Moisture sorption properties of freeze-dried arazá ( Eugenia stipitata Mc Vaugh) powder: Effect on physicochemical and thermodynamic properties

Abstract

BACKGROUND:

Arazá, a flavorsome berry rich in antigenotoxic and antimutagenic antioxidants (phenolics/carotenoids/flavonoids/ascorbic acid) could be used for preparing functional powdered instant-beverages. Moisture sorption can cause quality problems therefore a thorough analysis of powder sorption properties is required.

OBJECTIVE:

To study freeze-dried arazá moisture sorption behavior using maltodextrin or Arabic-gum as drying-aids and its effect on isosteric-sorption-heat, differential-entropy, mechanism controlling moisture-sorption, color, solubility and antioxidant content/activity.

METHODS:

To determine arazá/maltodextrin and arazá/Arabic-gum moisture-sorption capacities and their effect on powders physicochemical properties, samples were exposed to relative humidities ranging 11–76 and 10/20/40°C until reaching equilibrium. Changes in equilibrium water content (W_c), cold water solubility, antioxidant (polyphenols, flavonoids, carotenoids and ascorbic acid) concentration and activity were recorded.

RESULTS:

GAB model satisfactorily predicted W_m; the samples monolayer-moisture concentrations were lower than 0.1 kg H₂O (kg dried matter)^–1 indicating good powder stability. Thermodynamic properties/W_c relationships and the sorption process mechanism were determined. At 10/20°C arazá/maltodextrin had better color/solubility, higher total polyphenols/carotenoids/flavonoids and lower ascorbic acid contents than arazá/Arabic-gum, however, antioxidant activities were similar. Temperature/water activity (a_w) combinations for optimum color, solubility, antioxidant content/activity were 10°C/a_w ≤0.11 (arazá/Arabic-gum) and 10°C/a_w ≤0.23 (arazá/maltodextrin).

Keywords

Arazá moisture sorption drying-aid type physical properties antioxidant activity

Abbreviations

AMD

Arazá/Maltodextrin

MD10

Maltodextrin DE10

AAG

Arazá/Arabic gum

Arabic gum

d.m.

Dry matter

a_w

Water activity

Relative humidity

W_c

Equilibrium moisture content (kg H₂O (kg d.m)^–1)

W_m

Monolayer moisture content (kg H₂O (kg d.m)^–1)

K/C

GAB model constants

Mean relative deviation modulus

Absolute temperature (°K).

T_hm

Harmonic mean temperature (°K)

T_β

Isokinetic temperature (°K)

Universal gas constant (0.008314 kJ mol^–1 °K^–1)

q_stn

Net isosteric heat (kJ mol^–1)

Q_st

Total heat of sorption (kJ mol^–1)

q₀

Net isosteric heat of the first water molecule (kJ mol^–1)

ΔS_d

Differential entropy (kJ mol^–1 K^–1)

ΔG_β

Free energy (kJ mol^–1) at T_β

L*a*b*

CIEL*a*b* color coordinates

saturation index

hue angle (°)

Ascorbic acid content (mg AA (100 g d.m (dry matter))^–1)

Total polyphenols content (mg GAE (gallic acid eq) (g d.m)^–1)

Total carotenoids content (mg β-carotene eq) (g d.m)^–1

Total flavonoids (mg CAT (catechin eq) (g d.m)^–1)

ARA

Antiradical activity

FRAP

Ferric reducing antioxidant power (μM Fe⁺²) (g d.m)^–1

1 Introduction

Arazá (Eugenia stipitata Mc Vaugh) is an amazonian berry highly appreciated because of its attractive yellow color, exotic flavor and high levels of antioxidants with proven antimutagenic and antigentoxic effects [1]. These properties combined with its high popularity suggest that it could be a very promising ingredient for preparing functional beverages.

Arazá has a high content of heat sensitive aroma, flavor and bioactives compounds like phenolic acids, flavonoids, ascorbic acid and β-carotene [2], in this cases, Marques, Ferreira and Freire [3] recommended to use freeze-drying to obtain a high-quality dried product.

Although drying is one of the most common means of preserving fruit juices; the storage conditions used may cause undesirable sensory and nutritional changes due to inadequate temperature and powder moisture sorption [4].

Moisture sorption isotherms are important for calculating the moisture level corresponding to optimum food stability [5] and other thermodynamic functions like the isosteric heat of sorption, the differential entropy and the mechanisms that control the sorption process [6]. These properties are used in the design, modeling and optimization of the drying process as well as for predicting the powder stability and quality during packaging and storage [7]. Variation of the heat of sorption with moisture provides valuable information for energy requirements calculations and knowledge of the extent of the water/solid vs. water/water interaction [8].

Arazá color depends on its carotenoids concentration, these compounds, as well as ascorbic acid, flavonoids and phenolic compounds are the most important antioxidants present in the fruit [2]. Martinelli et al. [9] reported that their stability is highly affected by temperature and the powder humidity content therefore, to obtain a high quality freeze-dried arazá powder a thorough analysis of the color, antioxidants concentration and activity relationship with temperature and the product moisture sorption capacity is needed.

The objectives of this study were:

To determine and model the adsorption isotherms of freeze-dried arazá at three temperatures (10, 20 and 40°C) using MD10 or AG as drying-aids.

To calculate the isosteric heat of sorption, the differential entropy and to determine the mechanism controlling arazá moisture sorption process and

To analyze the interaction between moisture sorption capacity, temperature and the physical (color, solubility in cold water) or chemical, (antioxidant concentration and activity) properties of the freeze-dried fruit.

