Copyright is owned by the Author of the thesis.  Permission is given for 
a copy to be downloaded by an individual for the purpose of research and 
private study only.  The thesis may not be reproduced elsewhere without 
the permission of the Author. 
 
  
 
 
 
 
 
 
Ultra Filtration (UF) Process Development for  
the Production of Camembert Cheese 
 
 
A thesis presented in partial fulfilment of the requirements for the degree of 
 
 
Master of Technology 
in 
Food Technology 
 
 
at Massey University, Albany, 
New Zealand. 
 
 
Ming Ho Edwin Law 
 
 
2010 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  i 
ABSTRACT 
 
The application of UF technology in cheese production has several potential 
advantages; product consistency, yield, lower costs and more automation. This study 
investigated the effects of four processing variables in the manufacture of Camembert 
cheese using UF and their impact on cheese quality. Using an incomplete block 
design, sixteen unique treatments were produced with combined processing variables 
(high-fat or low-fat; brine-salted or retentate-salted; acidified to pH 5.2 or pH 4.9; set 
in tubular moulds and small moulds). The cheeses were matured for seven weeks at 
4?1 ?C and were analysed for total solids, fat, salt, non-protein nitrogen (NPN) and 
soluble nitrogen (SN) contents during the maturation period (seven weeks). Major 
defects were evaluated by experienced cheese graders in the fourth week. pH was 
measured and instrumental analysis was also conducted. Sensory evaluation on 
consumer acceptance was also conducted in the fourth week. 
 
All the cheese samples exhibited similar increases in rind and core pH, NPN/TN and 
SN/TN ratios, and were generally characterised by thick rind and softness. The low-
fat cheese samples had significantly lower NPN/TN ratio and higher overall 
acceptance in sensory evaluation. The salt content was also significantly higher. The 
retentate-salted cheese samples had significantly lower NPN/TN ratios and more 
defects in rind discolouration and deformation, and saltiness. The cheese samples 
acidified to pH 5.2 had significantly lower NPN/TN ratios and fewer defects in rind 
discolouration, softness, sourness, and bitterness. The cheese samples made using 
tube moulds were significantly firmer with fewer defects in rind deformation, core 
unevenness, and softness. 
 
The level of fat and extent of acidification was found to have a profound effect on 
cheese quality, and cheeses produced with low-fat retentate and/or acidified to pH 5.2 
generally had superior shelf-life with lower levels of proteolysis. The preference of 
the two salting methods may be debatable, but considering labour and time, retentate-
salting is preferable. Tube mould generally produced better cheese with fewer defects. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  ii 
ACKNOWLEDGEMENTS 
 
I would like to acknowledge Goodman Fielder (NZ) Ltd. for offering the opportunity 
and financial support to carry out this study. 
 
I would also like to thank: 
? Mr. Hamish Conway, Technical and Innovation Manager of Goodman Fielder 
(NZ) Ltd. for his kind advice and support in this study. 
? My supervisors Dr. Anthony Mutukumira and Dr. Marie Wong for their 
constant encouragement and guidance in this study. 
? Helen Matthews for her kind advice and assistance in this study. 
? Tim Meredith, David Pride, Caroline Edge, Divya Patel, Aaron De Jonge, 
Martyn Finlay, and Derick Bussell of Goodman Fielder (NZ) Ltd. Puhoi 
Valley Cheese, as well as all staff that assisted in this study for their support 
and generosity. 
? Dr. John Grigor and Dr. Daniel Walsh for their advice in sensory science and 
statistics. 
? Massey University and Foundation for Research, Science & Technology for 
their resources and financial support. 
? The people who spent time and effort for the participation of this study. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  iii 
TABLE OF CONTENTS 
 
ABSTRACT ................................................................................................................... i 
ACKNOWLEDGEMENTS ........................................................................................ ii 
TABLE OF CONTENTS ...........................................................................................iii 
LIST OF TABLES ...................................................................................................... vi 
LIST OF FIGURES ................................................................................................... vii 
LIST OF ABBREVIATIONS ..................................................................................xiii 
 
1. INTRODUCTION ................................................................................................ 1 
 
2. LITERATURE REVIEW ................................................................................... 3 
2.1. Introduction ......................................................................................................... 3 
2.2. Cheese ................................................................................................................. 4 
2.2.1. Types of cheeses .......................................................................................... 5 
2.3. Advances in the dairy industry ........................................................................... 7 
2.3.1. Membrane technology ................................................................................. 7 
2.3.2. Principles of membrane filtration ................................................................ 8 
2.3.3. Pressure-driven membrane processes .......................................................... 9 
2.3.4. Microfiltration .............................................................................................. 9 
2.3.5. Ultrafiltration ............................................................................................. 10 
2.3.6. Nanofiltration and reverse osmosis ............................................................ 10 
2.3.7. Application of membrane technology in cheese ........................................ 11 
2.3.8. Characteristics of cheeses manufactured using membrane processes ....... 11 
2.3.9. Major approaches in cheese-making using membrane processes .............. 12 
2.4. Camembert cheese ............................................................................................ 14 
2.4.1. Composition, properties and standards ...................................................... 15 
2.4.2. Manufacturing of Camembert .................................................................... 15 
2.4.3. UF in Camembert cheese-making .............................................................. 17 
2.5. Cheese ripening and quality .............................................................................. 18 
2.5.1. Effects of lactic fermentation ..................................................................... 18 
2.5.2. Proteolysis .................................................................................................. 19 
2.5.3. Lipolysis ..................................................................................................... 22 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  iv 
2.5.4. Flavour development ................................................................................. 23 
2.5.5. Texture development ................................................................................. 26 
2.6. Analysis of cheese ............................................................................................. 26 
2.6.1. Determination of pH .................................................................................. 26 
2.6.2. Total solids ................................................................................................. 27 
2.6.3. Salt content ................................................................................................. 27 
2.6.4. Fat content .................................................................................................. 27 
2.6.5. Nitrogen content ......................................................................................... 28 
2.6.6. Texture analysis ......................................................................................... 28 
2.6.7. Sensory analysis ......................................................................................... 29 
2.6.8. Microbiological analysis ............................................................................ 30 
2.7. Conclusions ....................................................................................................... 31 
 
3. MATERIALS AND METHODS ...................................................................... 32 
3.1. Experimental design .......................................................................................... 32 
3.1.1. Description of processing variables ........................................................... 33 
3.2. Production of Camembert ................................................................................. 34 
3.3. Analysis of nitrogen fractions ........................................................................... 42 
3.4. Measurement of pH ........................................................................................... 43 
3.5. Determination of total solids ............................................................................. 44 
3.6. Determination of fat content ............................................................................. 44 
3.7. Determination of salt content ............................................................................ 45 
3.8. Instrumental analysis ........................................................................................ 45 
3.9. Cheese grading .................................................................................................. 46 
3.10. Consumer acceptance ...................................................................................... 47 
3.11. Microbiological analysis ................................................................................. 48 
3.12. Statistical analysis of data ............................................................................... 48 
 
4. RESULTS ........................................................................................................... 49 
4.1. Change in total solids ........................................................................................ 49 
4.2. Salt content ........................................................................................................ 55 
4.3. Fat content ......................................................................................................... 56 
4.4. Changes in pH levels ........................................................................................ 58 
4.5. Level of proteolysis ........................................................................................... 64 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  v 
4.5.1. Non-protein nitrogen (NPN) ...................................................................... 64 
4.5.2. pH 4.4 soluble nitrogen (SN) ..................................................................... 70 
4.6. Texture analysis ................................................................................................ 76 
4.7. Cheese grading .................................................................................................. 79 
4.8. Consumer acceptance ........................................................................................ 84 
 
5. DISCUSSION ..................................................................................................... 89 
5.1. Moisture retention ............................................................................................. 89 
5.2. Salt uptake ......................................................................................................... 92 
5.3. Proteolysis ......................................................................................................... 94 
5.4. Textural and organoleptic properties ................................................................ 97 
5.4.1. Instrumental analysis ................................................................................. 97 
5.4.2. Cheese grading ........................................................................................... 99 
5.5. Consumer acceptance ...................................................................................... 109 
 
6. CONCLUSIONS .............................................................................................. 112 
 
7. RECOMMENDATIONS ................................................................................. 114 
 
8. REFERENCES ................................................................................................. 115 
 
APPENDICES 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  vi 
LIST OF TABLES 
 
Table 3.1: Incomplete block design of sixteen unique treatments for the manufacture 
of Camembert cheese ................................................................................................... 32 
 
Table 3.2: A list of attributes used in cheese grading and the total scores given in each 
subgroup. ...................................................................................................................... 47 
 
