Journal of Food Engineering 149 (2015) 153–158
Contents lists available at ScienceDirect
Journal of Food Engineering journal homepage: www.elsevier.com/locate/jfoodeng
Partitioning of calcium and magnesium (total divalent cations) during membrane filtration of milk M.-J. Lin, A.S. Grandison, M.J. Lewis ⇑ Department of Food and Nutritional Sciences, University of Reading, PO Box 226, Whiteknights, Reading RG6 6AP, UK
a r t i c l e
i n f o
Article history: Received 28 February 2013 Received in revised form 13 October 2014 Accepted 15 October 2014 Available online 22 October 2014 Keywords: Ultrafiltration Nanofiltration Reverse osmosis Total divalent cations Ionic calcium
a b s t r a c t Partitioning of total divalent cations (TDVC) during reverse osmosis (RO), nanofiltration (NF) and ultrafiltration (UF) has been investigated. During RO, there was an increase in TDVC and Ca2+, and a reduction in the ethanol stability of RO retentates. During UF of milk at its normal pH, there was an increase in total divalent cations, but only a slight increase in Ca2+ in the retentate. There was some loss of micellar calcium during UF. However, the ratio of amounts of soluble to total divalent cations in the retentate decreased as concentration factor increased. During NF, a small amount of TDVC was found in the permeate and TDVC rejection was estimated to be about 0.83. During UF of milk, the amount of TDVC in permeate increased significantly as the pH was reduced over the range 6.7–5.1 and the concentration of Ca2+ also increased both in the retentate and the permeate. However, this was not reversible, as when milk was restored to its original pH, its Ca2+ remained higher and ethanol stability was lower. In contrast, for whey and for UF permeate, changes in Ca2+ were reversible, when they were subject to similar pH changes. During UF, TDVC concentration in permeate decreased as temperature increased, due to the lower solubility of calcium phosphate at higher temperature. Ca2+ in permeate also decreased as UF temperature increased. Ó 2014 Published by Elsevier Ltd.
1. Introduction The main membrane techniques for processing milk are reverse osmosis (RO), nanofiltration (NF), ultrafiltration (UF) and microfiltration (MF). Factors affecting permeate rates and reducing concentration polarisation and membrane fouling have been well researched (Renner and AbD El-Salam, 1991; Field et al., 1995; Youravong et al., 2003). UF has been used for determining the amounts of calcium, magnesium and phosphorus, which are not associated with the casein micelle (Davies and White, 1960). At the normal pH of milk, approximately 30% of the calcium, 65% of the magnesium and 45% of the phosphorus is in the diffusible form. These increase as the pH is reduced, by whatever means. Glover (1985) reported that the ratio of soluble to total divalent cations (TDVC) decreased from 29% in raw milk down to 7% in milk concentrated 5-fold by UF. Changes in pH have been found to influence the amount of minerals in the final retentate during UF of whey and buttermilk (Hiddink et al., 1978). It was observed that maximum mineral removal was obtained by UF at pH 6.6, followed by diafiltration at pH 3–3.5. It has been observed that the rejection
⇑ Corresponding author. E-mail address:
[email protected] (M.J. Lewis). http://dx.doi.org/10.1016/j.jfoodeng.2014.10.018 0260-8774/Ó 2014 Published by Elsevier Ltd.