2 Materials and methods

2.1 Fruit selection and processing

Fresh arazá (moisture: 94.5±2.4%, soluble solids (SS): 3.9±2.4°Brix) was purchased at a local market in Ibagué (Colombia) and stored 7 days at 4°C until processing. The fruits, selected according to its color (green/yellow) and size (12–14 cm diameter), were washed, peeled, cut and manually press to form a paste which was mixed with different concentrations (0, 2.5, 5%) of (a) maltodextrin Dextrose Equivalent 10 (AMD; Productos de Maız S.A., Buenos Aires, Argentina) or (b) Arabic gum (AAG; Quimica Oeste S.A., Buenos Aires, Argentina). To ensure the highest possible water removal, the paste (≈300 g) placed in aluminium trays, was initially frozen at –35°C (48 h) followed by a 24 h period at –80°C. The frozen mixtures were freeze-dried (48 h) with a FIC L1-1-E300-CRT freeze dryer (vacuum pressure < 13.32 Pa, condenser temperature=–35°C, shelves temperature = 22°C; Buenos Aires, Argentina).

2.2 Water sorption isotherms

The adsorption isotherms of the AMD and AAG powders were determined by the static gravimetric method [10] at 10, 20 and 40°C. Triplicate samples (1.5 g), placed in previously weighed plastic containers, were kept in desiccators containing saturated solutions of different salts that provided environments with a constant relative humidity (RH). The salts selected and their water activities values at the different temperatures were:

LiCl [0.11 (10, 20, 40°C)]; CH₃COOK [0.23 (10, 20°C), 0.21 (40°C)]; MgCl₂ (0.33 (10, 20, 40°C)], K₂CO₃ [0.43 (10, 20, 40°C)], Mg(NO₃)₂ [0.57 (10°C), 0.54 (20°C), 0.48 (40°C)], NaCl [0.75 (10, 20, 40°C)] [10].

The desiccators were stored in temperature controlled chambers at 10, 20 and 40°C. To monitor the moisture gain, the pre-weighed container with the sample was taken from the desiccator and weighed every 3 days with an analytical scale until reaching equilibrium (difference between 2 consecutive weights < ±0.003 g), the equilibration period lasted 2–3 weeks. At this point, the water activity (a_w) of the equilibrated samples can be considered equal to the corresponding RH/100 [11].

2.3 Water activity and moisture content analysis

Water activity (a_w) was measured at 25°C in an AquaLab series 3 (Decagon Device, Pullman, Washington, USA), calibrated with the saline solutions used for the sorption experiments.

The moisture content was analyzed in triplicate as described by the AOAC [12] method.

2.4 Mathematical modelling of the water sorption isotherm

The relationship between the equilibrium moisture content (W_c; kg H₂O (kg d.m)^–1) and the a_w of the samples was predicted with the Guggenheim-Anderson-de Boer (GAB) model (Equation (1); [10]). $W_{c} = \frac{W_{m} {KCa}_{w}}{({1 - Ka}_{w}) ({1 - Ka}_{w} {+ CKa}_{w})}$ (1)

W_m (kg H₂O (kg d.m)^–1) represent the monolayer water content, while C and K are constants related to the monolayer (C) and multilayer (K) sorption heats. Although there are several equations describing water sorption in foods, the GAB equation has the advantage of giving a good description of the process for the whole a_w range and its parameters have physical meaning [13].

Parameters were estimated using nonlinear regression analysis with the OriginPro v 8.0 (OriginLab Corp., Northampton, MA USA) software. The model goodness of fit was evaluated with the mean relative deviation modulus (E; Equation 2) $E = \frac{100}{n} \sum_{i = 1}^{n} \frac{[W_{ci} {- W}_{cpi}]}{W_{ci}}$ (2)

“n” represents the number of observations while W_ci and W_cp are the experimental and predicted equilibrium water contents respectively. Lomauro et al. [14] reported that E < 10 could be considered indicative of a good fit.

Lewicki [15] concluded that for a good description of sigmoidal type isotherms and to assure that the difference between the true and predicted W_m results is less than±15%, the K and C values must be within the range: 0.24 < K≤1 and 5.67≤C≤∞.

2.5 Thermodynamic properties

The net isosteric heat or differential enthalpy of sorption (q_stn; kJ mol^–1), the total heat of sorption (Q_st, kJ mol^–1) and the differential entropy (ΔSd; kJ mol^–1 K^–1) were calculated from the equilibrium data using the following equations [16 –18]. ${lna}_{w} |_{W_{c}} = \frac{Δ Sd}{R} - \frac{q_{stn}}{RT}$ (2) $Q_{st} = q_{stn} + Δ H_{cond}$ (3)

a_w represents the predicted water activity value for a specific equilibrium moisture content (W_c); ΔH_cond is the free water latent heat of condensation (44.05 kJ mol^–1), calculated at the average temperature between 283 and 313°K (298°K), R is the universal gas constant (0.008314 kJ (mol °K)^–1) and T the absolute temperature (°K).

q_stn and ΔS_d were calculated from the slope (-q_stn/R) and the intercept (ΔS_d/R) of the ln(a_w) vs 1/T plot at each W_c.

The q_stn or Q_st relationship with the equilibrium moisture content was determined with the exponential empirical equation proposed by Tsami et al. ([19]; Equations (4, 5)). $q_{stn} {= q}_{0} e^{(\frac{- W_{c}}{W_{0}})}$ (4) $Q_{st} = q_{0} e^{(\frac{- W_{c}}{W_{0}})} + Δ H_{cond}$ (5)

Where q₀ = q_stn at W_c = 0 and W₀ = W_c at q_stn = 0.368 q₀.

The enthalpy-entropy compensation theory proposes a linear relationship between q_stn and ΔS_d (Equation 6) and could be used for determining the mechanisms controlling the sorption process [20]. $q_{stn} = Δ S_{d} T_{β} + Δ G_{β}$ (6)

T_β is the isokinetic temperature (°K) and represents the temperature at which all reactions proceed at the same rate and ΔG_β is the free energy (kJ mol^–1) at T_β. Both parameters were estimated by fitting Equation (4) to the q_stn and ΔS_d results calculated previously.