Table 4.1: p-values (p ? 0.05) for total solids within each type of treatment. ............. 49 
 
Table 4.2: p-values (p ? 0.05) for salt and fat levels within each type of treatment. ... 56 
 
Table 4.3: p-values (p ? 0.05) for core pH within each type of treatment. .................. 59 
 
Table 4.4: p-values (p ? 0.05) for NPN/TN ratios within each type of treatment. ...... 64 
 
Table 4.5: p-values (p ? 0.05) for SN/TN ratios within each type of treatment. ......... 70 
 
Table 4.6: Mean scores and standard deviations of the combined of the five sensory 
attributes. ...................................................................................................................... 85 
 
Table 4.7: p-values (p ? 0.05) of the four processing variables of the five sensory 
attributes. ...................................................................................................................... 85 
 
Table 5.1: Examples of important flavour components in Camembert cheese (Fox et 
al., 2004). ................................................................................................................... 108 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  vii 
LIST OF FIGURES 
 
Figure 2.1: Manufacturing protocol for Camembert cheese (Walstra et al., 2006). .... 16 
 
Figure 2.2: Overview of general protein conversion pathways for flavour formation in 
cheese (Walstra et al., 2006). ....................................................................................... 24 
 
Figure 2.3: Formation of flavour compounds from fat in cheese (Walstra et al., 2006).
 ...................................................................................................................................... 25 
 
Figure 3.1: Set-up of the four processing variables applied to four treatments per 
block of Camembert cheese production. ...................................................................... 36 
 
Figure 3.2: A small batch of retentate (80 L) was separated from the main batch to 
carry out trial experiment with four treatments (20L each). ........................................ 36 
 
Figure 3.3: A pour tank with pump which circulates and allows the mixing of retentate 
with other ingredients. The pre-cheese retentate was then pumped into different 
moulds. ......................................................................................................................... 37 
 
Figure 3.4: Small silicone tray moulds containing pre-cheese retentate. .................... 37 
 
Figure 3.5: Tube moulds containing pre-cheese retentate in plastic films. ................. 38 
 
Figure 3.6: Wire cutter used for slicing young cheese logs removed from tube moulds 
into approximately 125 g units. ................................................................................... 38 
 
Figure 3.7: Individual young Camembert cheeses loaded onto metal racks. ............... 39 
 
Figure 3.8: Brine tank used to immerse young cheeses. Stirring of the excess salt at 
the bottom of the tank was done prior to the immersion of the young cheeses, which 
ensures the salinity is above 95%. ............................................................................... 39 
 
Figure 3.9: Process flow chart of Camembert cheese production using ultrafiltration, 
from milk reception to culture addition. ...................................................................... 40 
 
Figure 3.10: Process flow chart of Camembert cheese production using ultrafiltration, 
continuing from culture addition to the finished product. (*) represents process 
variables which were manipulated. .............................................................................. 41 
 
Figure 4.1: Changes in total solids (TS) (mean ? SEM) of low-fat (n = 32) and high-fat 
(n = 16) Camembert cheese stored at 4 ? 1 ?C for six weeks post-packaging. ............ 51 
 
Figure 4.2: Changes in total solids (TS) (mean ? SEM) of brine-salted (n = 24) and 
retentate-salted (n = 16) Camembert cheese stored at 4 ? 1 ?C for six weeks post-
packaging. .................................................................................................................... 51 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  viii 
Figure 4.3: Changes in total solids (TS) (mean ? SEM) of Camembert cheese acidified 
to pH 5.2 (n = 24) and pH 4.9 (n = 24). The samples were stored at 4 ? 1 ?C for six 
weeks post-packaging. ................................................................................................. 52 
 
Figure 4.4: Changes in total solids (TS) (mean ? SEM) of Camembert cheese made 
using tube moulds (n = 24) and small moulds (n = 24). The samples were stored at 4 ? 
1 ?C for six weeks post-packaging. .............................................................................. 52 
 
Figure 4.5: Changes in total solids (TS) (mean ? SEM) of Camembert cheese for low-
fat brine-salted samples (n = 4 in each treatment). The samples were stored at 4 ? 1 ?C 
for six weeks post-packaging. ...................................................................................... 53 
 
Figure 4.6: Changes in total solids (TS) (mean ? SEM) of Camembert cheese for low-
fat retentate-salted samples (n = 4 in each treatment). The samples were stored at 4 ? 
1 ?C for six weeks post-packaging. .............................................................................. 53 
 
Figure 4.7: Changes in total solids (TS) (mean ? SEM) of Camembert cheese for high-
fat brine-salted samples (n = 4 in each treatment). The samples were stored at 4 ? 1 ?C 
for six weeks post-packaging. ...................................................................................... 54 
 
Figure 4.8: Changes in total solids (TS) (mean ? SEM) of Camembert cheese for high-
fat retentate-salted samples (n = 4 in each treatment). The samples were stored at 4 ? 
1 ?C for six weeks post-packaging. .............................................................................. 54 
 
Figure 4.9: Combined mean of salt content (mean ? SEM) in Camembert cheese 
samples for each treatment: high-fat (n = 32) and low-fat (n = 32); brine-salted (n = 
32) and retentate-salted (n = 32); final acidification pH of 5.2 (n = 32) and pH 4.9 (n = 
32); and tube mould (n = 32) and small mould (n = 32). Samples were analysed in the 
first week post-packaging. ........................................................................................... 55 
 
Figure 4.10: Salt content (mean ? SEM) in Camembert cheese samples of sixteen 
treatments (n = 4 in each treatment) analysed in the first week post-packaging. ........ 56 
 
Figure 4.11: Combined mean of fat content (mean ? SEM) in Camembert cheese 
samples for each treatment: low-fat (n = 16) and high-fat (n = 16); brine-salted (n = 
16) and retentate-salted (n = 16); final acidification pH of 5.2 (n = 16) and pH 4.9 (n = 
16); and tube mould (n = 16) and small mould (n = 16). Samples were analysed in the 
first week post-packaging. ........................................................................................... 57 
 
Figure 4.12: Fat content (mean ? SEM) in Camembert cheese samples of sixteen 
treatments (n = 4 in each treatment) analysed in the first week post-packaging. ........ 57 
 
Figure 4.13: Changes in the core and rind pH (mean ? SEM) of low-fat (n = 64) and 
high-fat (n = 32) Camembert cheese stored at 4 ? 1 ?C for six weeks post-packaging.
 ...................................................................................................................................... 60 
 
Figure 4.14: Changes in the core and rind pH (mean ? SEM) of brine-salted (n = 48) 
and retentate-salted (n = 48) Camembert cheese stored at 4 ? 1 ?C for six weeks post-
packaging. .................................................................................................................... 60 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  ix 
Figure 4.15: Changes in the core and rind pH (mean ? SEM) of Camembert cheese 
acidified to pH 5.2 (n = 48) and pH 4.9 (n = 48). The samples were stored at 4 ? 1 ?C 
for six weeks post-packaging. ...................................................................................... 61 
 
Figure 4.16: Changes in the core and rind pH (mean ? SEM) of Camembert cheese 
made using tube moulds (n = 48) and small moulds (n = 48). The samples were stored 
at 4 ? 1 ?C for six weeks post-packaging. .................................................................... 61 
 
Figure 4.17: Changes in the core and rind pH (mean ? SEM) of Camembert cheese for 
low-fat brine-salted samples (n = 8 in each treatment). The samples were stored at 4 ? 
1 ?C for six weeks post-packaging. .............................................................................. 62 
 
Figure 4.18: Changes in the core and rind pH (mean ? SEM) of Camembert cheese for 
low-fat retentate-salted samples (n = 8 in each treatment). The samples were stored at 
4 ? 1 ?C for six weeks post-packaging. ........................................................................ 62 
 
Figure 4.19: Changes in the core and rind pH (mean ? SEM) of Camembert cheese for 
high-fat brine-salted samples (n = 8 in each treatment). The samples were stored at 4 
? 1 ?C for six weeks post-packaging. ........................................................................... 63 
 
Figure 4.20: Changes in the core and rind pH (mean ? SEM) of Camembert cheese for 
high-fat retentate-salted samples (n = 8 in each treatment). The samples were stored at 
4 ? 1 ?C for six weeks post-packaging. ........................................................................ 63 
 
Figure 4.21: Changes in the NPN/TN ratios (mean ? SEM) of low-fat (n = 32) and 
high-fat (n = 16) Camembert cheese stored at 4 ? 1 ?C for seven weeks post-
packaging. .................................................................................................................... 66 
 
