State Diagram of Spray Dried Bovine Colostrum Powder

Abstract

The state diagram of spray dried bovine colostrum powder was constructed by determining its freezing curve (initial freezing point versus solids content), the glass line (glass transition temperature versus solids content), maximal-freeze-concentration conditions [end point of freezing (), and glass transition temperature of maximally-freeze-concentrated solution ()] by using differential scanning calorimetry. The freezing curve and the glass line were modeled using the Chen equation by incorporating concept of unfrozen water and Gordon-Taylor equation, respectively. The conditions of the maximal-freeze-concentration obtained were −32.14°C () and −50.52°C (), respectively, and the corresponding solids contents were 0.807 and 0.839 kg dry solids/kg samples, respectively. The state diagram of spray dried bovine colostrum powder obtained can be useful for predicting its stability during storage as a function of temperature and moisture content.

Keywords:

• State diagram
• Bovine colostrum
• Glass transition temperature
• Freezing curve

INTRODUCTION

Bovine colostrum is the initial milk secreted by a cow during the first 4 days postparturition.^[1] Colostrum-based products are commercially available as a health food supplement in Australia, New Zealand, USA, Europe, and China,^[2] because bovine colostrum contains abundant bioactive components, including growth factors, immunoglobulins (Igs), lactoperoxidase, lysozyme, lactoferrin, cytokines, nucleosides, vitamins, peptides, and oligosaccharides, which are of increasing relevance to human health.^[3] In addition, colostrum-based products are marketed as a general “health promoting” product, particularly suitable for athletes.^[4] Drying is a traditional method for preservation of foods, which can reduce transportation and storage costs and extend the shelf life. Spray drying and freeze drying are commonly employed methods for dehydrating liquid bovine colostrum into powder.^[5] During the processing of bovine colostrum and storage of bovine colostrum powder, choosing optimum processing parameters and storage conditions is of great importance to the dairy industry.

A state diagram of food presents different physical states of food as a function of solids content and temperature and a simplified state diagram of food consists of a freezing curve of initial freezing point versus solids content, a glass line of glass transition temperature (T_g) versus solids content, and conditions of maximal-freeze-concentration,^[6] which is a helpful tool for determining its processing and storage stability. In the last decades, the state diagrams of food materials, including pure components and real foods, have been extensively studied.^[7−12] In the literature, it has been reported that foods can be considered very stable at the glassy state, since below glass temperature compounds involved in deterioration reactions take many months or even years to diffuse over molecular distances and approach each other to react.^[13] Formation of a glassy state results in a significant arrest of translational molecular motion, and chemical reactions become very slow.^[14] The glassy state concept and water activity concept have been widely tested in food as evident from the literature and a detailed review about the relationships between glass temperature and water activity and stability of food materials have been given by Rahman.^[15−17] The glass line (glass transition temperature versus total solids content) and maximal-freeze-concentration conditions of skim milk and whole milk with different fat contents have been reported in the literature.^[18] In addition, the glass transition lines of whole milk powder and skim milk powder and a universal relationship between the glass transition temperature and water activity have been reviewed and the glass line has been found to be a new tool for optimizing the drying process of milk powders.^[19] However, the comprehensive data about the state diagram of bovine colostrum, including the freezing curve, the glass line, and maximal-freeze-concentration conditions, have not been reported in the literature.

The objectives of this study were to construct a state diagram of spray dried bovine colostrum powder by determining the glass line (glass transition temperature (T_g) versus total solids content), the freezing curve (freezing point versus total solids content), and maximal-freeze-concentration conditions (end point of freezing for maximal-freeze-concentration condition , glass transition temperature of maximal-freeze-concentration solution ) and to predict the optimum storage conditions of spray dried bovine colostrum (SDBC) powder according to the estibilished state diagram of bovine colostrum.