Krug et al. [21] concluded that a linear chemical compensation pattern also requires that the isokinetic and the harmonic mean temperature (T_hm; Equation (7)) should be different (P < 0.05) $T_{hm} = \frac{n}{\sum_{i = 1}^{n} (1 / T)}$ (7)

n = total number of isotherms.

2.6 Cold water solubility

The moisture sorption effect on AMD and AAG solubilities (Equation (8)) at 4°C was determined in duplicate according to Chen and Jane [22] $Solubility (%) = \frac{(p_{ss} * FD * 100)}{p_{l}}$ (8)

pss and pl = dried weights (g) of the supernatant and the sample respectively; FD = dilution factor.

2.7 Color analysis

Color was measured on triplicate samples with a Minolta CR-400 Chroma Meter (Minolta, Osaka, Japan), each value was the average of 9 measurements. Color was expressed by its lightness (L*), saturation index (SI), a measure of color intensity. $SI = \sqrt{(a^{* 2} {+ b}^{* 2})}$ (9)

and hue angle (HA) representing the psychrometric hue [23]. ${HA = \tan}^{- 1} (\frac{b^{*}}{a^{*}})$ (10)

Hue angle values equal to 0, 90, 180 and 270° correspond to red-purple yellow, bluish-green and blue respectively.

2.8 Antioxidant concentration and activity

2.8.1 Sample preparation and extraction of antioxidants

Some biologically effective antioxidants, especially those associated with complex carbohydrates and proteins, may remain in the extraction residue and consequently, the real and calculated antioxidant content and activity may not agree [24]. To solve this problem, AMD and AAG antioxidants extraction was done following the procedure recommended by Saura-Calixto et al. [25].

0.5 g of powder were extracted with 20 ml of methanol/water (50 : 50, v/v; pH 2.0) for 60 min under constant shaking. To separate the residue (R1) from the supernatant (S1), the mixture was centrifuged at 2500 g for 10 min. following this step, R1 was extracted with 20 ml of acetone/water (70 : 30, v/v) for 60 min. The methanolic and acetonic supernatants were combined and used for analyzing the antioxidant content and activity associated with extractable compounds.

Non-extractable polyphenols (condensed tannins-proanthocyanidins and hydrolysable polyphenols) were determined in the residue of the methanol/acetone/water extraction [25].

To extract the condensed tannins/proanthocyanidins, the residue (R1) was treated with 5 ml/l HCl–butanol (3 h at 100 °C), the soluble fraction was separated by centrifugation (2500 g, 10 min) and used for analyzing condensed tannins/proanthocyanidins concentration and antioxidant activity (Secs. 2.8.2 and 2.8.3).

To determine hydrolysable polyphenols: the residue (R1) was hydrolysed with methanol/H₂SO₄ 90 : 10 (v/v) at 85°C for 20 h. After centrifugation (2500 g, 10 min), the supernatant was collected and used for total polyphenols content and antioxidant activity determinations as described in Secs. 2.8.2 and 2.8.3.

2.8.2 Total polyphenols, condensed tannins-proanthocyanidins, flavonoids, carotenoids and ascorbic acid contents

The total phenolics content of the extracts (TP; mg GAE (gallic acid equivalents)(g d.m)^–1 was assessed with the Folin-Ciocalteau method [26].

The condensed tannins-proanthocyanidins content of the supernatant from the HCl–butanol treatment was quantified with the protocol published by Saura-Calixto et al. [25].

Total flavonoids (TF; mg CAT (catechin equivalents)(g d.m)^–1 and carotenoids levels (TC; mg β-carotene equivalents)(g d.m)^–1 were determined according to Chang et al. [27] and Ordoñez-Santos et al. [28] respectively while the ascorbic acid (AA, mg AA (100 g d.m)^-1) was analysed as described by Nováková et al. [29].

2.8.3 Antiradical activity and ferric reducing antioxidant power (FRAP)

The AMD and AAG antiradical activity (ARA; Equation (11)) were determined with the DPPH scavenging assay [30]. $ARA = [\frac{{Abs}_{b}^{0 -} {Abs}_{s}^{t}}{{Abs}_{b}^{0}}] * 100$ (11)

Absb0 and Absst are the absorbancies of the blank (b) and the sample (s) at time = 0 and at t = 120 min.

Ferric Reducing Antioxidant Power (FRAP; μM Fe⁺² (g d.m)^–1) was evaluated according to Pulido et al. [31] after 30 min of reaction time.

2.9 Statistical analysis

The effect of temperature, drying-aid type and water activity on the color, solubility, antioxidant content, antiradical activity and FRAP were analyzed with the Infostat (v. 2013) software. Significant differences among means were determined by analysis of variance followed by pairwise comparisons with the Tuckey test. P values < 0.05 were considered statistically significant. Model parameters were calculated using regression analysis with the OriginPro v 8.0 software.

3 Results and discussion

The samples prepared with 0 or 2.5% MD10 (AMD) or AG (AAG) had a rubbery texture with a_w values of 0.36 and 0.13 respectively, caused probably by an insufficient drying-aid level and/or jellification of the pectins present in the fruit [32]. In contrast, the samples with 5% drying-aid content had good flowing properties with a_w = 0.08–0.09 (Table 1), therefore, this concentration was selected for the rest of the experiments.