Figure 4.22: Changes in the NPN/TN ratio (mean ? SEM) of brine-salted (n = 24) and 
retentate-salted (n = 24) Camembert cheese stored at 4 ? 1 ?C for seven weeks post-
packaging. .................................................................................................................... 66 
 
Figure 4.23: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese 
acidified to pH 5.2 (n = 24) and pH 4.9 (n = 24). The samples were stored at 4 ? 1 ?C 
for seven weeks post-packaging. ................................................................................. 67 
 
Figure 4.24: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese made 
using tube moulds (n = 24) and small moulds (n = 24). The samples were stored at 4 ? 
1 ?C for seven weeks post-packaging. ......................................................................... 67 
 
Figure 4.25: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese for 
low-fat brine-salted samples (n = 4 in each treatment). The samples were stored at 4 ? 
1 ?C for seven weeks post-packaging. ......................................................................... 68 
 
Figure 4.26: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese for 
low-fat retentate-salted samples (n = 4 in each treatment). The samples were stored at 
4 ? 1 ?C for seven weeks post-packaging. ................................................................... 68 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  x 
Figure 4.27: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese for 
high-fat brine-salted samples (n = 4 in each treatment). The samples were stored at 4 
? 1 ?C for seven weeks post-packaging. ...................................................................... 69 
 
Figure 4.28: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese for 
high-fat retentate-salted samples (n = 4 in each treatment). The samples were stored at 
4 ? 1 ?C for seven weeks post-packaging. ................................................................... 69 
 
Figure 4.29: Changes in the SN/TN ratio (mean ? SEM) of low-fat (n = 32) and high-
fat (n = 16) Camembert cheese stored at 4 ? 1 ?C for seven weeks post-packaging. .. 72 
 
Figure 4.30: Changes in the SN/TN ratio (mean ? SEM) of brine-salted (n = 24) and 
retentate-salted (n = 24) Camembert cheese stored at 4 ? 1 ?C for seven weeks post-
packaging. .................................................................................................................... 72 
 
Figure 4.31: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese 
acidified pH 5.2 (n = 24) and pH 4.9 (n = 24). The samples were stored at 4 ? 1 ?C for 
seven weeks post-packaging. ....................................................................................... 73 
 
Figure 4.32: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese made 
using tube moulds (n = 24) and small moulds (n = 24). The samples were stored at 4 ? 
1 ?C for seven weeks post-packaging. ......................................................................... 73 
 
Figure 4.33: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese for low-
fat brine-salted samples (n = 4 in each treatment). The samples were stored at 4 ? 1 ?C 
for seven weeks post-packaging. ................................................................................. 74 
 
Figure 4.34: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese for low-
fat retentate-salted samples (n = 4 in each treatment). The samples were stored at 4 ? 
1 ?C for seven weeks post-packaging. ......................................................................... 74 
 
Figure 4.35: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese for 
high-fat brine-salted samples (n = 4 in each treatment). The samples were stored at 4 
? 1 ?C for seven weeks post-packaging. ...................................................................... 75 
 
Figure 4.36: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese for 
high-fat retentate-salted samples (n = 4 in each treatment). The samples were stored at 
4 ? 1 ?C for seven weeks post-packaging. ................................................................... 75 
 
Figure 4.37: Change in force (N) (mean ? SEM) at 50% deformation for low-fat brine-
salted cheese samples (n = 8 in each treatment) stored at 4 ? 1 ?C for six weeks post-
packaging. .................................................................................................................... 77 
 
Figure 4.38: Change in force (N) (mean ? SEM) at 50% deformation for low-fat 
retentate-salted cheese samples (n = 8 in each treatment) stored at 4 ? 1 ?C for six 
weeks post-packaging. ................................................................................................. 77 
 
Figure 4.39: Change in force (N) (mean ? SEM) at 50% deformation for high-fat 
brine-salted cheese samples (n = 8 in each treatment) stored at 4 ? 1 ?C for six weeks 
post-packaging. ............................................................................................................ 78 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  xi 
Figure 4.40: Change in force (N) (mean ? SEM) at 50% deformation for high-fat 
retentate-salted cheese samples (n = 8 in each treatment) stored at 4 ? 1 ?C for six 
weeks post-packaging. ................................................................................................. 78 
 
Figure 4.41: Percentage of occurrence in nine major sensory defects for high-fat (n = 
8) and low-fat (n = 8) Camembert cheese at the age of four weeks. The samples were 
stored at 4 ? 1 ?C post-packaging. ............................................................................... 82 
 
Figure 4.42: Percentage of occurrence in nine major sensory defects for brine-salted 
(n = 8) and retentate-salted (n = 8) Camembert cheese at the age of four weeks. The 
samples were stored at 4 ? 1 ?C post-packaging. ......................................................... 82 
 
Figure 4.43: Percentage of occurrence in nine major sensory defects for Camembert 
cheese acidified to pH 5.2 (n = 8) and pH 4.9 (n = 8) at the age of four weeks. The 
samples were stored at 4 ? 1 ?C post-packaging. ......................................................... 83 
 
Figure 4.44: Percentage of occurrence in nine major sensory defects for Camembert 
cheese made using tube moulds (n = 8) and small moulds (n = 8) at the age of four 
weeks. The samples were stored at 4 ? 1 ?C post-packaging. ..................................... 83 
 
Figure 4.45: Consumer overall acceptance of high-fat (n = 225) and low-fat (n = 620) 
Camembert cheese at the age of four weeks. The cheese samples were stored at 4 ? 1 
?C post-packaging. ? represents the mean; horizontal lines represent upper quartile, 
median, and lower quartile respectively from top to bottom; * represents outliers. .... 86 
 
Figure 4.46: Consumer overall acceptance of brine-salted (n = 420) and retentate-
salted (n = 425) Camembert cheese at the age of four weeks. The cheese samples were 
stored at 4 ? 1 ?C post-packaging. ? represents the mean; horizontal lines represent 
upper quartile, median, and lower quartile respectively from top to bottom; * 
represents outliers. ....................................................................................................... 86 
 
Figure 4.47: Consumer overall acceptance of Camembert cheese acidified to pH 5.2 
(n = 413) and pH 4.9 (n = 432) at the age of four weeks. The cheese samples were 
stored at 4 ? 1 ?C post-packaging. ? represents the mean; horizontal lines represent 
upper quartile, median, and lower quartile respectively from top to bottom; * 
represents outliers. ....................................................................................................... 87 
 
Figure 4.48: Consumer overall acceptance of Camembert cheese made using small 
mould (n = 414) and tube mould (n = 431) at the age of four weeks. The cheese 
samples were stored at 4 ? 1 ?C post-packaging. ? represents the mean; horizontal 
lines represent upper quartile, median, and lower quartile respectively from top to 
bottom; * represents outliers. ....................................................................................... 87 
 
Figure 4.49: Consumer overall acceptance of sixteen treatments of Camembert cheese 
at the age of four weeks. The cheese samples were stored at 4 ? 1 ?C post-packaging. 
The sample size of consumer panellists varied between 21 and 43 panellists per 
treatment. ? represents the mean; horizontal lines represent upper quartile, median, 
and lower quartile respectively from top to bottom; * represents outliers. ................. 88 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  xii 
Figure 5.1: The rind discolouration defect, shown by browning of the mould, with 
some discolouration on the edges. ............................................................................. 101 
 
Figure 5.2: The rind deformation defect, shown by unevenness and concaving of the 
rind. ............................................................................................................................ 102 
 
Figure 5.3: The thick rind defect, shown in the UF Camembert cheese (left) 
comparing to the Camembert made conventionally (right). ...................................... 102 
 
Figure 5.4: The core unevenness defect, shown by the overripened soft and flowing 
texture in the outer portion and firm (still chalky) texture in the inner portion of the 
cheese paste. ............................................................................................................... 103 
 
Figure 5.5: Results of softening at a mild degree (left), and over softening (right) with 
the development of a flowing texture of the cheese paste. ........................................ 104 
 
 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  xiii 
LIST OF ABBREVIATIONS 
 