MATERIALS AND METHOD

Materials and Samples Handling

Fresh bovine colostrums used in this experiment were provided by a dairy farm of Bright Dairy and Food Co., Ltd, Shanghai, China, tranferred to our lab, and then stored at −20°C for 2 to 4 weeks prior to use in this study. Frozen bovine colostrum was thawed at room temperature, filtrated to remove non-milk ingredients by medical gauzes, and defatted by a SE 02.0V centrifugal separator (SEITAL S.r.l., Santorso (VI), Italy). The samples used in this experiment were employed with a constant total solids mass concentration of 9.59 ± 0.01, containing fat of 0.74 ± 0.02, protein of 4.93 ± 0.01, lactose of 2.96 ± 0.01, and ash of 0.956 ± 0.002 g/100 g samples, respectively. Bovine colostrum powders were obtained by spray drying fresh bovine colostrums in a mini–scale spray dryer (Shandong Tianli Drying Equipment Inc., Jinan, Shandong Province, China) and the spray-drying conditions were the atomizer pressure of 0.7 ± 0.01 MPa, the feed temperature of 32 ± 1°C, the feed rate of 160 ± 20 mL/h, air inlet temperature of 125 ± 1°C, drying air flow rate of 0.78 ± 0.01 m³/min, and compressed air flow rate of 300 ± 10 L/h. SDBC powders collected from the spray dryer were kept immediately in vacuum desiccators over P₂O₅ at room temperature. Desiccated powders were further dried in a vacuum oven at 30°C for 5–7 days before experiments so as to remove residual moisture.

For the adsorption experiment, three replications of SDBC powder samples (1 ± 0.001) g were placed in previously weighed glass sorption jars, and then quickly placed inside desiccators over eight saturated salt solutions of known relative humidity.^[20] The saturated salts used were as follows: KOH, LiCl, CH₃COOK, MgCl₂, K₂CO₃, Mg(NO₃)₂, NaNO₂, and NaCl with equilibrium relative humidities of 8.2, 11.3, 22.5, 32.8, 43.2, 52.9, 65.8, and 75.0%. A test tube containing thymol was placed inside the desiccators with high relative humidity to prevent mold growth during storage. The desiccators were placed in a temperature-controlled cabinet, with an accuracy of ± 1°C at the selected temperatures of 25°C. The samples were weighed at intervals of 72 h. Equilibrium was reached when the sample weight difference between two successive measurements was less than the balance accuracy of 0.001 g. This involved a period of 21–24 days. The total time for removal, weighing, and putting back the samples in the desiccators was about 30 s. This minimized the degree of atmospheric moisture sorption during weighing. The equilibrium moisture contents of the samples were determined using AOAC method 927.05.^[21]

Determination of Thermal Transition

The thermal transition experiments in all the samples were conducted with a differential scanning calorimeter (DSC) (Diamond DSC, PerkinElmer, Boston, MA, USA) in inert atmosphere (20 mL/min of N₂). The DSC was calibrated for temperature and heat flow using n-hexane (melting point, −95.6°C; latent heat of melting, 140.16 J/g), double-distilled water (melting point, 0°C; latent heat of melting, 333.88 J/g), and indium (melting point, 156.6°C; latent heat of melting, 28.45 J/g). An empty sealed aluminium pan was used as a reference in each test. Liquid nitrogen was used for sample cooling before the run.

The samples (10–20 mg) containing unfrozen water, reaching equilibration in different saturated salt solutions mentioned above, were transferred in pre-weighed differential scanning calorimeter aluminium pans (volume 30 μL), hermetically sealed, then cooled from room temperature to −100° at 20°C/min, equilibrated for 10 min and scanned from -100 to 150°C at 10°C/min according to the recommendation by Sablani and his co-authers.^[22] The glass transition temperature (T_g) is identified as a (vertical) shift in the heat flow curve of the thermogram. Pyris Diamond DSC analysis software (PerkinElmer, Boston, MA, USA) was used to analyze the onset, mid, and end-points of the glass transition. Three replicates were used for the determination of glass transition temperature at each water content/water activity.

The samples containing freezable water with different moisture contents were scanned from −100 to 20°C at 10°C/min to initially locate the T_g of the maximally freeze-concentrated solutes () and onset temperature of ice melting within the maximally freeze-concentrated solutes (). Triplicate samples (10–20 mg) of each concentration (0.198–0.660 kg dry solids/kg samples) were cooled to −100°C, equilibrated for 5 min, heated at 10°C/min to −33°C ( - 1°C), annealed for 30 min to allow ice formation, re-cooled to −100°C at 20°C/min, equilibrated for 5 min and then scanned from −100 to 20°C at 10°C/min to determine the freezing point, enthalpy of ice melting, and . Three transitions (peak, maximum slope, and start of melting) were determined from the ice melting endotherm. The initial or equilibrium freezing point was considered as the point of maximum slope of the endothermic peak.^[6] The end point of freezing () is taken as the initial point of ice melting at endothermic peak.^[7,17] All the results were given as an average for three samples.