Table 1
AAG and AMD physicochemical properties before equilibration

Initial parameters

AAG AMD

a_w 0.09±2x10^–3b 0.08±2x10^–3a

H (kg H₂O (kg d.m)^–1) 0.03±2x10^–3b 0.02±2x10^–3a

L* 83.92±2.29^a 86.79±0.79^a

SI 33.35±0.49^a 34.97±0.32^a

HA (°) 86.14±0.33^b 93.26±0.27^a

Hydrolyzable polyphenols ND ND

Non-extractable polyphenols ND ND

TP (mg GAE (g d.m)^–1) 92.36±0.34^b 99.72±0.79^a

TF (mg (CAT (g d.m)^–1) 14.27±0.48^b 18.01±0.28^a

TC (mg β-carotene (g d.m)^–1) 29.14±0.28^b 44.85±0.78^a

AA (mg AA (100 g d.m)^–1) 92.56±1.63^b 42.97±0.19^a

ARA (%) 69.43% ±0.21^a 72.96% ±2.04^a

FRAP (μM Fe⁺²(g d.m)^–1) 1302.56±3.98^a 1311.37±5.69^a

Solubility (Sol %) 80.06±0.12^a 81.96±0.38^a

	Initial parameters
a_w	0.09±2x10^–3b	0.08±2x10^–3a
H (kg H₂O (kg d.m)^–1)	0.03±2x10^–3b	0.02±2x10^–3a
L*	83.92±2.29^a	86.79±0.79^a
SI	33.35±0.49^a	34.97±0.32^a
HA (°)	86.14±0.33^b	93.26±0.27^a
Hydrolyzable polyphenols	ND	ND
Non-extractable polyphenols	ND	ND
TP (mg GAE (g d.m)^–1)	92.36±0.34^b	99.72±0.79^a
TF (mg (CAT (g d.m)^–1)	14.27±0.48^b	18.01±0.28^a
TC (mg β-carotene (g d.m)^–1)	29.14±0.28^b	44.85±0.78^a
AA (mg AA (100 g d.m)^–1)	92.56±1.63^b	42.97±0.19^a
ARA (%)	69.43% ±0.21^a	72.96% ±2.04^a
FRAP (μM Fe⁺²(g d.m)^–1)	1302.56±3.98^a	1311.37±5.69^a
Solubility (Sol %)	80.06±0.12^a	81.96±0.38^a

a_w: Water Activity; AA: Ascorbic Acid; ARA: Antiradical Activity; FRAP: Ferric Reducing Antioxidant Power; H: Water Content; HA: Hue Angle; L*: Lightness; SI: Saturation Index; TP: Total Polyphenols; TF: Total Flavonoids; TC: Total Carotenoids; Within each line, properties with different superscripts are significantly different (P < 0.05).

3.1 Characterization of the samples

Table 1 shows AMD and AAG powder’s physicochemical properties before equilibration. Statistical analysis indicated that drying-aid type influence on the samples solubility, lightness and saturation index was not significant (P > 0.05), however, the hue angle values in AAG were 7% lower (P < 0.05) than those from AMD indicating that the former was slightly redder than the later.

The TP, TF and TC contents were 8, 26 and 54% higher (P < 0.05) in AMD than in AAG (Table 1), however, AA behavior was the opposite, as its content was 115% greater in the latter. Ballesteros et al. [33] and Silva et al. [34] reported similar drying-aid effects in TP, TF and AA contents of spent ground coffee and camu-camu spray-dried with AG or MD10.

The antiradical activity (ARA) and the Ferric Reducing Antioxidant Power (FRAP) levels were not affected by the drying-aid type (P > 0.05).

3.2 Moisture sorption isotherms

A comparison between the AMD (Fig. 1a) and AAG (Fig. 1b) isotherms indicated that at 20 and 40°C and a_w ranging 0.11–0.33, W_c levels in both samples were similar ((P _> 0.05); 0.053 kg H₂O (kg d.m)^–1). However, at 10°C, W_c (AMD) remained stable up to a_w = 0.43 whereas W_c (AAG) behavior was similar to that observed at higher temperatures. At a_w = 0.43 or 0.75, no drying-aid effect was detected, the equilibrium water content mean values were: 0.13 and 0.25 kg H₂O (kg d.m)^–1 respectively.

Fig.1

Temperature influence on the sorption isotherms of freeze–dried AMD (a) and AAG (b) Lines represent the predicted equilibrium moisture contents.

Figures 1(a) and (b) show the predicted results of the adsorption isotherms for AMD and AAG freeze-dried samples. R² and E values (Table 2) indicated that the model fit was good and satisfied previous reports regarding goodness of fit criteria ((E < 10; [14]). Since K≈1 (P > 0.05), Lewicki’s [15] recommendations regarding the validity ranges of the K and C values (Sec. 2.5) were also fulfilled.

Table 2

Regression parameters and statistical tests (R²; E) of the GAB model used for predicting the AAG and AMD sorption isotherms at 10, 20 and 40°C

Parameters	10 °C				20 °C				40 °C
	AAG	Error	AMD	Error	AAG	Error	AMD	Error	AAG	Error	AMD	Error
W_m (kg H₂O (kg d.m)^–1)	0.06	6.52x10^–3	0.05	2.62x10^–3	0.05	2.45x10^–3	0.05	3.56x10^–3	0.04	5.17x10^–3	0.03	3.27x10^–3
K	1.02	0.03	1.03	0.01	1.05	0.01	1.04	0.01	1.07	0.02	1.09	0.02
C	200		200		200		200		200		200
R ²	0.91		0.97		0.98		0.97		0.96		0.98
E	6.63		5.64		6.39		4.94		5.95		5.09

W_m : monolayer water content (kg H₂O (kg d.m)^–1); K: sorption heats constant; C: GAB constant; R²: coefficient of determination; E% : relative deviation modulus.

The monolayer moisture contents (W_m) represent the moisture level corresponding to optimum food stability [5]. Labuza and Ball [35] concluded that W_m > 0.1 kg H₂O (kg d.m)^–1 may compromise food stability. Results showed that for all temperatures, the predicted W_m values were equal or less than that limit (Table 2) therefore the powders had a good stability.

W_m (AAG) values were higher than those from the AMD samples, this could be due to AG higher concentration of hydrophilic groups in comparison with MD10 that favors a greater interaction solute/sorbent [36]. Perez-Alonso et al. [37] and Tonon et al. [38] reported a similar behavior for pure MD10 and AG and spray-dried açai.