HB4.9S High-fat, brine-salted, pH 4.9, small mould treatment 
HB4.9T High-fat, brine-salted, pH 4.9, tube mould treatment 
HB5.2S High-fat, brine-salted, pH 5.2, small mould treatment 
HB5.2T High-fat, brine-salted, pH 5.2, tube mould treatment 
HR4.9S High-fat, retentate-salted, pH 4.9, small mould treatment 
HR4.9T High-fat, retentate -salted, pH 4.9, tube mould treatment 
HR5.2S High-fat, retentate-salted, pH 5.2, small mould treatment 
HR5.2T High-fat, retentate-salted, pH 5.2, tube mould treatment 
LAB  Lactic acid bacteria 
LB4.9S Low-fat, brine-salted, pH 4.9, small mould treatment 
LB4.9T Low -fat, brine-salted, pH 4.9, tube mould treatment 
LB5.2S Low -fat, brine-salted, pH 5.2, small mould treatment 
LB5.2T Low -fat, brine-salted, pH 5.2, tube mould treatment 
LR4.9S Low -fat, retentate-salted, pH 4.9, small mould treatment 
LR4.9T Low -fat, retentate -salted, pH 4.9, tube mould treatment 
LR5.2S Low -fat, retentate-salted, pH 5.2, small mould treatment 
LR5.2T Low -fat, retentate-salted, pH 5.2, tube mould treatment 
MF  Microfiltration 
N  Nitrogen 
NF  Nanofiltration 
NPN  Non-protein nitrogen 
RO  Reverse osmosis 
SN  Soluble-nitrogen 
TCA  Trichloroacetic acid 
TN  Total nitrogen 
UF  Ultrafiltration 
UHT  Ultra-high temperature 
 
 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  1 
1. INTRODUCTION 
 
Camembert is a surface mould-ripened cheese with a soft consistency in a flat 
cylindrical form (Scott, 1998). It originated in the Normandy region of France and has 
been produced industrially since the beginning of the twentieth century (Fox et al., 
2004). The manufacture of Camembert and other surface mould-ripened cheeses has 
since become widespread in France and other regions in Europe. 
 
Traditionally, Camembert is made from raw milk with the addition of a mesophilic 
starter. Rennet is added at about pH of 6.4 and coagulation occurs for 30-45 min. The 
coagulum is transferred to the moulds and allowed for whey drainage starting at 26 ? 
28 ?C and gradually cooled to 20 ?C. At this stage, a curd with pH of 4.6-4.7 and low 
mineral content is obtained. The cheese is dry-salted and ripened for 21 days or more 
at 11-13 ?C and 90% relative humidity (Fox et al., 2000). With the development of 
industrial methods to accommodate large milk throughput, Alpma-type processes and 
highly mechanised equipments are now used (Fox et al., 2004).  
 
The emergence of membrane processes has unlocked many opportunities in the dairy 
industry, leading to significant new process and product development (Fox et al., 
2004). The introduction of membrane processes in cheese manufacture by the 
Maubois, Mocquot and Vassal (MMV) process has enabled significant advances in 
cheese-making (Maubois and Mocquot, 1975). The benefits include improved plant 
efficiency, increased cheese yield, and the development of automated processes (Fox 
et al., 2004). Camembert was the first cheese to be made according to the MMV 
principle using ultrafiltration (UF), and its application has been well accepted in the 
industry using a continuous moulding and demoulding system ?Camatic? (Maubois, 
2002; Fox et al., 2004). 
 
Although UF cheeses provide advantages such as yield increase, reduced cost of 
labour, and increased plant capacity, they possess unique ripening characteristics and 
texture qualities (Fox et al., 2004). These properties may cause quality defects which 
impact on the preference of the consumer in purchasing the product (Walstra et al., 
2006). It is therefore of paramount importance to study the factors contributing to 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  2 
undesirable characteristics, and modify the existing manufacturing methods aimed at 
improving the overall cheese quality and shelf-life. 
 
The main objective of this project was to determine the optimum processing 
parameters for the production of Camembert cheese from bovine milk utilising a 
hollow-fibre UF system, with respect to improving the overall quality and shelf-life of 
the cheese. Four processing variables were studied. The cheeses made were monitored 
for chemical, microbiological, physical, and sensory aspects during maturation to 
evaluate the quality of the cheeses. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  3 
2. LITERATURE REVIEW 
2.1. Introduction 
For millennia, bovine milk has been used to produce products such as cream, butter, 
kefir, yogurt, and cheese. With the advances of modern technology, it has enabled us 
to produce ice cream, powdered milk and infant formulae, concentrated milks, lactose, 
casein and functional protein products for food-additive and industrial uses (Walstra 
et al., 2006).  
 
In many cultures of the world, the consumption of milk by humans continues beyond 
infancy, especially in the Western world. Animal milk, particularly cows? milk, is 
used as a food product due to its nutritional value. 
 
The principle components in bovine milk include lactose, fat, protein, minerals, 
organic acids, and water. Lactose, or milk sugar, is the major carbohydrate of milk 
and is a disaccharide of glucose and galactose. It is mainly converted into lactic acid 
in many fermented milk products such as yogurt and cheese (Walstra et al., 2006). 
The fats in milk are primarily triglycerides, varying in lengths (2 to 20 carbon atoms) 
and saturation (0 to 4 double bonds). Other milk lipids include diglycerides, 
monoglycerides, phospholipids, cholesterol, and free fatty acids. Nearly all fat in milk 
is in fat globules, which varies in diameter (0.1 to 15 ?m) and consists of a membrane 
to prevent them from coalescence (Walstra et al., 2006). Casein, present in micelles, 
accounts for about four fifths of milk protein. It includes ?s1-, ?s2-, ?-, and ?-casein. 
The behaviour of casein micelles has great importance in determining the properties 
of milk and subsequent dairy products, especially in cheese making. Other proteins in 
milk consists of serum proteins, ?-lactoglobulin, and a wide range of enzymes 
(Walstra et al., 2006). The mineral substances in milk mainly consist of K, Na, Ca, 
Mg, Cl, and phosphate. Milk also contains partly ionized salts and organic acids, such 
as citrate. 
 
The chemical composition of milk varies between species, and is affected by the 
season, stage of lactation, and health of animal. This subsequently determines the 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  4 
nutritional value of the milk, its flavour, the growth of microorganisms and chemical 
reactions that can occur in milk, as well as the quality of processed dairy products. 
 
Today, the consumption of milk has become part of a regular human diet in many 
regions. Countries such as Finland, Sweden, and Ireland are among the top fluid milk 
drinkers, consuming 183.9 litres, 145.5 litres, and 129.8 litres per capita in 2006, 
respectively (FIL/IDF, 2007). On the other hand, cheeses in Greece, France, and Italy 
dominate milk consumption, with 28.9 kg, 23.9 kg, and 23.7 kg of cheese consumed 
per capita in 2006, respectively. Considering the climate of these regions, it would 
appear that cultures in warmer climates produce more stable products such as cheese 
as means of preservation. 
 
2.2. Cheese 
The history of cheese is believed to have evolved around 6000-7000 BC when the 
ancient human learnt to domesticate animals, a period referred to as the ?Agricultural 
Revolution? (Fox, 1993). This allowed human to secure the provision of milk, leading 
to the discovery of fermented dairy products by accident. 
 
Due to the highly perishable nature of milk, especially in warmer climates, it is prone 
to microbial contamination. Some species of bacteria utilise milk lactose as a source 
of energy to produce lactic acid as a by-product. When sufficient acid has been 
produced to allow the caseins, the principal protein in milk, to reach the region of 
their isoelectric point (? pH 4.6), aggregation occurs and a gel that entraps fat and 
other milk constituents is formed. As this gel structure is disturbed and broken, the 
aqueous phase of the gel, the whey, is separated from the curd. Whereas the whey 
may be consumed immediately, it was realized that the storage of the curd could be 
extended by dehydration and adding salt. Hence, the result is a preserved milk product 
that is durable and easily transportable, and perhaps the very first form of cheese (Fox 
et al., 2000). 
 
Over 2000 varieties of cheeses exist to date and it accounts for about 35% of total 
milk production (Fox et al., 2000). In 2005, the Food and Agricultural Organisation of 
the United Nations (FAO) estimated that the worldwide production of cheese was 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  5 
18.7 million tonnes, and Europe was the largest producer which contributed 9.9 
million tonnes of cheese (FAO, 2008). 
 
The definition of cheese given by the Australia New Zealand Food Standards Code is 
that ?Cheese means the ripened or unripened solid or semi-solid milk product which 
may be coated and is obtained by one or both of the following processes ? (a) 
coagulating wholly or partly milk, and/or materials obtained from milk, through the 
action of rennet or other suitable coagulating agents, partially draining the whey 
which results from such coagulation; or (b) processing techniques involving 
concentration of coagulation of milk and/or materials obtained from milk which give 
an end-product with similar physical chemical and organoleptic characteristics as the 
product described in (a)? (FSANZ, 2008). 
 