of calcium, sodium and phosphorus was higher during diafiltration than UF and that diafiltration of acidified milk gave rise to lower rejections of calcium, phosphorus and sodium (Bastian et al., 1991). Calcium and P recovery in UF concentrates (CF = 5) were 84% for calcium and 66% for P. These were reduced by diafiltration and UF of acidified milk. Premaratne and Cousin (1991) reported the following concentration factors for some different divalent and trivalent cations resulting from a five-fold concentration of milk by UF: Zn (4.9), Fe (4.9), Cu (4.7), Ca (4.3), Mg (4.0) and Mn (3.0). This suggested that there were differences in their binding capacities to casein and whey proteins. Retentates produced by UF of skim milk were able to withstand sterilisation at 120 °C for 7 min (Sweetsur and Muir, 1985). The heat stability was improved by procedures which reduced the levels of salts in the retentates. In milk concentrated two-fold by UF, it was noted that Ca2+ in the retentate immediately after production was slightly lower than in the original milk, but increased slightly during storage, by up to 15%. Ca2+ in permeate was reported to be only one third of that in the milk (May and Smith, 1998). Partitioning of minerals in milk when pH is reduced has been well researched (Holt et al., 1981; Holt, 2004). Both pH reduction and heating will result in movement of calcium and P between the micelle and the soluble phase. The reversibility of this move-
154
M.-J. Lin et al. / Journal of Food Engineering 149 (2015) 153–158
ment and how it might influence the stability of the casein micelle has been less studied. Ultrafiltration at high temperatures has not been studied in depth. Rose and Tessier (1959) and Pouliot et al. (1989a,b,c) used UF to look at mineral partitioning at high temperature and found that soluble calcium decreased as temperature increased. Sood and Kosikowski (1979) reported that UF at 60 °C resulted in a higher permeate rate than at 50 °C. This paper investigates factors affecting calcium losses of TDVC from UF retentates during membrane processing, and how TDVC and Ca2+ in UF permeate are influenced by pH and temperature. 2. Materials and methods Raw milk was obtained from The Centre for Dairy Research, University of Reading (CEDAR). It was centrifuged using a separator (R.A. Lister and Co., Ltd., Dursley, UK) to produce skim milk which was stored at 4 °C prior to concentration by a variety of membrane processes. 2.1. Membrane processing 2.1.1. RO and UF Skim milk was ultrafiltered using an Aquious PCI (Hants, UK) tubular system, surface area 0.8 m2, fitted with ES625 membranes (MWCO, 25 kDa,) at 50 °C. The inlet and outlet pressures were 6 bar and 2 bar, respectively. Samples were taken from the retentate when concentration factors reached 1.25, 1.75 and 2.65. Skim milk was concentrated with an Aquious PCI RO tubular unit at 50 °C, fitted with AFC99 membranes. Its surface area was 2.6 m2 and the inlet pressure was 40 bar. Samples were taken from the retentate and permeate when concentration factors reached 1.25, 1.50, 1.75, 2.00 and 2.40. 2.1.2. Combined UF and NF Skim milk was UF treated using the PCI tubular Module, at conditions described earlier, until the solids content in the milk retentate increased by a factor of about 4. TDVC and Ca2+ concentrations were measured in both retentates and permeates at different concentration factors. The UF permeate was collected and nanofiltered the following day at 20 °C–5.6-fold total solids concentration in NF retentate, using PCI AFC30 NF membranes of surface area 0.8 m2 at a pressure of 25 bar. 2.1.3. Effect of pH on calcium loss in UF-reversibility studies Two batches of skim milk were ultrafiltered by the PCI tubular module in the constant recycle mode. In the first batch, UF commenced at 7 °C and increased to 25 °C, while in the second batch a constant temperature of 30 °C was maintained throughout. During the UF process, milk pH was reduced in stepwise fashion by adding 5 M HCl and then raised by adding 5 M NaOH back to the original milk pH level. Samples of milk retentate and permeate (25 mL) were collected for measuring TDVC and Ca2+ concentrations at different pH values. After adjusting pH by adding either HCl or NaOH, permeates were collected after 5 and 10 min at each pH, as well as after 30 min for both lowest and highest pH levels. TDVC concentration in permeate and pH value in milk were measured to observe how quickly they came to equilibrium after pH adjustment. To understand more clearly the reversibility of movement of TDVC when altering pH, the pH of milk was adjusted in stepwise fashion during UF at 30 °C to pH 5.40 with 5 M HCl and then adjusted back to its original pH with 5 M NaOH Ca2+ and ethanol stability were measured. For comparison with skim milk, samples of cheese whey and UF permeate were subjected to the same changes.