Temperature increase reduced monolayer water content from both samples (Table 2), Caballero-Cerón et al. [39] reported a similar effect for dehydrated mango.

3.3 Thermodynamic properties

In accordance with previous publications [40], we detected a strong reduction in Q_st with increasing W_c levels in both samples (Fig. 2a) that were satisfactorily fitted with Equation (5) (R² = 0.99; [19]); regression parameters are presented in Table 3.

Fig.2

Changes of the AAG and AMD sorption heat (a) and differential entropy (b) with equilibrium moisture content. Lines represent the values predicted by Eqns. (5) and (6) respectively.

Table 3

Regression parameters of the Equations (5), (12) and (6) for freeze-dried AAG and AMD

Parameters	AAG	AMD
q₀ (kJ mol^–1)	486.47	192.96
W₀ (kg H₂0 (kg d.m)^–1)	0.02	0.03
R ²	1	0.97
a	0.45	0.26
b	30.28	27.81
R ²	0.96	0.98
T_β (°K)	337.02	342.69
ΔG_β (kJ mol^–1)	1.19	1.02
R ²	1	1

q₀: net isosteric heat of the first water molecule; W₀: first water molecule; R²: coefficient of determination; a/b: constants Equation 12; T_β: Isokinetic temperature; ΔG_β: Free energy at T = T_β.

For all the W_c range tested Q_st (AAG) was higher than Q_st (AMD, indicating that water binding was stronger in the AAG than in the AMD sample (Fig. 2a). This could be attributed to AG higher content of hydrophilic groups capable of forming strong hydrogen bonds than MD10.

The predicted Q_st (kJ mol^–1) results at each W_m (kg H₂O (kg d.m.)^–1) were:

AMD: 73.29 (0.05), 80.41(0.05) and 100.24 (0.03)

AAG: 80.51 (0.06), 94.17 (0.05) and 72.10 (0.04)

The values between parentheses correspond to the monolayer moisture contents (Table 2). Increasing W_c from 0.04 (AAG) or 0.03 (AMD) kg H₂O (kg d.m.)^–1 to 0.07 kg H₂O (kg d.m.)^–1 reduced Q_st of both samples by 43–48%. This behavior was expected since part of this range corresponds to the monolayer moisture contents where the sorbate/sorbent binding energy is very high [18].

A further W_c increment (0.07–0.12 kg H₂O (kg d.m.)^–1) diminished Q_st of both samples, although it was still higher than ΔH_cond, indicating that the sorbate/sorbent energy is greater than that between water molecules favoring multilayer sorption.

At 0.15–0.17 kg H₂O (kg d.m.)^–1, Q_st reached an asymptotic level similar to ΔH_cond indicating that the samples bound water limit will be within this range [17].

The relationship between ΔS_d and W_c in each sample was modeled satisfactorily with Equation (12), (Fig. 2b; Table 3). $S_{d} = a^{(- {bW}_{c})}$ (12)

To analyze the applicability of the isokinetic theory to the samples moisture sorption process Equation (6) was fitted to their respective q_stn and ΔS_d values (Fig. 3; Table 3). AMD and AAG isokinetic temperatures were 342.07±0.45°K (AMD) and 337.02±0.62°K (AAG), both values were different (P < 0.05) from T_hm (295.81°K) satisfying Krug et al. [21] requirements for applying the linear compensation theory. This fact combined with the high linearity degree (R² > 0.99) obtained confirms the existence of q_stn/ΔS_d compensation; hence the isokinetic theory is a valid mean for describing the water sorption mechanism in both samples within the experimental conditions used.

Fig.3

Enthalpy-entropy linear relationship for freeze-dried AAG and AMD samples.

The samples presented only one line of compensation each (Fig. 3), therefore there is no change of mechanism within the experimental conditions used in the current study. Since both isokinetic temperatures were higher than the harmonic mean temperature the moisture sorption is controlled by the energy of interactions related to the chemical composition of the samples [41].

3.4 Temperature, water activity and drying-aid type effects on samples physicochemical properties

3.4.1 Color

Figure 4 (a, b & c) presents the relationship between (a) lightness (L*), (b) saturation index (SI) and (c) hue angle (HA) with the samples a_w at 10, 20 and 40°C. Results showed that using MD10 as the drying-aid had advantages regarding the powder color since within the same T/a_w conditions, AMD had higher (P < 0.05) L*, SI and HA values than AAG, indicating that the former’s color was yellower, lighter and more saturated than the latter’s.

Fig.4

Temperature and water activity influence on Lightness (L*; a), Saturation Index (SI; b) and Hue Angle (HA °; c) of AAG and AMD freeze-dried samples.

At each equilibration temperature, a_w initial increments did not affect the color attributes (P > 0.05); statistical analysis showed that the extension of this period was strongly dependent of temperature and type of drying-aid (P < 0.05). Table 4 presents the water activity ranges for L*, SI and HA highest levels (P > 0.05) and their respective values. Although at 10 and 20°C AMD’s a_w intervals were equal or higher than AAG’s, at 40°C, the difference was not significant (P > 0.05) suggesting that MD10 positive influence could not compensate the temperature increments’ negative effects.

Table 4

Water activity ranges corresponding to Lightness (L*); Saturation Index (SI) and Hue Angle (HA) highest values

Sample	Temperature (°C)	a_w range/color parameters
		a_w	L*	a_w	SI	a_w	HA
AAG	10	0.11–0.23	82.68	0.11–0.54	33.09	0.11–0.33	85.21
	20	0.11–0.23	80.04	0.11–0.43	30.05	0.11–0.33	84.03
	40	0.11–0.23	78.44	0.11–0.43	31.41	0.11	84.58
AMD	10	0.11–0.54	85.47	0.11–0.76	35.88	0.11–0.54	92.20
	20	0.11–0.54	85.02	0.11–0.43	35.24	0.11–0.43	92.43
	40	0.11–0.23	84.07	0.11–0.43	33.53	0.11	88.49

The water activity/temperature combinations necessary for preventing moisture sorption undesirable effects were:

AAG	AMD
•10, 20°C a_w≤0.23	•10°C a_w≤0.57
•40°C a_w≤0.11	•20°C a_w≤0.43
	•40°C a_w≤0.11

Using MD10 instead of AG extended the a_w interval corresponding to the best color attributes from 0.23 (10–20°C) to 0.57 (10°C) or 0.43 (20°C).