2.2.1. Types of cheeses 
The types of cheeses are generally categorised based on their rheological properties 
and moisture content. An often used scheme classifies cheese into hard, semi-hard, 
semi-soft, and soft cheeses. Common examples of hard cheeses include Parmesan and 
Cheddar; semi-hard cheeses include Gouda, Edam, Emmental, and Roquefort; semi-
soft cheeses include Havarti, Mozzerella, and Feta; and soft cheeses include Brie, 
Camembert, Cottage, and Cream (Scott, 1998). 
 
In many types of cheese, some are consumed immediately after manufacture, namely 
fresh cheese. This type of cheese undergoes minimal, if any, ripening after the 
fermentation of lactose. They generally have limited shelf life of about two weeks 
under refrigeration. Examples of fresh cheese include Quarg and Cottage cheese 
(Walstra et al., 2006). 
 
In the majority of the cheese varieties other than the fresh cheeses, the cheeses are 
generally allowed to ripen before consumption. Cheese ripening involves chemical 
changes in the cheese which often begin before curd making finishes. In non-fresh 
cheese types, the process of ripening is allowed to continue, which causes change in 
the composition, structure, and organoleptic properties of the cheese. The 
phenomenon of cheese ripening involves chemical, physical, and micriobiological 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  6 
interactions which are highly dynamic and complex, of which the specific reactions 
and metabolic activities are still not fully understood (Fox et al., 2000). 
 
A common type of hard or semi-hard ripened cheese is Cheddar where the cheddaring 
process is used during manufacture. In Cheddar cheese production, after a sequence of 
manufacturing processes which include curd-heating and curd-cutting, cheddaring is 
applied where the curd is stacked and turned (Walstra et al., 2006). Gouda-type 
cheeses are semi-hard ripened cheeses involving curd-washing and eye-formation. 
Curd-washing in these cheeses enables the lowering of lactose content, which affects 
subsequent acid development. Citric acid fermenting starter bacteria produces CO2 in 
the cheese, which causes eye formation and produces a cheese with holes in the 
cheese matrix (Walstra et al., 2006; Scott, 1998). Swiss-type cheeses utilise high 
temperatures during curd treatment in manufacture and produces a cheese with low 
water content and relatively high pH, such as Emmental, a semi-hard cheese with eyes 
(Walstra et al., 2006; Scott, 1998). Another common type of cheese is white brine 
cheese. They are rindless, highly salted, drained-curd, and low pH (Scott, 1998). 
These cheeses originated from the regions in the Middle East where the climate is hot 
and unsuitable for cheese storage. Hence, they are stored in strong brine for improved 
shelf-life, producing cheeses such as Feta and Domiati (Scott, 1998). 
 
Soft cheeses comprise of a broad range of different curd types, including very soft 
Petit Suisse and Gervais, to firmer Camembert and Brie, and firmer Limburger and 
Romadur (Scott, 1998). These types of cheese are generally discriminated by the 
method of ripening and moisture content. Camembert is one of the most popular 
surface mould ripened cheese in France and elsewhere. It has white mould on the rind 
with a soft and spreadable texture (Scott, 1998; Kristensen, 1999). 
 
Various factors contribute to the development of a type of cheese, which depends on 
the local conditions, including climate, storage environments, the availability and type 
of milk. Additionally, the knowledge in traditional methods that have evolved and 
accumulated experience over generations as well as new findings play a significant 
role (Walstra et al., 2006). 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  7 
2.3. Advances in the dairy industry 
With the advancement in dairy technology over the past century, many process steps 
have been developed and employed in manufacturing of various cheese types. 
Examples of which include the pasteurisation of cheese milk, use of starter cultures, 
use of industrial coagulants, as well as washing of the curd. New data promotes the 
use of surface flora inocula, acceleration of ripening, application of latex to cheese 
rind, and also the use of membrane technology (Walstra et al., 2006). 
 
The emergence of membrane technology has uncovered innovative applications and 
new products in the dairy industry. Its application in cheese making has become 
popular since the invention of the MMV process, which produced cheese by UF of the 
cheese milk (Fox, 1993; Maubois and Mocquot, 1975). 
 
2.3.1. Membrane technology 
The applications in membrane and membrane separation techniques have advanced 
from laboratory to large scale industrial processes, introducing considerable technical 
and commercial challenges. Membrane technology is currently utilised in many 
industrial-sized applications, such as the use of reverse osmosis in potable water 
extraction from the ocean, ultrafiltration in fractionating macromolecular solutions in 
the food and drug industry, recovery of valuable constituents and cleaning of 
industrial effluents by electrodialysis, application in artificial kidney using dialysis in 
removing urea and other toxins in the blood stream, as well as releasing drugs at a 
predetermined rate in medical treatments (Porter, 1990). 
 
Membrane processes can also be customised to modify their properties in specific 
separation applications. The basic principle of a membrane process involves a 
separating barrier and a driving force used to transport chemical components, which is 
similar in all different modes of operations (Porter, 1990). When compared to 
conventional separation methods, membrane processes are generally faster, more 
efficient, and economical. There are numerous advantages associated with the use of 
membrane technology, which includes lower energy consumption, the ability to 
combine with other separation processes to form hybrid processing, separation under 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  8 
mild conditions, the possibility for setting a continuous process, the ability to up-scale 
easily, the high variability and adjustability in membrane properties, and the use of 
additives to products are not required (Mulder, 1996).  
 
The range of expertise involved in membrane technology is broad. Examples include 
the development of membrane structures by polymer chemists; defining transport 
properties of different membrane and their separation characteristics by physicist and 
mathematicians; and designing of large scale separation processes for industrial use 
by chemical engineers (Porter, 1990). 
 
2.3.2. Principles of membrane filtration 
Membrane filtration is applied in many processes which is based on the principle of 
selective permeability of one or more components of a liquid mixture through a 
membrane barrier (FIL/IDF, 1991; Mulder, 1996). There are a number of ways for 
mass transportation through a membrane, commonly facilitated by a significant flux 
of matter such as convection or diffusion, electrical potential differences, 
concentration, pressure, or temperature gradients (Mulder, 1996; Porter, 1990). The 
rate of permeation through a membrane is usually proportional to the driving force, 
hence the flux-force relationship can be described by a linear phenomenological 
equation (Mulder, 1996). 
 
Most membrane processes are driven by a pressure or concentration difference across 
the membrane, or by electrical potential difference. Membrane processes which are 
driven by pressure difference include microfiltration (MF), ultrafiltration (UF), 
nanofiltration (NF), and reverse osmosis (RO) (FIL/IDF, 1991). Membrane processes 
driven by concentration include pervaporation, gas separation, vapour permeation, 
dialysis, diffusion dialysis, and carrier-mediated transport. Membrane processes can 
also be driven by electrical potential differences known as electrodialysis, electro-
osmosis and membrane electrolysis (Mulder, 1996). 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  9 
2.3.3. Pressure-driven membrane processes 
The applications of pressure-driven membrane processes are generally based on 
concentrating a dilute solution and separating out specific molecules (aqueous or non-
aqueous) (FIL/IDF, 1991; Mulder, 1996). In these processes, the solvent is usually in 
a continuous phase and the concentration of the solute is relatively low (Mulder, 
1996). With specific pore size and pore size distribution, the membrane used is 
dependent on the molecular size and chemical properties of the solute. Hence, the 
types of membrane processes are defined by the molecular size of the solute and the 
consequent membrane structure (Porter, 1990; Walstra et al., 2006). 
 
The principle of pressure-driven membrane processes is that due to applied pressure, 
the solvent and various solute molecules permeate through the membrane, while other 
molecules are retained depending on the structure of the membrane (Mulder, 1996). 
Several membrane processes are commonly used, including MF, UF, NF, and RO, 
where the size of the molecules being separated decreases and the pore size in the 
membrane becomes smaller, respectively (Porter, 1990). 
 
2.3.4. Microfiltration 
Microfiltration (MF) is the separation process which is similar to conventional coarse 
filtration. The membrane pore sizes are between 10 to 0.05 ?m with cross-flow 
pressure kept at low levels, usually under 2 bars (Mulder, 1996). ?Cross flow? or 
?tangential flow? filtration is used in order to offset deposition on the membrane 
layers where the fluid flow is tangential to the membrane surface and perpendicular to 
the permeate flow through the membrane (FIL/IDF, 1991). MF is used to selectively 
separate particles with molecular weights greater than 200,000 Da, allowing the 
process to retain suspensions and emulsions. In dairy applications, MF is involved in 
removing somatic cells, bacteria (Elwell and Barbano, 2006), fat globules, casein 
micelles, aggregated whey components (?-casein and ?-lactoglobulin) in milk (Fox et 
al., 2004; Saboya and Maubois, 2000). 
 