2.1.4. Effect of temperature on calcium loss during UF Skim milk was ultrafiltered in the constant recycle mode using the PCI module. Temperature was changed throughout the process by means of a heat exchanger in the circuit; ranges used were 5– 25 °C, 26.5–55.5 °C and 60–10 °C, over a period of about 90 min. Approximately 25 mL samples of milk, retentate and permeate were taken at various temperatures from the system for measuring both TDVC and Ca2+ concentrations (at 20 °C). 2.1.5. Analytical procedures Retentates and permeates were analysed for TDVC and Ca2+. TDVC concentration was shown as the average of three measurements by the EDTA titration method. The procedure involved titrating 5 mL milk, 1 mL ammonia buffer solution (7 g ammonium chloride and 25 g ammonia solution, specific gravity 0.88, made up to 100 mL with distilled water) and 0.02 mL calmagite indicator against 0.01M EDTA solution until the colour of milk changed from pink to blue (On-Nom et al., 2010). Ca2+ concentration was measured using a Ciba Corning 634 ISE 2+ Ca analyser (Siemens Diagnostic, Newbury, UK) (Lin et al., 2006a). The electrode was calibrated using five Ca2+ standards on a daily basis. A Ciba Corning 250 analyser combined with a Patterson calcium directION flow cell (Patterson Scientific Ltd., Luton, UK) was used for acidified samples, as they had higher Ca2+ values. Ethanol stability was determined by mixing milk samples with equal volumes of different strength ethanol solutions, at 1% intervals. The ethanol stability was recorded as the highest concentration which just failed to cause the milk to coagulate. 3. Results 3.1. Comparison of UF and RO retentates Results are shown for how various milk properties change as milk is concentrated by RO and UF in Table 1. TDVC increased in a linear fashion with concentration factors for both UF and RO retentates (R2 = 0.964 for UF). During RO, the Ca2+ concentration in milk increased by a factor of 1.31, from 1.74 mM to 2.28 mM at concentration factor 2.40. Ethanol stability decreased from 82% to 68% at a concentration factor 1.25, then at a reduced rate to 60% at a concentration factor of 2.40. In contrast, Ca2+ level increased by much less during UF, by a factor of 1.04, from 2.15 mM to 2.23 mM at a concentration factor 2.65. Also, as expected, increases in TDVC were lower for UF than for RO at equivalent concentration factors. The ethanol stability of UF retentate remained at 82% up to a concentration factor 1.75, decreasing slightly to 80% at a concentration factor 2.65 and appeared to be influenced more by the composition of the soluble phase than by the protein. In a second UF trial, TDVC concentration in UF retentate increased by 2.77-fold, from 34.8 mM to 96.4 mM at a concentration factor of 4 (Table 2). Loss of TDVC from the retentate was calculated to be 30.3%. During UF, TDVC in the permeate increased only slightly, from 9.0 mM to 11.6 mM. The ratio of soluble to TDVC was 12% at this concentration factor. However, Ca2+ concentrations in both UF retentate and permeate remained constant throughout the concentration process, with only a slight increase in milk concentrated to 3-fold and 4-fold. Also, Ca2+ in permeate was lower than in retentate for samples taken at different concentration factors. Colloidal calcium and magnesium can be estimated by subtracting TDVC in permeate from that in the retentate. Colloidal TDVC in the feed was thus estimated to be 25.8 mM, whereas in the 4-fold retentate it was 84.8 mM. Thus, if no colloidal TDVC were lost, the expected concentration in the retentate would be 103.2 mM, indicating a considerable loss resulting from UF. This loss must arise
155
M.-J. Lin et al. / Journal of Food Engineering 149 (2015) 153–158 Table 1 Changes in TDVC, Ionic calcium and ethanol stability in retentates during reverse osmosis and ultrafiltration at different concentration factors (CF). Reverse osmosis
Ultrafiltration
CF
Total divalent cations
Ionic calcium (mM)
Ethanol stability (%)
CF
Total divalent cations (mM)
Ionic calcium (mM)
Ethanol stability (%)
1.0 1.25 1.5 1.75 2.0 2.4
37.0
1.74 2.07 2.05 2.13 2.19 2.28
82 68 69 65 62 60
1.0 1.25
36.6 45.6
2.15 2.08
82 82
1.75 2.65
57.6 83.6
2.10 2.23
82 80
96.0
Table 2 Changes in TDVC and Ionic calcium during ultrafiltration of milk.
Table 3 Calcium measurement and ethanol stability in UF milk subjected to pH adjustment.