3.4.2 Solubility

Results regarding water activity influence on AMD and AAG solubilities in water at 4°C are shown in Fig. 5. Although equilibrating the samples at 10/20°C and a_w = 0.11 did not affect AAG and AMD solubilities (P > 0.05), further a_w enhancements caused a significant loss (P < 0.05).

Fig.5

Temperature and water activity influence on cold-water solubility of AAG and AMD freeze-dried samples.

This behavior can be explained considering that water sorption increases the powders concentration of crystalline particles which are less soluble and have a lower dissolution rate than their amorphous counterparts [42].

Experimental results showed that at 10/20°C and a_w between 0.23–0.76, AMD’s solubility was 5.5–13% higher than that from AAG (P < 0.05), this difference may be attributed to maltodextrin higher solubility when compared to Arabic gum [43].

Increasing the equilibration temperature to 40°C resulted in a 15–46% solubility loss in both products; in these conditions, both samples presented similar values (P > 0.05) indicating that MD10 positive influence was compensated by the temperature increment.

The ranking of decreasing order was: $AMD (10^{\circ} C) \approx AMD (20^{\circ} C) > AAG (10^{\circ} C) \approx AAG (20^{\circ} C) > > AMD (40^{\circ} C) \approx AAG (40^{\circ} C)$

As in color, using MD10 as drying-aid extended the a_w limit for maximum solubility at the 3 equilibration temperatures from 0.11 to 0.33. In these conditions, AMD’s solubilities were 77.68% (10/20°C) and 65.89% (40°C) while the corresponding AAG’s were: 71.64% (10/20°C) and 61.32% (40°C).

The temperature/water activity conditions for preventing moisture sorption effects on the samples solubilities were 10/20°C/0.11.

3.4.3 Antioxidant content

Figures 6 (a–d) show the water activity influence on the TP (a), TF (b), TC (c), and AA (d) levels at 10, 20 and 40°C. Although temperature increments reduced (P < 0.05) AMD’s or AAG’s antioxidants contents, the effect was strongly dependent on the type of drying-aid used (P < 0.05). At 20°C, MD10 over-compensated the negative effect of the temperature enhancement on TP and TC so the order of decreasing stability (Figs. 6a, b) was:

Fig.6

Temperature and water activity effect on total polyphenol (TP, a), total flavonoids (TF, b), total carotenoids (TC, c), ascorbic acid (AA; d) of AAG and AMD freeze-dried samples.

$AMD (10^{\circ} C) > AMD (20^{\circ} C) > AAG (10^{\circ} C) > AAG (20^{\circ} C) > AMD (40^{\circ} C) > AAG (40^{\circ} C)$

In the case of TF (10–40°C) (Fig. 6c), the drying-aid influence followed a slightly different pattern and the stability order was:

$AMD (10^{\circ} C) > AMD (20^{\circ} C) > AAG (10^{\circ} C) > AAG (20^{\circ} C) > AMD (40^{\circ} C) \approx AAG (40^{\circ} C)$

The benefits of encapsulating with AG on the ascorbic acid stability were extremely significant (P < 0.05; Fig. 6d) since the drying-aid*equilibration temperature interaction also changed the stability order at 40°C and the ranking was: $AAG (10^{\circ} C) \approx AAG (20^{\circ} C) > > AAG (40^{\circ} C) > > AMD (10^{\circ} C) \approx AMD (20^{\circ} C) \approx AMD (40^{\circ} C)$

The drying-aid degree of protection varied between the different bioactives, the differences (%) in antioxidant content between AMD and AAG at all temperatures were: 0–44% (TP) and 21–77% (TC) whereas for TF, the range was 14–52%.

Across the whole a_w range, AA(AAG) content was 7–213% (10°C), 126–255% (20°C) and 8–112% (40°C) higher than in AMD. Silva et al. [34] reported a similar drying-aid effect in the AA content of camu-camu spray-dried with AG or MD10.

The temperature/a_w conditions necessary for preventing antioxidant losses produced by moisture sorption in the arazá powders were:

AMD	AAG
•TP 10°C a_w≤0.23	•TP 10°C a_w≤0.11
•TF 10°C a_w≤0.11	•TF 10°C a_w≤0.11
•TC 10°C a_w≤0.33/ 20°C a_w≤0.23	•TC 10°C a_w≤0.11
•AA 10/20°C a_w≤0.11	•AA 10/20°C a_w≤0.11

3.4.4 Antioxidant activity

Figure 7 (a-b) show FRAP and ARA relationship with a_w at the 3 temperatures.

Fig.7

Temperature and water activity effect on the reducing power (FRAP; a) and antiradical activity (ARA; b) of AAG and AMD freeze-dried samples.

Ferric Reducing Antioxidant Power results (Fig. 7a) showed that for a_w = 0.11–0.57, the drying-aid effect and its interaction with temperature were not significant (P > 0.05), however at a_w = 0.75, a strong T*a_w effect was detected (P < 0.05). The FRAP stability order was:

a_w = 0.11–0.57 $AMD (10 ° C) \approx AMD (20 ° C) \approx AAG (10 ° C) \approx AAG (20 ° C) > > AMD (40 ° C) \approx AAG (40 ° C)$

a_w = 0.76 $AMD (10 ° C) \approx AMD (20 ° C) > AAG (10 ° C) \approx AAG (20 ° C) > > AMD (40 ° C) \approx AAG (40 ° C)$

Antiradical activity experimental results (Fig. 7b) showed that although at each temperature, ARA(AMD) was higher than ARA(AAG), this difference was not significant (P > 0.05) except at a_w = 0.54–0.75 and 20°C where AAG’s activity was 11–26% smaller (P < 0.05). Increasing temperature from 10 to 20°C caused a non-significant loss in ARA’s levels in both samples (P > 0.05), in contrast, at 40°C, the temperature effect was extremely detrimental (P < 0.05) since the ARA values across the whole a_w range dropped 50 to 75%.