Modeling of Water Adsorption Isotherm and Thermal Transition

Many theoretical, empirical and semi-empirical equations are available for modeling of sorption isotherm data.^[23] In this study, water adsorption data of SDBC powders were modeling using most commonly employed Guggenheim-Andersen-de Boer (GAB) and Brunauer-Emmett-Teller (BET) equations.^[11] GAB and BET equations have a sound theoretical background and their parameters provide physical meaning related to the sorption process. The GAB and BET equations are listed as follows:

(1)

(2)

where M represents the equilibrium moisture content (EMC) (g H₂O/100 g dry solids), a_w represents water activity (decimal), M₀ is monolayer moisture content (g H₂O/100 g dry solids), and C, K are coefficients of models. In order to evaluate how accurate each model was in fitting experimental data, the mean relative percentage deviation modulus (MRE) value, the coefficient of determination (R²), root mean square error (RMSE), and the graph of residuals were most commonly applied for the criteria.^[23] It is generally considered that MRE values and RMSE below 10%^[24,25] and R² > 0.9 indicate an adequate fit for practical purposes.

The glass transition temperatures of amorphous foods are influenced by water content. The influence of water content on glass transition temperature is commonly modeled by the Gordon and Taylor (G-T) equation:^[26]

(3)

where T_gm, T_gs, and T_gw are the glass transition temperatures of the mixture, solids, and water, respectively; X_w and X_s are the mass fraction of water (kg H₂O/kg samples) and total solids (kg dry solids/kg samples), and k is the G-T model parameter. From the thermodynamic standpoint, the k parameter is equivalent to the ratio of the change in specific heat of the components of the mixture at their T_g. The model parameters (k and T_gs) of Eq. (3) were estimated using nonlinear regression analysis while considering T_gw as −135°C.

The Clausius-Clapeyron equation was used to model the freezing curve of bovine colostrum with change in water content. The Clausius-Clapeyron equation is expressed as:^[11]

(4)

where δ is the freezing point depression (T_w - ) relative to increasing total solid content, is the freezing point of the food material (°C), T_w is the freezing point of water (°C), β is the molar freezing point constant of water (1860 kg °C/(kg mol)), λ_w is the molecular mass of water, X_s is the solid mass fraction (kg dry solids/kg samples), and E is the molecular mass ratio of water to solids (λ_w/λ_s). Use of Clausius-Clapeyron equation is limited to ideal and dilute solutions. The Chen model is an extension of Clausius-Clapeyron equation by the introduction of a new parameter B, which is the ratio of unfreezable water of the total solid content. The Chen model is expressed as:^[27]

(5)

In this study, the Chen model was selected to fit the freezing point data.

Statistical Analysis

A nonlinear regression analysis by the least square method was carried out to estimate the parameters of Eqs. (1), (2), (3), and (5) using the SAS9.2 software program (SAS Institute Inc., Cary, NC, USA).

TABLE 1 Model parameters for BET and GAB equations

	Models
Model parameters	GAB	BET^a
M0 (g H2O/100 g dry solids)	6.189	5.142
C	16.445	24.935
K	0.845
MRE (%)	3.25	0.92
RMSE (%)	3.94	1.00
R^2	0.995	0.999
Residual plots	Random	Random

^aOnly experimental data with a_w < 0.50 were fitted to the BET model.