In other applications, MF is used in sterilisation and clarification of beverages 
(Hajipour et al., 2010) and pharmaceuticals (Sirkar, 2000), cell harvesting (Roush and 
Lu, 2008), membrane bioreactors for continuous fermentation (Lee et al., 2008), 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  10 
separation of oil-water emulsions (Vahid et al., 2006), plasmaphersis (Mulder, 1996), 
waste-water treatment (Herzberg et al., 2010; Lorain et al., 2010), as well as metal ion 
recovery (Matis et al., 2005). 
 
2.3.5. Ultrafiltration 
Ultrafiltration (UF) membranes generally have pore sizes from 0.05 ?m to 1 nm with 
the cross-flow pressure applied between 1 to 10 bars. UF is often used to retain 
macromolecules and colloids from a solution, typically with molecular weights 
between 1000 to 200,000 Da (Mulder, 1996). The application of UF in milk involves 
separation and fractionation of proteins, fat and colloidal salts from lactose, soluble 
minerals, non-protein nitrogens (NPN), water-soluble vitamins and water (FIL/IDF, 
1991, Fox et al., 2004, Hinrichs, 2001). Other industrial applications of UF include 
the recovery of whey proteins (Yorgun et al., 2008) and potato starch (Zwijnenberg et 
al., 2002), production of ultrapure water for semiconductor industries (Kang, 1999), 
treatment of oil and latex emulsions (J?nsson and Tr?g?rdh, 1990), and paper waste 
treatment (Porter, 1990). 
 
2.3.6. Nanofiltration and reverse osmosis 
Nanofiltration (NF) utilises membranes with pore sizes typically under 2 nm, with 
cross-flow pressures between 10 to 25 bars. NF is used to separate low molecular 
weight solutes between 1000 to 200 Da, such as inorganic salts or small organic 
molecules (e.g. glucose) (Mulder, 1996). Its applications are also dependent on the 
availability of special membranes, allowing it to selectively reject ions based on their 
charge. In the dairy industry, NF is often utilised in partial demineralisation of whey-
like materials (Su?rez et al., 2006). The small pore sizes of the membranes are also 
sufficient to retain most of the lactose, hence NF is often referred as an ultratight UF 
or loose RO (FIL/IDF, 1991). 
 
Reverse osmosis (RO) is used to separate molecular weight solutes with 150 Da or 
less. Hence it is generally utilised in the removal of water or to concentrate a dilute 
solution (FIL/IDF, 1991; Luce and Rex Dieterle, 2005). The operating cross-flow 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  11 
pressures in RO are about 15 to 80 bars, which is relatively high due to the tight 
nature of its membranes (Mulder, 1996). 
 
2.3.7. Application of membrane technology in cheese 
The use of membranes in cheese making began in the late 1960s when the MMV 
process, named after its inventors - Maubois, Mocquot and Vassal, was invented 
(Maubois and Mocquot, 1975). This innovative discovery allowed an opportunity for 
advancement in cheese making technology. The benefits from the incorporation of 
membrane processes in cheese making included improvements in plant efficiency, 
reduced cost of labour and materials, increased plant capacity (Sharma et al., 1989), 
increases in cheese yield (Hinrichs, 2001), development of a continuous process, 
reduced pollution from whey discharge, and development of new cheese varieties 
(Fox et al., 2004).  
 
Practical applications of membranes in cheese include the manufacture of most types 
of cheese from the milk of cows, goats, ewes or water buffaloes, as well as production 
of milk powders with excellent cheese-making properties (Fox et al., 2004). It is also 
used in restoring rennet coagulation properties of ultrahigh temperature-treated (UHT) 
milk (Ferron-Baumy et al., 1991), on-farm concentration of milk, removal of somatic 
cells and bacteria from cheese milk (Maubois, 2002; Thomann et al., 2008), and 
casein enrichment of cheese milk (Fox et al., 2004; Thomann et al., 2008).  
 
2.3.8. Characteristics of cheeses manufactured using membrane processes 
In 1989, it was recorded that over 400,000 tonnes of cheese were made using UF 
(Mistry and Maubois, 1993), and it is currently the most widely used membrane 
process for cheese-making. In the manufacture of cheese using UF or MF, the quality 
of the end cheeses are often determined by specific properties of the protein-enriched 
product (Fox, 1993). 
 
During the UF of milk at its normal pH (about 6.7), mineral salts bound to casein 
micelles are concentrated in the same proportion as proteins. This results in a retentate 
with increased buffering capacity (Fox et al., 2000), which affects a number of basic 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  12 
parameters in cheese-making, such as acidification by lactic acid bacteria (LAB), final 
pH value, coagulation by rennet, rheological properties of the curd, activity of 
ripening enzymes, growth and rate of survival of spoilage micro-organisms, as well as 
water-holding capacity of the cheese throughout ripening (McSweeney, 2007). Hence, 
higher lactic acid production is needed by LAB to obtain the optimal pH, which often 
results in acid-tasting products for most cheese varieties (McSweeney, 2007). 
 
A notable aspect of UF cheese is that whey proteins are incorporated in the curd. The 
amount of whey proteins retained is dependent on the degree of UF concentrate and 
variety of the cheese (Lo and Bastian, 1998). The whey proteins account for 
approximately 20% of the total protein in the cheese if it is retained completely. It 
substitutes part of the casein and acts as an inert filler, which may soften the cheese 
(Harper et al., 1989). Additionally, whey proteins have much higher water-binding 
capacity than casein, thus UF cheeses are often less susceptible to drying than 
traditionally-made cheeses during their shelf-life (Harper et al., 1989; Walstra et al., 
2006). 
 
During the acidification of UF cheese, large quantities of calcium salts are released 
into the aqueous phase of the cheese curd, which increases ionic strength and modifies 
casein aggregation. This sometimes results in cheese with a crumbly or sandy texture, 
as well as poor spreadability and stretching properties (Fox, 1993). Moreover, there is 
significant difference in the rheological properties of UF retentate and milk. Milk is 
considered as a Newtonian liquid. Whereas UF retentate exhibits a pseudoplastic 
behaviour, as its viscosity increases with decreasing shear stress (Walstra et al., 2006). 
Therefore, specific design and operating parameters of the UF system should be 
customised for cheese-making, using suitable UF flux and monitoring the degree of 
membrane fouling (Fox, 1993).  
 
2.3.9. Major approaches in cheese-making using membrane processes 
The production of cheese using UF can be divided into three major categories, which 
include (1) use of protein-standardised cheese milk (Kelly et al., 2008); (2) 
intermediate or medium concentrate retentates (Schreier et al., 2010); and (3) liquid 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  13 
pre-cheeses or often highly-concentrated retentates, where the UF retentate possesses 
the composition of the cheese variety to be made (Fox, 1993). 
 
In the collection of milk at dairy plants, the protein content varies and is affected by 
multiple factors, such as the stage of lactation (Mio? et al., 2008), weather (Roupas, 
2001), feeding and breed of lactating animal (Auldist and Hubble, 1998). These 
variations in the composition of the incoming milk may require adjustment by 
cheesemakers in making consistent quality products (Roupas, 2001). For example, 
low protein content produces weak curds and subsequently high losses of caseins in 
whey (Walstra et al., 2006). In Europe, protein-standardised milk is applied in the 
manufacture of Camembert cheeses utilising the Alpma coagulator (Fox, 1993), 
which produces a continuous supply of curd for moulding. 
 
Various cheese varieties ranging from soft to hard have been made using medium or 
intermediate concentrated retentates, which include the manufacture of UF Cheddar 
(Agrawal and Hassan, 2008), Feta (Jana and Thakar, 1996; Karami et al., 2009), and 
Havarti cheeses (Lo and Bastian, 1998). 
 
When a higher concentration factor is used in UF or MF, a highly-concentrated 
retentate or liquid pre-cheese is produced. The cheese milk is often concentrated to 
the composition of the drained curd being made before the addition of rennet in soft 
cheeses (Maubois, 2002). Hard cheeses on the other hand require a series of UF and 
MF processes, as UF alone does not sufficiently result in the high retentate 
concentration required for hard cheeses without membrane fouling (Schreiera et al., 
2010). Whey drainage in this process is minimal and the use of cheese vats is not 
required. This principle was first applied in the manufacture of Camembert cheese 
(Maubois and Mocquot, 1975), and was successfully developed in producing other 
cheese varieties (Brandsma and Rizvi, 2001; Schreier et al., 2010). Nowadays, a range 
of membrane processing modules are commercially available for the manufacture of 
soft cheeses, and these membranes have been developed based on cellulose acetate 
(Nagendran and Mohan, 2008), polycarbonates (Ko et al., 1993), polysulphonates, 
and zirconium oxide (Shaw, 1981). A continuous procedure known as the ?Camatic? 
process is utilised in a Camembert cheese plant by Alfa Laval, which makes use of the 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  14 
MMV patent and employs a Romicon hollow fibre UF plant with polysulphonate 
membranes (Shaw, 1981; Fox et al., 2004). 
 