The stability rankings were:

a_w = 0.11–0.43 $AMD (10 ° C) \approx AAG (10 ° C) \approx AMD (20 ° C) \approx AAG (20 ° C) > > AMD (40 ° C) \approx AAG (40 ° C)$

a_w = 0.54–0.76 $AMD (10 ° C) \approx AAG (10 ° C) \approx AMD (20 ° C) > AAG (20 ° C) > > AMD (40 ° C) \approx AAG (40 ° C)$

In comparison with the non-equilibrated samples, the losses in antioxidant activity due to water sorption across the whole a_w interval were:

ARA	FRAP
AMD/AAG(10°C) 3.5–28%	AMD(10//20°C) 3–24%
AMD(20°C) 10–27%	AAG(10/20°C) 3–48%
AAG(20°C) 10–42%
AMD/AAG(40°C) 51–78%	AMD/AAG(40°C) 49–79%

The advantages of using MD10 as the drying-aid were detectable only at 20°C.

The T/a_w limits for preventing moisture sorption negative effects on the ARA and FRAP levels were 10°C/0.11.

4 Conclusions

The drying-aids type (maltodextrin DE10, Arabic Gum) and concentration (5%) used in the current study allowed the production of powders with good stability.

Temperature, water activity and drying-aid type had a strong impact in the powders color and solubility as well as in their antioxidant content and activity. AMD (10/20°C) had better color and solubility, higher TP/TC/TF and lower AA concentrations than AAG (10/20°C) however, their ARA and FRAP values were similar. These results suggest that using an optimized mix of MD10 and AG as drying-aid may result in enhancing AA concentration and consequently the product activity without affecting its color and solubility.

For obtaining optimum levels of color, solubility, antioxidant content and activity both samples must be kept at temperatures≤10°C and water activities≤0.11.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflict of Interest

The authors have no conflict of interest to report.

Footnotes

Acknowledgments

We thank Eng. Chris Young for assisting with the english translation, Claudio Reyes for the ascorbic acid analysis and Juliana Orjuela-Palacio for her contribution to this work. C.A. Reyes-Álvarez was supported by the Argentinean Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET).

References

de Araújo

, Neri-Numa

, de Paulo Farias

, da Cunha

GRMC

, Pastore

. Wild Brazilian species of Eugenia genera (Myrtaceae) as an innovation hotspot for food and pharmacological purposes. Food Res Int. 2019;121:57–72.

Garzón

, Narváez-Cuenca

C-E

, Kopec

, Barry

, Riedl

, Schwartz

. Determination of carotenoids, total phenolic content, and antioxidant activity of arazá (Eugenia stipitata McVaugh), an amazonian fruit. J Agric Food Chem. 2012;60(18):4709–17.

Marques

, Ferreira

, Freire

. Freeze-drying of acerola (Malpighia glabra L.). Chem Eng Process Process Intensif. 2007;46(5):451–7.

Bhandari

, Howes

. Implication of glass transition for the drying and stability of dried foods. J Food Eng. 1999;40(1-2):71–9.

Orjuela-Palacio

, Lanari

. The impact of moisture sorption properties on the color and bioactives concentrations of black currant-yerba mate instant drinks. J Berry Res. 2016;6(3):303–19.

Vega-Gálvez

, López

, Miranda

, Di Scala

, Yagnam

, Uribe

. Mathematical modelling of moisture sorption isotherms and determination of isosteric heat of blueberry variety O^′Neil. Int J Food Sci Technol. 2009;44(10):2033–41.

Gabas

, Telis

VRN

, Sobral

PJA

, Telis-Romero

. Effect of maltodextrin and arabic gum in water vapor sorption thermodynamic properties of vacuum dried pineapple pulp powder. J Food Eng. 2007;82(2):246–52.

Al-Muhtaseb

, McMinn

WAM

, Magee

TRA

. Water sorption isotherms of starch powders: Part mathematical description of experimental data. J Food Eng. 2004;61(3):297–307.

Martinelli

, Gabas

, Telis-Romero

. Thermodynamic and quality properties of lemon juice powder as affected by maltodextrin and arabic gum. Dry Technol. 2007;25(12):2035–45.

10.

Demarchi

, Quintero Ruiz

, Giner

. Sorptional behaviour of rosehip leather formulations added with sucrose or polydextrose. Biosyst Eng. 2014;118(1):83–94.

11.

Labuza

. Sorption Phenomena in Foods: Theoretical and Practical Aspects. In: Theory, Determination and Control of Physical Properties of Food Materials. Springer, Netherlands; 1975. p. 197–219.

12.

AOAC. Official methods of analysis of AOAC international. 16th. Gaithersburg; 1998.

13.

Al-Muhtaseb

, McMinn

WAM

, Magee

TRA

. Moisture sorption isotherm characteristics of food products: A review. Food Bioprod Process. 2002;80(2):118–28.

14.

Lomauro

, Bakshi

, Lebensmittel-Wissenschaft

. Evaluation of food moisture sorption isotherm equations. Part I. Fruit, vegetable and meat products. LWT – Food Sci Technol. 1985;18(2):111–7.

15.

Lewicki

. The applicability of the GAB model to food water sorption isotherms. Int J Food Sci Technol. 1997;32(6):553–7.

16.