RESULTS AND DISCUSSION

Sorption Isotherm of SDBC Powder

The water adsorption isotherm of SDBC powder at 25°C was presented in Fig. 1 and showed a sigmoid shape, classified to a typical type II curve. The sigmoid shape of sorption isotherm was commonly found in many food materials containing amorphous solids.^[28] Similar results were reported in other dairy products, including powdered milk^[29] and yogurt powder.^[30] The sorption isotherm data were modeled using BET and GAB equations, and parameters for two models were shown in Table 1. It was clearly shown that BET and GAB equations fitted well with experimental data in the water activity range of 0.08–0.43 and 0.08–0.75, respectively, based on MRE and RMSE below 5%, R² > 0.9, and random distribution of residual plots. Similarly, BET and GAB models were reported to be suitable for describing sorption isotherm of other dairy products.^[18,29,30] The monolayer moisture content, M₀, obtained from BET model, was recognized as BET monolayer moisture content, and it has been found to be a reasonable guide to various aspects of interest in dried foods, related to the limit from which the rate of deteriorative reactions began to increase significantly. The limitation of BET model is that it is not able to predict sorption behavior accurately when a_w is beyond 0.5.^[23] In this study, M₀ of BET equation at 25°C were 5.142 g H₂O/100 g dry solids, comparable to 5.5 g H₂O/100 g dry solids at 24°C for freeze dried skim milk powder^[18] and higher than 4.5 g H₂O/100 g dry solids for whole milk powder at 25°C.^[19] The GAB model was widely recommended to fit well with many food materials, since it was particularly successful for water activities up to 0.9. The monolayer moisture content, M₀, obtained from GAB models at 25°C, was 6.189 g H₂O/100 g dry solids, higher than that of freeze dried skim milk powder at 25°C,^[18] which was probably attributed to the difference of the milk components. K and C values obtained from GAB equation belonged to the ranges of 1–20 and 0.7–1, respectively, which was commonly reported in many food materials.^[11]

FIGURE 1 Water adsorption isotherm of SDBC powder at 25°C.

Thermal Transition of SDBC Powder Containing Unfreezable Water

The glass transition temperatures of foods depend mainly on the quantity of water, constituents, and molecular weight of solutes present in the food. The glass transition temperature in foods are not sharp but occur over a range of temperatures.^[17] The T_gi and T_gm of glass transition region determined from thermograms were commonly employed in the literature.^[9] All the thermograms of samples with moisture content less than 0.14 kg H₂O/kg samples showed only one transition and no formation of ice. Due to the effects of fat melting on thermograms,^[18,19] the glass transition temperatures of the samples reaching the equilibrium moisture content in saturated salt solutions with a_w (0.225 and 0.328) were difficult to detect. The T_gi and T_gm of SDBC powders containing different unfreezable water contents were listed in Table 2. Obviously, T_g values decreased with increasing water content due to the plasticization effect of water on the amorphous constituents of the matrix. The T_gi values in Table 2 were fitted using the G-T equation and the values of T_gs and k were estimated for 102.5°C and 9.44, respectively. The estimated T_gs in this study was close to T_g of lactose (101°C),^[18] comparable to T_g of whole milk powder (100.6°C).^[19]

TABLE 2 Glass transition temperature of SDBC powder as a function of moisture content (when there is no formation ice)

aw	Xs (kg solids/kg samples)	Xw (kg H2O/kg samples)	Tgi (°C)	Tgm (°C)
0.0823	0.962	0.038	34.87 (0.95)^a	39.98 (1.39)
0.1130	0.958	0.042	34.73 (1.01)	38.59 (1.44)
0.4316	0.920	0.080	−1.18 (0.53)	1.26 (0.52)
0.5289	0.902	0.098	−11.76 (3.07)	−9.04 (2.42)
0.6500	0.883	0.117	−33.88 (1.87)	−28.68 (2.65)
0.7529	0.860	0.140	−45.20 (2.98)	−39.64 (4.24)

^aValues in the parentheses are standard deviations of three samples.