In France, UF Camembert cheeses have been associated with psychological and 
commercial difficulties. Although there is no significant differences between the 
organoleptic qualities of UF and traditionally made Camembert cheese, the bulk 
density of the UF cheese paste is much higher than that of traditional cheese due to 
the lack of mechanical openings (Fox, 1993). Therefore, French consumers get the 
impression that they received less cheese for their money, as they are accustomed to 
buying Camembert cheese by the size of the piece and not by weight (Fox, 1993). 
Although UF Camembert was not highly successful on the French market, it led to the 
development of other UF varieties, and the subsequent success in the manufacture of 
UF Feta. In Denmark, Feta accounts for 35% of the total cheese produced and more 
than 90% is produced by UF (Fox, 1993). 
 
While UF cheeses offer considerable yield benefits and have been well established on 
the consumer market, they do inherit unique composition, ripening characteristics, 
organoleptic and textural qualities (Maubois, 2002). It has been suggested that 
separate standards of identity should be used for UF cheeses, and new cheese varieties 
should be developed instead of duplicating traditional varieties (Fox, 1993). 
 
2.4. Camembert cheese 
A small proportion of world cheese production consists of mould-ripened cheeses, 
which have a growing popularity with consumers and increasing demand on the 
market (Fox et al., 2004). Surface mould-ripened soft cheeses such as Camembert and 
Brie; and blue veined cheeses such as Roquefort, Gorgonzola, and Danish Blue are 
examples of mould-ripened cheess. The mould present on the surface (Penicillium 
camemberti) or within the cheese (P. roqueforti) gives the products different 
appearances and produce very distinctive aromas and tastes (Fox et al., 2000). 
Camembert is well known and is the most widely produced surface mould-ripened 
soft cheese, with the growth of a white mould species Penicillium camemberti 
attributing to its characteristic surface rind (Shaw, 1981). 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  15 
2.4.1. Composition, properties and standards 
Under French food law, Camembert is a non-scalded and non-pressed cheese. The 
cheese generally weighs between 210 and 260 g, and must have at least 110 g of total 
solids (TS). The fat content in dry matter is about 45%, but there may be variants with 
lower or higher fat contents (Shaw, 1981). The dimensions of a Camembert explained 
by Shaw (1981) are 10.5 to 11.0 cm diameter with 3.0 cm thickness. 
 
2.4.2. Manufacturing of Camembert 
The manufacture of Camembert cheese and most cheese varieties is fundamentally a 
method of preserving milk over the short and medium term (Shaw, 1981; Walstra et 
al., 2006). The main purpose of this preservation method is to lower the pH and water 
activity, which discourages the growth of spoilage microorganisms (Linton et al., 
2008). The manufacturing process of Camembert and other soft cheese varieties has 
been well-documented (Fox et al., 2000; Walstra et al., 2006; Gunasekaran and Ak, 
2003; Shaw, 1981; Scott, 1998). A schematic of the manufacturing protocol for 
Camembert cheese is shown in Figure 2.1. In comparison to other cheese varieties, 
Camembert has a relatively high pH value and high water activity. Therefore, it 
results in the product having a short shelf-life with maturation being achieved in a 
relatively short time (Shaw, 1981). The lowering of the pH is achieved by controlled 
lactic fermentation, and draining and salting of the curd reduces the water activity. In 
addition to the control of pH and water activity to produce a stable product, further 
controlled transformation of the product during maturation is attained by the micro-
flora and enzymes (Fox, 1993). Coagulation of milk in the manufacture of Camembert 
cheese is obtained by a combination of the lactic acid production by lactic acid 
bacteria (LAB) and a coagulant protease. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  16 
 
Figure 2.1: Manufacturing protocol for Camembert cheese (Walstra et al., 2006). 
 
In the various stages of manufacture, including milk ripening, coagulation, curd 
draining, and maturation, the acidification profile is of key importance (Gunasekaran 
and Ak, 2003). The rate and quantity of lactic acid production affects (1) the level of 
calcium solubilisation in milk, and consequently the rheological properties of the 
coagulum; (2) the rate of whey draining in curd cutting; and ultimately (3) the mineral 
composition and texture of the final cheese (Fox et al., 2000). Additionally, LAB 
starters are important during the maturation of the cheese, as they contribute to the 
overall flavour of the cheese. Therefore in cheese-making, the acid production phase 
should be of a predictable and consistent nature with defined methods of control. 
Hence, it is crucial to select and use LAB cultures of good quality (Fox et al., 2000). 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  17 
The level of proteolysis obtained by the coagulating enzymes is another major factor 
which influences a sequence of events, including the rheological properties of the 
coagulum, whey drainage, and maturation of the final cheese (Walstra et al., 2006). 
The ripening or maturation period for Camembert varies between 15 and 60 days (Fox 
et al., 2000), depending on a number of factors including the moisture content of the 
cheese, the activity of proteases and lipases derived from the rennet, LAB starters, 
surface or internal microflora of the cheese (Fox, 1993), temperature as well as the 
relative humidity of the maturing rooms (Walstra et al., 2006; Gunasekaran and Ak, 
2003). 
 
The methods for the manufacture of Camembert and other soft cheeses have evolved 
over the years, where various recipes may be found showing variations in the amount 
of LAB starters or rennet addition, time or temperature control in different stages, etc. 
The addition of various technological innovations to the traditional methods have 
allowed them to evolve into large-scale industrial processes that are highly 
sophisticated, mechanised, and automated with large milk throughput (Fox, 1993). 
 
2.4.3. UF in Camembert cheese-making 
The application of UF processes in Camembert cheese-making since the invention of 
the MMV process has led the production of Camembert cheese to industrial scale with 
high capacity and automation. An example is the ?Camatic? process which was 
developed by Alfa-Laval for continuous cheese manufacture and is located in its 
Nantes Camembert plant. It has a capacity of 50,000 litres per day or 4,000 
Camembert cheeses per hour (Shaw, 1981). 
 
The ?Camatic? consist of three main stages, comprising the preparation of the pre-
cheese, automated moulding operations and ripening of the cheese (Fox, 1993). The 
procedure of manufacture gives about 12 to 15 % yield increase (Shaw, 1981), and 
involves the following: 
? High-temperature short-time pasteurised milk is ultrafiltered at 50 ?C to a pre-
cheese retentate of 5:1 and a TS content of 35%; 
? The pre-cheese is cooled to 30 ?C and 2 % mesophilic LAB starter and 0.75% 
NaCl are added; 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  18 
? The mixture is acidified to pH 5.5 and is filled into forms by an automated 
process which utilises on-line inoculation with rennet; 
? Curd wheels develop rapidly and are continuously and gently moved for 45 
minutes; 
? The cheeses are turned over once, and a continuous electric current is applied to 
each cheese between the air-exposed surface in contact with a carbon electrode 
and the stainless steel cup holding the cheese. Any whey discharged from the 
cheese is demineralised in this process allowing the whey for uses in other 
foodstuffs (Harju, 1989); 
? Air injector is used to demould the wheels onto cheese trays; 
? The fresh Camembert cheeses are brined for approximately 30 minutes and then 
removed from the brine solution; 
? The cheeses are sprayed with Penicillium camemberti spores and held for 12 
days at 11-12 ?C to allow mould development on the cheese surface (Shaw, 
1981). 
 
2.5. Cheese ripening and quality 
The ripening of cheese includes all the biochemical changes occurring in cheese, 
catalysed by the metabolic activity of the microflora and enzymes derived from these 
organisms or from other sources (Fox et al., 2000). During cheese ripening, the 
structure and composition of the cheese changes, as well as its organoleptic properties 
(Walstra et al., 2006), thus affecting the ultimate quality of the cheese. 
 
2.5.1. Effects of lactic fermentation 
The growth of LAB is important in the manufacture and maturation of cheese. It has 
been indicated that in the fresh cheese, starter bacteria replicate only a few times, and 
their growth in cheese ceases at relatively high pH levels (e.g. 5.7) with variations 
depending on the species and the strains involved (Walstra et al., 2006). Although the 
starter culture ceases to grow, fermentation in the cheese continues and allows further 
reduction in pH. The growth of most LAB may be reduced or stopped by the addition 
of salt to the cheese, with the exception of salt-tolerant LAB contained in cheese milk 
added with high salt content (Walstra et al., 2006). The rate of acid production in 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  19 
cheese is of particular importance, where a predictable and consistent rate of 
acidification is necessary in eliminating variations in subsequent curd syneresis, 
moisture content and texture profile of the cheese (Kristensen, 1999; Walstra et al., 
2006). 
 