Concentration factor
Retentate total divalent cations (mM)
Retentate ionic calcium (mM)
Permeate total divalent cations (mM)
Permeate ionic calcium (mM)
1.00 1.50 2.04 2.60 3.00 4.00
34.8 43.4 55.4 68.9 80.6 96.4
2.07 2.07 2.09 2.06 2.19 2.23
9.0 9.4 10.2 10.6 10.9 11.6
1.66 1.73 1.64 1.70 1.61 1.57
from movement of colloidal calcium and magnesium to their soluble forms, in order to restore equilibrium which was disturbed by removal of TDVC during the UF process. It is noteworthy that calcium is also transferred from the colloidal to the soluble phase when calcium is removed from milk by ion exchange. (Lin et al., 2006b). The relatively small increase in Ca2+, combined with removal of calcium from the micelle may help explain why a high heat stability of UF retentates was reported by Muir and Sweetsur (1978). 3.2. UF and NF Skim milk was UF treated at 50 °C to a 3.6-fold concentration factor. The concentrations of TDVC concentrations in milk, UF retentate and UF permeate were 34.0 mM, 96.2 mM and 8.60 mM, respectively. Therefore, the loss of TDVC in this UF process was 21.4%. The UF permeate was NF treated at 50 °C to a concentration factor 5.6, to investigate calcium retention. The concentrations of TDVC in NF retentate and permeate were 35.7 mM and 1.80 mM, respectively; in addition, a further 26.2% TDVC was lost in the NF process, corresponding to a rejection factor of about 0.83. The Ca2+ concentration in both UF permeate and NF permeate was the same at 0.90 mM. Thus, about 50% of the calcium was ionised in NF permeate, compared to about 10% of the calcium in UF permeate, although it is not clear what is responsible for this difference.
pH
Ionic calcium milk (mM)
Total divalent cations permeate (mM)
Ionic calcium permeate (mM)
Ethanol stability (%)
6.84 6.53 6.29 6.20 5.46 5.78* 6.73*
2.52 3.26 4.71 5.36 12.1 10.0 2.93
10.6 11.9 13.2 13.8 23.7 22.2 11.3
2.69 3.23 4.28 4.92 11.2 10.0 3.02
84 55 37 34 0 17 60
* Following pH reduction to 5.46, the pH of these samples was then increased by addition of 5M NaOH.
Ca2+ concentration increased substantially as pH was reduced, both in milk and its permeate. The ethanol stability of UF milk decreased from 84% to 34% as pH decreased from 6.84 to 6.20 and Ca2+ concentration increased from 2.52 mM to 5.36 mM, and the percentage of Ca2+ to TDVC in milk increased from 7.2% to 15.0%. When the pH of the milk reached 5.46, its Ca2+ level was 12.1 mM and its ethanol stability was lower than 20%. Restoring milk pH back to near its starting value decreased both TDVC and Ca2+ concentrations in permeate. However, the procedure for reducing and restoring pH was not reversible as at any pH, Ca2+ values for pH-restored milk were higher than the original milk and ethanol stabilities were lower. This reversibility was further investigated by reducing milk pH from 6.74 to 5.09 and then increasing it to 7.15. Changes in Ca2+ in milk are shown in Fig. 1. The trends found earlier were reproducible and showed that at any particular pH, Ca2+ was lower in the milk whilst decreasing the pH than whilst increasing it. In a third replication milk pH was reduced to 5.40. TDVC concentration in the permeate increased from 9.6 mM to 23.1 mM and then decreased to 15.3 mM when milk pH was adjusted back to 6.28
12
10.66
Decrease pH
10
Ca2+(mM)
10.12 10.66
8.73
Increase pH
3.3. Effect of pH on calcium removal UF was performed at constant composition, by returning permeate back to the feed tank, whilst changing the pH by stepwise addition of acid or alkali. In these experiments, samples were taken 5 min after pH adjustment, as it was established that no further changes in pH and Ca2+ took place after this time and changes in TDVC were also very small. TDVC in the UF milk remained constant throughout the process, confirming that its composition was constant throughout. Ca2+ concentration in the milk increased from 2.52 mM to 12.1 mM as milk pH decreased from 6.84 to 5.46 (Table 3). When pH was increased back to 6.73, Ca2+ was reduced to 2.93 mM. TDVC in permeate increased considerably as pH was reduced due to transfer of colloidal calcium to the soluble phase.
9.77
8 8.07
6.27
6 6.00
4.24
4 2 0
4.91 2.74
3.89 3.13
1.45 1.57
7.00
1.90
6.50
6.00
5.50
5.00
pH Fig. 1. Effect of stepwise acidification of milk by addition of HCl, followed by restoring pH by addition of NaOH on ionic calcium concentration in milk during ultrafiltration at 30 °C.