Thermal Transition of Bovine Colostrum Containing Freezable Water

It has been verified that annealing is necessary to allow the formation of maximum amount of ice and leads to a maximally-freeze-concentrated solid matrix.^[6] An annealing time of 30 min was taken as optimal for further analysis since annealing was able to eliminate devitrification of amorphous solids. A typical DSC thermogram of the sample containing freezable water with 30 min annealing was shown in Fig. 2 and the three characteristic temperatures (onset as , maximum slope of heat flow as , and peak as ),^[31,32] end point of freezing (), and glass transition temperature of the maximally-freeze-concentrated solution () could be determined from Fig. 2. The three characteristic temperatures, end point of freezing (), and enthalpy of ice melting (△H_m) of bovine colostrums with different freezable moisture contents were obtained and listed in Table 3. As expected, the decreased with increasing solids content and the magnitude of freezing point temperature depression depended on the molecular weight of the solids. Accordingly, the molecular weights of food materials were estimated using this technique.^[27] The - X_s data in Table 3 were modeled using the Chen model, and E and B values were estimated as 0.0338 and 0.1464, respectively. According to the E value, the molecular weight of bovine colostrum was evaluated for 532 g/mol, which was higher than that of skim milk.^[27] The enthalpy of ice melting was plotted against water content to determine the quantity of unfreezable water. A linear relationship was developed between enthalpy and water content data for bovine colostrum and the results were presented in Fig. 3. The amount of unfreezable water was 0.216 kg H₂O/kg samples by extending the line to △H_m equal to zero.

TABLE 3 Freezing point temperature and maximal-freeze-concentration conditions of bovine colostrum as a function of initial solid concentration

Xs (kg solids/kg samples)	(°C)	(°C)	(°C)	(°C)	(°C)	△Hm (kJ/kg)
0.198	2.13 (0.58)	−0.45 (0.15)	−3.26 (0.63)	−31.32 (1.11)^n	−50.48 (1.21)	257.81 (4.66)
0.279	0.59 (0.82)	−0.84 (0.07)	−4.26 (0.17)	−32.76 (0.94)	−48.95 (0.98)	221.63 (4.30)
0.399	−1.68 (0.56)	−3.15 (0.25)	−5.65 (1.08)	−31.70 (0.85)	−50.39 (0.85)	167.48 (6.35)
0.501	−2.68 (0.31)	−3.75 (0.33)	−6.21 (0.25)	−32.24 (0.79)	−50.38 (1.02)	109.48 (13.18)
0.591	−3.55 (1.1)	−6.49 (2.81)	−14.94 (1.13)	−32.78 (1.18)	−48.90 (0.56)	82.84 (5.84)
0.660	−5.77 (0.42)	−8.95 (2.12)	−15.38 (0.58)	−32.43 (0.16)	−51.76 (0.85)	63.29 (0.18)

ⁿValues in the parentheses are standard deviations of three samples.

^aPeak of ice melting endotherm.

^bMaximum slope of ice melting endotherm.

^cOnset of ice melting endotherm.

FIGURE 2 A typical DSC thermogram of bovine colostrum containing freezable water (0.501 kg solids/kg samples) for 30 min annealing (, , and are the peak, maximum slope, and start of ice melting endotherm, and is the end point of freezing for maximal-freeze-concentration condition).

FIGURE 3 Enthalpy change of ice melting as a function of water content.

In general, the and are independent of the initial solids concentration of food materials,^[18] but influenced by the molecular weight of total solids present in food materials. The - X_s and - X_s data were listed in Table 3 and statistical analysis showed that there was no significant difference between the and and the initial solids concentration of bovine colostrum. The and of bovine colostrum were −32.14 and −50.52°C, respectively, which were comparable to −32 and −50°C of skim milk obtained by Jouppila and Roos.^[18]

State Diagram of Spray Dried Bovine Colostrum

The simplified state diagram of spray dried bovine colostrum was constructed by the freezing curve (AB), the glass line (CDE), and maximal-freeze-concentration conditions (points F and G) and is presented in Fig. 4. The stability and shelf life of low moisture and frozen food materials can be evaluated through state diagram. The freezing curve (AB) was modeled using the Chen equation. The B value obtained from nonlinear regression represented the unfreezable water content and the B value in this study corresponded to the unfreezable water of 0.1277 kg H₂O/kg samples, which was lower than 0.216 kg H₂O/kg samples obtained from enthalpy of ice melting. The value of (total solids content corresponding to ) was determined by extending the freezing curve into −32.14°C (point F or B). The estimated value and the moisture content corresponding to were calculated as 0.807 kg solids/kg samples and 0.193 kg H₂O/kg samples, respectively. , also considered as the unfreezable water, was higher than the B value. To identify the glass transition temperature corresponding to the maximally-freeze-concentrated liquid bovine colostrum (), the value of was calculated by incorporating into the glass line (point G) modeled using the G-T equation. The was estimated for -50.52°C and was 0.839 kg solids/kg samples. The value of is of great importance for frozen foods to design the optimal storage conditions because the rates of diffusion controlled reactions in frozen foods may decrease considerably at temperatures less than . Point H was determined from the intersection point of the glass line and a vertical line passing through point B () and the temperature corresponding to point H was recognized as the theoretical glass transition temperature () of maximally-freeze-concentrated solution and calculated as -62.12°C based on the predicted G-T model, which was lower than . In theory, the value of (total solids content corresponding to ) is equivalent to the value of (total solids content corresponding to ).^[16] However, in practice, a difference between the value of and the value of commonly occurred. Point D was defined as and as the intersection of the freezing curve to the glass line by maintaining the similar curvature of the freezing curve (line AB).^[31] The values of and were estimated as −52.35°C and 0.835 kg solids/kg samples, respectively.