Al-Muhtaseb

, McMinn

, Magee

. Water sorption isotherms of starch powders. Part Thermodynamic characteristics. J Food Eng. 2004;62(2):135–42.

17.

Iglesias

, Chirife

. Isosteric heats of water vapor sorption on dehydrated foods. I. analysis of the differential heat curves. Leb und Technol. 1976;9:116–22.

18.

Telis-Romero

, Kohayakawa

, Silveira

Jr , Pedro

MAM

, Gabas

. Enthalpy-entropy compensation based on isotherms of mango. Ciência e Tecnol Aliment. 2005;25(2):297–303.

19.

Tsami

, Maroulis

, Marinos-Kouris

, Saravacos

. Heat of sorption of water in dried fruits. Int J Food Sci Technol. 1990;25(3):350–9.

20.

Beristain

, Garcia

, Azuara

. Enthalpy-entropy compensation in food vapor adsorption. J Food Eng. 1996;30(3-4):405–15.

21.

Krug

, Hunter

, Grieger

. Enthalpy-entropy compensation. 1. Some fundamental statistical problems associated with the analysis of van’t Hoff and Arrhenius data. J Phys Chem. 1976;80(21):2335–41.

22.

Chen

, Jane

. Properties of granular cold-water-soluble starches prepared by alcoholic-alkaline treatments. Cereal Chem. 1994;71(6):623–6.

23.

Minolta. Precise color communication: color control from perception to instrumentation. Minolta, editor. 1998.

24.

Pérez-Jiménez

, Arranz

, Tabernero

, Díaz-Rubio

, Serrano

, Goñi

, et al. Updated methodology to determine antioxidant capacity in plant foods, oils and beverages: Extraction, measurement and expression of results. Food Res Int. 2008;41(3):274–85.

25.

Saura-Calixto

, Serrano

, Goñi

. Intake and bioaccessibility of total polyphenols in a whole diet. Food Chem. 2007;101(2):492–501.

26.

Schlesier

, Harwat

, Böhm

, Bitsch

. Assessment of antioxidant activity by using different in vitro methods. Free Radic Res. 2002;36(2):177–87.

27.

Chang

, Yang

, Wen

. Estimation of total flavonoid content in propolis by two complementary colorimetric methods. J Food Drug Anal. 2002;10(3).

28.

Ordoñez-Santos

, Hurtado Aguilar

, Ríos Solarte

, Arias Jaramillo

. Concentración de carotenoides totales en residuos de frutas tropicales. Prod+Limpia. 2014;9(1):91–8.

29.

Nováková

, Solich

, Solichová

. HPLC methods for simultaneous determination of ascorbic and dehydroascorbic acids. TrAC Trends Anal Chem. 2008;27(10):942–58.

30.

Brand-Williams

, Cuvelier

, Berset

. Use of a free radical method to evaluate antioxidant activity. LWT – Food Sci Technol. 1995;28(1):25–30.

31.

Pulido

, Bravo

, Saura-Calixto

. Antioxidant activity of dietary polyphenols as determined by a modified ferric reducing/antioxidant power assay. J Agric Food Chem. 2000;48(8):3396–402.

32.

Zambrano-Alcívar

. Elaboración de pulpa a base de arazá (Eugenia Stipitata), utilizando tratamientos térmicos para su conservación natural. Universidad Laica “Eloy Alfaro” de Manabi; 2014.

33.

Ballesteros

, Ramirez

, Orrego

, Teixeira

, Mussatto

. Encapsulation of antioxidant phenolic compounds extracted from spent coffee grounds by freeze-drying and spray-drying using different coating materials. Food Chem. 2017;237:623–31.

34.

Silva

, Cornejo

, Gomes

, Pontes

, Matta

, Freitas

. Influence of shell material on vitamin c content, total phenolic compounds, sorption isotherms and particle size of spray-dried camu-camu juice. Fruits. 2013;68(3):175–83.

35.

Labuza

, Ball

. Moisture sorption: Practical aspects of isotherm measurement and use. 2nd ed. Minessota: American Association of Cereal Chemists; 2002.

36.

Silva

, Kurozawa

, Park

, Hubinger

. Water sorption and glass transition temperature of spray-dried mussel meat protein hydrolysate. Dry Technol. 2012;30(2):175–84.

37.

Pérez-Alonso

, Beristain

, Lobato-Calleros

, Rodríguez-Huezo

, Vernon-Carter

. Thermodynamic analysis of the sorption isotherms of pure and blended carbohydrate polymers. J Food Eng. 2006;77(4):753–60.

38.

Tonon

, Baroni

, Brabet

, Gibert

, Pallet

, Hubinger

. Water sorption and glass transition temperature of spray dried açai (Euterpe oleracea Mart.) juice. J Food Eng. 2009;94(3-4):215–21.

39.

Caballero-Cerón

, Guerrero-Beltrán

, Mújica-Paz

, Torres

, Welti-Chanes

. Moisture Sorption Isotherms of Foods: Experimental Methodology, Mathematical Analysis, and Practical Applications. In Springer, New York, NY; 2015. pp. 187–214.

40.

Vega-Gálvez

, Lemus-Mondaca

, Tello-Ireland

, Miranda

, Yagnam

. Kinetic study of convective drying of blueberry variety o’ neil (Vaccinium corymbosum L.). Chil J Agrigultural Res. 2009;69(2):171–8.

41.

Leffler

. The enthalpy-entropy relationship and its implications for organic chemistry. J Org Chem. 1955;20(9):1202–31.

42.

. Amorphous pharmaceutical solids: Preparation, characterization and stabilization. Int. J. Pharm. 2001;48:27–42.

43.

Cano-Chauca

, Stringheta

, Ramos

, Cal-Vidal

. Effect of carriers on the microstructure of mango powder obtained by spray drying and its functional characterization. Innov Food Sci Emerg Technol. 2005;6(4):420–8.