M.-J. Lin et al. / Journal of Food Engineering 149 (2015) 153–158
3.4. Reversibility of Ca2+ and ethanol stability in milk
[tCa]-pH [tCa]-pH
[iCa]-pH [iCa]-pH
Fig. 3 shows how ethanol stability and Ca2+ of pH restored milk samples were affected by how much the pH was reduced before it was restored. When the pH of acidified milk was returned to its original pH, neither Ca2+ concentration nor ethanol stability were restored to their original values. For example, when pH was decreased to values of 5.98 and above, the maximum ethanol stability of the pH-neutralised milk was 71%, compared to its original value of 83%. When readjusting the pH of milk acidified to even lower pH values, the ethanol stability was low when the pH was restored to its original value. The relationship between the ethanol stability and the Ca2+ concentration in milk which was acidified by adding 5 M HCl showed a non-linear relationship (see Fig. 4). The relationship between the ethanol stability and the Ca2+ concentration in acidified milk neutralised by adding 5 M NaOH to re-adjust it to near its original pH appeared closer to linear (correlation coefficient was 0.8743, for N = 8). The Ca2+ in milk increased when the milk pH decreased, which caused a corresponding increase in ethanol stability.
Calcium (ionic and total) mM
25
23.1
20
15
10
15.3 9.6
5
10.7 5.0
1.2 0
6.50
6.30
6.10
5.90
5.70
5.50
5.30
pH Fig. 2. Effect of stepwise acidification of milk by addition of HCl, followed by restoring pH by addition of NaOH on TDVC and ionic calcium concentration in permeates during ultrafiltration at 30 °C (iCa = ionic calcium: tCa = TDVC).
(Fig. 2). Ca2+ in permeate followed a similar pattern, increasing from 1.22 mM to 10.7 mM when milk pH was reduced to 5.40 and decreasing to 5.03 mM when milk pH was returned back to 6.30. For both TDVC and Ca2+, hysteresis was observed similar to that found in the earlier trial. This suggests that once calcium had been removed from the colloidal phase following pH reduction, its reintroduction into the micelle led to an alteration of these properties when pH was restored. Ezeh and Lewis (2011) reported that such changes had a more pronounced effect on rennet coagulation time than on heat stability.
(A)
3.5. Recovery of Ca2+ in cheese whey and UF permeate Cheese whey and UF permeate from cheese whey were also used to investigate partitioning of TDVC as pH was reduced and restored in order to compare their behaviour with skim milk. In cheese whey and UF permeate TDVC concentrations were 14.20 mM and 11.50 mM, respectively. Separate cheese whey and UF permeate samples were each adjusted to different pH values, followed by readjustment to their original pH (see Fig. 5). As the pH of UF permeate decreased from 6.49 to 5.04, the Ca2+
[iCa]
Eth (%)
100%
14.00
11.53
Ethanol Stability (%)
90% 80%
9.82
70%
40% 30% 10% 0%
8.00
6.34
50%
20%
10.00
7.78
60%
2.33
2.72
3.34
3.99
12.00
6.00
4.85
4.00
1.62
Ionic Calium (mM)
156
2.00
83%
65%
53%
37%
29%
25%
21%
16%
0%
0%
6.75
6.58
6.38
6.16
5.98
5.80
5.59
5. 40
5.19
4.97
-
Milk pH after HCl Added
Eth (%)
100%
Ethanol Stability (%)
90%
83%
82%
80%
78%
[iCa]
75%
14.00 12.00
71%
70%
10.00
63% 55%
60% 50%
8.00
50%
6.00
40% 30% 20%
1.62
1.85
1.94
2.02
2.14
2.31
2.42
2.23
6.75
6.58
2.55
0%
0%
5.19
4.97
4.00 2.00
10% 0%
2.51
Ionic Calium (mM)
(B)
6.38
6.16
5.98
5.80
5.59
5.40
-
Milk pH after HCl added Fig. 3. Changes of Ca2+ concentration and ethanol stability in milk (A) after adding 5 M HCl to decrease milk pH to various levels; and (B) after adding 5 M NaOH to neutralise the pH of the acidified milk to their original values, pH 6.75.
