment specifically designed to work at 50 MPa has higher flow rate when compared with the equipment at 100 MPa, which improves the process productivity. In addition, the maintenance costs are also lower due to less wear and tear of valves and gaskets. MP-HPH is a good alternative in the production of FCOJ, once it uses lower pressures to reduce product consistency.

Fig. 5 shows the thixotropic behaviour of FCOJ as a function of MP-HPH processing. The results showed that the effect of MP-HPH on the thixotropy of FCOJ was similar to the tendency observed for the flow behaviour, with asymptotic behaviour for both homogenisation pressure (P_H) and number of passes.

Figure 5. Shear stress against time (thixogram) for FCOJ during shearing at 300 s-1 at −10°C: Effect of MP-HPH.

The thixotropy of FCOJ was well described by the Figoni–Shoemaker model (Eq. (3)), whose parameters are shown in Table 2 (R² > 0.97). Similar to the flow characterisation, multiple passes at 50 MPa caused reductions in both the initial (σ_i) and equilibrium stresses (σ_e), which were similar or even greater than those caused by a single process at 100 MPa. In addition, the multiple passes at 100 MPa caused similar or even greater reductions than single processing at 150 MPa.

Table 2. Effect of single and multiple HPH processing on the thixotropic properties of FCOJ: Figoni–Shoemaker model (Eq. (3)) at −10°C.

PH	Passes
	1	2	3
0		–	–
50
100
150		–	–

As the PSD decreased (Fig. 1 and Fig. 2), the initial resistance to shear also decreased, explaining the reduction on the initial stress parameter (σ_i). In addition, the bimodal PSD resulted in a smaller friction during flowing, less drag by the small particles through the serum and a smaller alignment factor, which explains the decrease in equilibrium stress (σ_e).

Thixotropy is related to the destruction or disruption of the internal structure due to flowing.^[34,35] Although the internal structure can be formed by strong attractive interactions, this is not the case of FCOJ, which is a fluid with no yield stress (σ_o). However, even the random non-alignment distribution and entanglement of the constituents could show an initial resistance to flow. Another factor that could result in thixotropic behaviour is related to the initial particle inertia to flow, especially due to the high consistency of FCOJ. These factors contribute to an initial resistance to flow, gradually changing during flowing to an optimal alignment, until reaching equilibrium stress.

Corroborating to the results for single-pass HPH ^[16], it was not possible to observe a tendency for the thixotropic kinetic factor (k_FS) in relation to neither the homogenisation pressure (P_H) nor the number of passes. Similar to the flow behaviour, both the initial and equilibrium stresses of the FCOJ processed with multiple passes at 50 MPa showed similar results when compared with one or two passes at 100 MPa. Three passes at 100 MPa caused a greater reduction when compared with a single processing at 150 MPa. The physical limitation of the gap is responsible for the non additive effect of multiple processes.

The MP-HPH can reduce the FCOJ thixotropic behaviour, which is also high relevant during the industrial process. It is due to the reduction in the initial and equilibrium stresses, which is highly desirable from an energy saving point of view, as it also reduces the process complexity.

Finally, all results showed in this section indicated that the FCOJ processed by MP-HPH requires less energy during processing than a non-homogenised one. Therefore, it is suggested that the process of the MP-HPH can be used on FCOJ industry to save energy, reduce the processing time and reduce maintenance costs. However, a specific analysis considering a real situation is needed to confirm it, also considering the energy needs for running MP-HPH process.

Conclusion

MP-HPH affected the microstructure of FCOJ and, consequently, its rheological behaviour. The PSD changed from a monomodal distribution to a bimodal distribution, changing the ratio between small and large particles. The consistency of FCOJ decreased with both homogenisation pressures (P_H) and number of passes. It was concluded that similar results could be obtained with 2 or 3 passes using a lower homogenisation pressure, when compared to single processes at higher pressures. The use of 2 passes at 50 MPa to achieve an effect similar to that observed at 100 MPa could be industrially advantageous, since a small reduction in pressure decreases the need for highly specific equipment, energy consumption and wearing, enabling the use of equipment with higher capacity. Therefore, the use of multiple process is a good alternative to obtain results similar to the single pass process for this juice, using low cost equipment and with lower energy consumption.

Nomenclature

= shear rate [s⁻¹]

σ = shear stress [Pa]

σ₀ = yield stress [Pa]

σ_e= equilibrium stress in the Figoni–Shoemaker model (Eq. (3)) [Pa]

σ_i = initial stress in the Figoni–Shoemaker model (Eq. (3)) [Pa]

d = particle diameter (Eq. (1): Eq. (2)) [μm]

dp/dz = linear pressure gradient (Eq. (5)) [Pa/m]

D[4,3] = particle volume-based diameter (Eq. (1)) [μm]

D[3,2] = particle area-based diameter (Eq. (2)) [μm]

k = consistency index, power law model (Eq. (4), Eq. (5)) [Pa·sⁿ]

k_FS = kinetic parameter in the Figoni–Shoemaker model (Eq. (f)) [s⁻¹]

n = flow behaviour index, power law model (Eqs. (4); (5) and (6)) [-]

P_H = homogenisation pressure [MPa]

t = time (Eq. (3)) [s]

Funding

The authors thank the São Paulo Research Foundation (FAPESP) for funding project no. 2012/15253-9 and TS Leite scholarship (2012/17381-4).

Additional information

Funding

The authors thank the São Paulo Research Foundation (FAPESP) for funding project no. 2012/15253-9 and TS Leite scholarship (2012/17381-4).