FIGURE 4 State diagram of spray dried bovine colostrum.

Assessing Water Activity and Glass Transition Concepts for Food Stability

The concept of water activity is an important tool in predicting microbial growth, enzymatic and non-enzymatic activities, and other deteriorative reactions in foods.^[33] According to water activity concept, a food product is most stable at its monolayer moisture content, which varies with the chemical composition, structure and environment conditions. Recently, the glass transition concept was also recommended to explain selected reaction kinetics in foods during production and storage.^[17] However, these two concepts mentioned above have their limitations.^[16,17] Therefore, this is now being recommended to combine the water activity concept with the glass transition concept in assessing process-ability, deterioration, food stability, and shelf life predictions. In this study, storage stability of SDBC powder was evaluated based on the glass transition temperature and the adsorption isotherm and presented in Fig. 5. In this case, if the product is commercialized at room temperature (25°C), the critical moisture content (CMC) of SDBC powder that takes the product from the glassy to rubbery state is 5.129 g H₂O/100 g dry solids and the corresponding critical a_w is 0.158. This means that the SDBC powder is apt to become the rubbery state when stored at the environment relative humidity (RH) over 15.8%. On the other hand, the SDBC powder will lead to the rubbery state when the moisture content of SDBC powder is over 5.129 g H₂O/100 g dry solids. The estimated CMC was much lower than M₀, obtained from GAB equation, meaning that the glass transition concept always underestimates CMC, compared to the water activity concept, which was in line with other findings.^[12,34,35]

FIGURE 5 Relationships between glass transition temperature, water activity, and water content of SDBC powder.

The monolayer moisture content, M₀, obtained from GAB equation was 6.189 g H₂O/100 g dry solids. As can be observed from Fig. 5, the corresponding T_g and was 14.89°C and 0.234, respectively. It implied that according to the glass transition concept, the safe storage or processing temperature selected for SDBC powder with moisture content of 6.189 g H₂O/100 g dry solids was below 14.89°C in order to prevent the samples from caking or collapse. From this result, it can be deduced that the state diagram of a food is useful for predicting the optimum storage or processing conditions.

CONCLUSIONS

The Chen and G-T models were suitable for fitting the freezing curve and the glass line, respectively. The estimated parameters E and B of the Chen model were 0.0338 and 0.1277 kg H₂O/kg samples, respectively. The model parameters T_gs and k of the G-T model were estimated for 102.5°C and 9.44, respectively. The and of maximal-freeze-concentration conditions were −32.14 and −50.52°C, respectively, and the corresponding and were 0.807 and 0.839 kg solids of kg samples, respectively. The state diagram of spray dried bovine colostrum powder was constructed by determining the freezing curve, the glass line, and the maximal-freeze-concentration conditions. The water sorption data provided the monolayer moisture content values of 6.189 g H₂O/100 g dry solids and 5.142 g H₂O/100 g dry solids in GAB and BET models, respectively. Based on the state diagram and sorption isotherm, the estimated critical water activity and critical moisture content were 5.129 g H₂O/100 g dry solids and 0.158, respectively, when the storage temperature of SDBC powder was selected as room temperature (25°C).

FUNDING

The authors gratefully acknowledge the National Department of Science and Technology of China under the National Key Technology R&D Program (2006BAD04A14) for providing the financial support.