157
M.-J. Lin et al. / Journal of Food Engineering 149 (2015) 153–158
(A) Acidified Milk 100%
Ethanol stability (%)
Ethanol stability (%)
Table 4 TDVC and ionic calcium in permeates from milk subject to ultrafiltration in the total recycle mode at different temperatures.
(B) Neutralised Milk
100% 80% 60% 40% 20% 0% -
2.00
4.00
6.00
80%
60%
40% 1.00
8.00
Ionic Calcium (mM)
1.50
2.00
2.50
3.00
Ionic Calcium (mM)
Fig. 4. Relationships between ethanol stability and ionic calcium concentration in pH adjusted milk. (A) Acidified milk. (B) Neutralised milk.
3.5
3.1
(A) Acidified
3.0
Ionic Calcium (mM)
2.5 2.2
2.1
2.0
2.1
2.0
1.9
1.8
1.5 1.0
Perm-[iCa] 6.7
2.5
2.7 2.4
6.5
6.3
6.1
Whey-[iCa] 5.9
5.7
5.5
5.3
0.5 4.9 -
5.1
Temperature (°C)
Total divalent cations (mM)
Ionic calcium (mM)
26.5 31.0 36.2 41.0 45.5 50.5 55.5
9.8 9.7 9.5 9.3 8.8 8.6 8.4
1.99 1.92 1.81 1.67 1.50 1.37 1.25
from 5 °C to 25 °C during UF in total recycle mode, and Ca2+ concentration in permeate decreased from 2.77 mM to 2.34 mM. Similar trends were found for the UF permeate over the temperature range 26.5–55.5 °C (Table 4). The same trend was found when milk was cooled from 60 °C down to 10 °C. The concentrations of TDVC and Ca2+ in the original milk were 39.4 mM and 1.39 mM, respectively. TDVC concentration increased from 8.1 mM to 10.1 mM on cooling from 60 °C to 10 °C in a linear manner, whilst Ca2+ concentration in permeate increased from 0.84 mM to 1.46 mM, again in a linear manner. The results are consistent with the fact that the solubility of non-dissociated soluble calcium salts in milk is known to increase when milk temperature decreases, which causes the calcium equilibrium to move from the colloidal phase to the soluble phase, and causes the Ca2+ concentration in milk to increase. Hence more TDVC will pass through the membrane.
pH (following HCl addition) 3.5
Ionic Calcium (mM)
(B) Neutralised 2.1
3.0
2.0
1.9
1.8
1.7
1.8
1.7
1.8
1.5 1.0
Perm-[iCa] 6.7
2.1
2.0
1.9
2.5
6.5
6.3
6.1
Whey-[iCa] 5.9
5.7
5.5
5.3
0.5
5.1
4.9
pH (following HCl addition) Fig. 5. Ionic calcium concentration in (A) acidified whey and UF permeate; and (B) in neutralised whey and UF permeate.
concentration increased from 1.98 mM to 2.70 mM. For cheese whey, as pH was decreased from 6.53 to 5.02, its Ca2+ concentration increased from 2.14 mM to 3.11 mM, which was much less than that found for milk. In contrast to milk, the relationship between Ca2+ and pH in both UF permeate and cheese whey showed no hysteresis when pH was restored and is thus considered to be much more reversible compared to that observed for milk. The fraction of Ca2+ to TDVC in milk increased from 0.072 to 0.33, as milk was acidified. For whey, this increase was from 0.172 to 0.234 and for UF permeate from 0.15 to 0.219. Proportional changes were much smaller for whey and UF permeate compared to milk. 3.6. Effects of temperature The concentrations of TDVC and Ca2+ in milk were 33.9 mM and 2.47 mM. The concentration of TDVC in permeate decreased slightly from 11.1 to 10.5 mM when temperature was increased
4. Discussion Concentration by RO leads to a reduction in ethanol stability, whereas concentration by UF results in little change in ethanol stability. This reduction in ethanol stability most probably arises because of the slight reduction in pH and increase in ionic calcium for RO concentrates. Although not evaluated, the decrease in alcohol stability will probably be accompanied by a decrease in heat stability to UHT processing Shew, (1981). The process of reducing and then restoring pH in milk leads to some disruption of the casein micelle, as indicated by the increase in Ca2+ and reduction in ethanol stability following pH restoration. Other methods for assessing the extent of this disruption were not investigated. For both parameters, this was more noticeable as the reduction in pH increased. This procedure would involve transfer of some calcium out of the micelle and then back into the micelle on pH readjustment. Since samples were measured 5 min after changing the pH, these results might suggest that the kinetics of calcium removal from the micelle are faster than the process of calcium movement back into the micelle. Lucey et al. (1996) found that the increase in buffering capacity produced when milk was acidified to pH 5.5 and then neutralised to pH 6.6, was reversible. However, it became irreversible when the milk pH was adjusted to the range 4.60–5.00. The effects of acidification followed by pH restoration on heat stability and rennet coagulation time have been recently reported by Ezeh and Lewis (2011). Their results suggested that changes became irreversible in pH restored milks once the pH was reduced below about 5.8. Increasing UF temperature reduced TDVC in permeate. It was also shown that Ca2+ in permeate decreased as temperature increased. On-Nom et al. (2010) found that Ca2+ increased in both UF permeates and dialysates as temperature increased. It has been proposed that both pH and ionic calcium in milk show considerable temperature dependence. However, for UF permeate, pH and ionic calcium show little temperature dependence below the temperature of UF, but considerable temperature dependence above that temperature. This has recently been proposed by Kaombe
158
M.-J. Lin et al. / Journal of Food Engineering 149 (2015) 153–158
et al. (2012) and is influenced by whether precipitated calcium phosphate will dissolve when the temperature is reduced. Our results confirm these finding, since when UF was done below 15 °C, Ca2+ in permeates were higher than in milk, but above 25 °C were always lower than in milk. Reductions in Ca2+ in permeate were higher than changes in TDVC in percentage terms and the proportion of ionised calcium in permeate decreased as temperature increased. Whey and UF permeate both showed a difference in behaviour compared to milk, and the concentration of soluble calcium increased much less as pH was reduced (see Fig. 5). Also, increases in Ca2+ in both whey and permeate were small, compared to that in milk. It is suggested that this shows the equilibrium conditions prevailing between Ca2+ and soluble calcium in the absence of casein in the colloidal phase. TDVC retention in UF retentates can be manipulated by controlling pH during UF. From a nutritional standpoint it would be advantageous to minimise losses of TDVC when UF retentates are used for manufacture of quarg or soft cheese. However, in order to minimise off-flavour production in UF quarg, milk should be acidified prior to UF (Winwood, 1983), which would minimise retention of TDVC in the product. Thus one possible explanation is that higher amounts of TDVC in these products will have a detrimental effect on flavour. The flavour defect is often described as metallic, but a mechanism relating this to higher TDVC has not yet been elucidated. Processed cheeses with bitter notes were also observed to have higher levels of minerals and phosphorus (Mayer, 2001), although the source of this may have been an overdose of an emulsifying agent of high phosphorus content. 5. Conclusions During RO, there was an increase in TDVC and Ca2+, and a reduction in the ethanol stability of RO retentates. During UF of milk at its normal pH, there was an increase in TDVC, but only a slight increase in Ca2+ for the retentate but ethanol stability was only slightly reduced. During NF, a small amount of TDVC was found in the permeate and TDVC rejection was estimated to be about 0.83. When UF was performed whilst reducing milk pH and then restoring it, the movement of TDVC from and back into the micelle was not reversible. Over the pH range 6.8–5.1, TDVC and Ca2+ in permeate were always higher as the pH was restored. In contrast, similar experiments performed on whey and UF permeate showed that movements of TDVC occurred to a lesser extent and were also reversible. TDVC and Ca2 in UF permeate decreased as UF temperature increased, due to the lower solubility of calcium phosphate at higher temperature. These opportunities to manipulate TDVC in UF milk concentrates may be beneficial in terms of manipulating their composition to improve heat stability or other functional properties.
References Bastian, E.D., Colling, S.K., Ernstom, C.A., 1991. Ultrafiltration partitioning of milk constituents into permeate and retentate. J. Dairy Sci. 74, 2423–2434. Davies, D.T., White, J.C.D., 1960. The use of ultrafiltration and dialysis in isolating the aqueous phase of milk and determining the partitioning of milk constituents between the aqueous and disperse phases. J. Dairy Res. 27, 171– 190. Ezeh, V.N., Lewis, M.J., 2011. Milk reversibility following reduction and restoration of pH. Int. J. Dairy Technol. 64, 179–187. Field, R.W., Wu, D., Howell, J.A., 1995. Critical flux for microfiltration fouling. J. Membr. Sci. 100, 259–272. Glover, F., 1985. Ultrafiltration and Reverse Osmosis for the Dairy Industry, Technical Bulletin No. 5. NIRD, Reading, England. Hiddink, J., de Boer, R., Romijn, D.J., 1978. Removal of salts during ultrafiltration of whey and buttermilk. Netherlands Milk Dairy J. 32, 80–93. Holt, C., 2004. An equilibrium thermodynamic model of the sequestration of calcium phosphate by casein micelles and its application to the calculation of the partition of salts in milk. Eur. Biophys. J. 33, 421–434. Holt, C., Dalgleish, D.G., Jenness, R., 1981. Calculation of the ion equilibria in milk diffusate and comparison with experiment. Anal. Biochem. 113, 154–163. Kaombe, D.D., Du, Y., Lewis, M.J., 2012. Mineral partitioning of milk at high temperature. J. Dairy Res. 79, 1–6. Lin, M.-J., Lewis, M.J., Grandison, A.S., 2006a. Measurement of ionic calcium in milk. Int. J. Dairy Technol. 59, 192–199. Lin, M.-J., Grandison, A., Chryssanthou, X., Goodwin, C., Tsioulpas, A., Koliandris, A., Lewis, M., 2006b. Calcium removal from milk by ion exchange. Milchwissenschaft 61, 370–374. Lucey, J.A., Gorry, C., O’Kennedy, B., Kalab, M., Tan-Kinita, R., Fox, P.F., 1996. Effect of acidification and neutralisation of milk on some physico-chemical properties of casein micelles. Int. Dairy J. 6, 257–272. May, R.J., Smith, D.E., 1998. Effect of storage and various processing conditions on the amount of ionic calcium in milk. Milchwissenschaft 53, 605–608. Mayer, H.K., 2001. Bitterness in processed cheese caused by an overdose of a specific emulsifying agent. Int. Dairy J. 11, 533–542. Muir, D.D., Sweetsur, A.W.M., 1978. The effect of concentration on the heat stability of skim milk. J. Dairy Res. 45, 37–45. On-Nom, M., Grandison, A.S., Lewis, M.J., 2010. Measurement of ionic calcium, pH and soluble divalent cations in milk at high temperature. J. Dairy Sci. 93 (2), 515–523. Pouliot, Y., Boulet, M., Paquin, P., 1989a. Observations on the heat induced salt balance changes in milk 1 effect of heating time between 4 and 90 °C. J. Dairy Res. 56, 185–192. Pouliot, Y., Boulet, M., Paquin, P., 1989b. An experimental technique for the study of milk salt balance. J. Dairy Sci. 72, 36–40. Pouliot, Y., Boulet, M., Paquin, P., 1989c. Observations on the heat induced salt balance changes in milk. J. Dairy Res. 56, 193–199. Premaratne, R.J., Cousin, M.A., 1991. Changes in chemical composition during ultrafiltration of skim milk. J. Dairy Sci. 74, 788–795. Renner, E., AbD El-Salam, M.H., 1991. Application of Ultrafiltration in the Dairy Industry. Elsevier Applied Science, London. Rose, D., Tessier, H., 1959. Composition of ultrafiltrates from milk heated at 80 to 230 F in relation to heat stability. J. Dairy Sci. 42, 969–980. Shew, D.I., 1981. New Monograph in UHT Milk, Document No. 133. International Dairy Federation, Brussels, pp. 115–121. Sood, V.K., Kosikowski, F.V., 1979. Ultrafiltration of milk at high temperature. J. Food Protect. 42, 958–960. Sweetsur, A.W.M., Muir, D.D., 1985. Effect of fat incorporation on properties of sterile concentrates prepared by ultrafiltration of skimmed milk. J. Soc. Dairy Technol. 38, 88–93. Winwood, J., 1983. Quarg production methods – past, present and future. J. Soc. Dairy Technol. 36, 107–109. Youravong, W., Lewis, M.J., Grandison, A.S., 2003. Critical flux in ultrafiltration of skimmed milk. Trans. Inst. Chem. Eng. 81, 303–308.