The production of lactic acid from lactose is essential for the preservation of cheese. It 
reduces pH and lactose content in the cheese, which help prevent the growth of 
undesirable microorganisms (Walstra et al., 2006). 
 
2.5.2. Proteolysis 
In the process of cheese ripening, proteins are broken down by proteolytic enzymes 
into smaller molecules, ranging from large peptides to free amino acids or ammonia. 
The concentrations of these products differ depending on the type of cheese, starter 
and coagulant used, and conditions of ripening (McSweeney, 2007). 
 
Milk proteinases 
In cheese, milk proteinases (alkaline and acid proteinases) decompose ?s1-, ?s2-, and 
?-casein. Most of the alkaline milk proteinase, or plasmin, is found as inactive 
plasminogen with a small portion being active (Walstra et al., 2006). It has a high 
affinity for casein at the pH of milk, and dissociates from casein at low pH (Bastian 
and Brown, 1996). The enzyme becomes active at high pH and decomposes ?-casein 
more quickly than ?s1-casein, and the metabolism of ?-casein yields various ?-caseins 
(?1-, ?2- , ?3- caseins). Acid milk proteinases have an optimal pH of 5.1 to 5.6 in 
cheese, and metabolise ?s1-casein faster than ?-casein (Walstra et al., 2006). 
 
Milk proteinases contribute to the proteolysis in cheese, which increases the amount 
of soluble nitrogen (SN) compounds, mainly consisting of peptides and small amounts 
of amino acids (Bastian and Brown, 1996). The proteinase content and its precursors 
in the milk can vary among milking times. Heat treatments or scalding processes 
induces partial inactivation of compounds which inhibit plasminogen activators, and 
consequently increases plasmin activity (Grappin et al., 1985). The pH conditions in 
cheese are often unfavourable for significant proteinase activity (Bastian and Brown, 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  20 
1996). However, a high pH cheese such as Camembert due to surface mould activity 
may activate the milk proteinases (Walstra et al., 2006). 
Enzymes from clotting agents 
Clotting enzymes not only have a specific function in milk clotting, but also have a 
significant effect on proteolysis in cheese and the subsequent quality of the product 
(Fox et al., 2000). The amount of clotting enzymes (e.g. calf rennet) retained in the 
cheese affects its action on further proteolysis. This is dependent on the adsorption of 
the enzyme onto the paracaseinate, determined by factors such as the amount of the 
enzyme added to the milk; the pH during curd making of which a higher quantity of 
the enzyme is adsorbed onto the paracasein at lower pH; and the scalding temperature 
of the curd for cheese varieties such as Emmentaler, where higher scalding 
temperatures inactivates a larger portion of the enzyme (Walstra et al., 2006). 
 
The optimum pH of most clotting enzymes, such as calf rennet is approximately 5. 
The metabolism of ?s1-casein is markedly faster than that of ?-casein. Proteolysis by 
clotting enzymes produces SN compounds with a large portion consisting of products 
with a relatively high molar mass, and only a small amount of amino acids (Walstra et 
al., 2006). 
 
Enzymes from lactic acid bacteria 
Starter bacteria produce proteolytic enzymes which play an important role in the 
maturation of ripened cheeses (Grappin et al., 1985). They target large peptides 
formed by clotting enzymes or plasmin, and produce small peptides and free amino 
acids which contribute directly or indirectly as precursors to flavour development 
(Walstra et al., 2006). Generally, any variations in the proteolysis of a cheese type is 
mostly due to variation in the enzyme activity of the starter bacteria. 
 
In the mesophilic starters, some Lactococcus strains are proteinase positive (Prt+), 
which contain a cell envelope proteinase required for growth of the bacteria in milk 
(Foucaud and Juillard, 2000; Walstra et al., 2006). In some strains which are 
proteinase negative (Prt-), they depend on the Prt+ strains to form peptides from milk 
proteins, primarily casein (Walstra et al., 2006; Foucaud and Juillard, 2000). In 
cheese, both ?s1- and ?-casein are targeted, but the rate of peptide formation is slow. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  21 
This is aided by action of clotting enzymes such as chymosin, which produces 
peptides at a rapid rate, and these are further hydrolysed by bacterial enzymes 
(Walstra et al., 2006). Additionally, smaller peptides are transported into the bacterial 
cell where specific peptidases immediately hydrolyse the peptides into amino acids, 
and they are subsequently diffused out of the cell. The peptidases in the bacterial cells 
include endopeptidases, aminopeptidases, di- and tripeptidases, and proline-specific 
peptidases (Walstra et al., 2006). 
 
The production of peptides and subsequently the formation of amino acids are of 
significant importance in the flavour profile of cheese, as accumulation of peptides 
might lead to a bitter off flavour (Kristensen, 1999). Some cultures have the ability to 
degrade bitter-tasting peptides, and are often used in the manufacture of various 
cheese types. This ability has been found to have positive relationship with the 
sensitivity of the cell to lysis (Walstra et al., 2006). 
 
In order for the bacteria to perform peptidolysis, energy is required to transport the 
peptides into the cell and convert them into amino acids. The energy source comes 
from lactose, which is usually fully consumed within 24 h in most cheeses. Therefore, 
lysis of the starter bacteria is required to release their intracellular peptidases into the 
cheese (Exterkate, 2006; Hannon et al., 2006). This can be achieved by bacterial 
autolysins with a subsequent osmotic shock to break down the cell wall, usually 
induced by a steep salt gradient (Walstra et al., 2006). Membrane permeabilisation 
techniques have also been studied to release the intracellular peptidases, through 
delipidating action of n-butanol (Exterkate, 2006). 
 
The quantity of enzymes produced by the starter bacteria can vary depending on the 
growth conditions. In the case of cheese milk that has been highly concentrated by 
ultrafiltration or undergone intense heat treatment (10 min at 120 ?C), the amount of 
cell envelope proteinases and some of the intracellular peptidases are significantly 
reduced (Walstra et al., 2006). Thermophilic starters, such as Streptococcus 
thermophilus (Simov and Ivanov, 2005) and various Lactobacillus species, such as L. 
helveticus (Richoux et al., 2009), L. delbrueckii ssp. bulgaricus (Moreira et al., 2003), 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  22 
and L. casei (Ong et al., 2007) also demonstrated similar proteolytic activities in 
cheese. 
 
Enzymes from non-starter microorganism 
Contamination of the cheese milk by non-starter LAB often involves the 
Lactobacillus species. They do not utilise sugar as a carbon source but feed on amino 
acids or the carbohydrate moiety of glycoproteins in the membrane of fat globules 
(Walstra et al., 2006). These bacteria are often highly proteolytic and can produce a 
wide variety of flavour compounds in cheese, of which they are usually undesirable. 
Therefore, milk pasteurisation and good hygienic practices during cheese-making can 
mostly prevent the growth of such microorganisms (Walstra et al., 2006). However, in 
the production of certain types of cheese using traditional methods, raw milk is used 
to enhance the more typical flavours of these cheese types (Fox et al., 2000).  
 
In most cheese varieties, a secondary microflora is inoculated in the cheese milk or 
curd. An example is Penicillium camemberti on surface mould cheeses such as 
Camembert and Brie (Fox et al., 2000). P. camemberti produces aspartyl and 
metalloproteinases which metabolises ?s1- and ?-caseins (Fox, 1993). Two 
extracellular proteinases, including an acid proteinase and a neutral one, are produced 
by P. camemberti. The latter one induces large increases in pH 4.6-SN, but does not 
have much effect on the production of free amino acids (Boutrou et al., 2006). 
Additionally, an exocellular acid carboxypeptidase and a neutral carboxypeptidase 
which is bound mycelially, and an exocellular aminopeptidase are synthesised by P. 
camemberti, of which have impact on the formation of free amino acids in soft 
cheeses such as Camembert (Boutrou et al., 2006). 
 
2.5.3. Lipolysis 
The concentration of active native lipoprotein lipases in milk is dependent on the 
pasteurisation process. Approximately 10 to 15% of the lipase remains after usual 
pasteurisation conditions (Walstra et al., 2006). The activity of lipase in raw milk is 
relatively stable, but acidification and increasing salt content of the milk or curd 
usually leads to its deactivation (Fox et al., 2000). Lipolysis significantly increases 
with temperature, and also with the homogenisation of the cheese milk. This is due to 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  23 
changing of the surface layer of fat globules, of which the rate of lipolysis returns to 
its normal level when the surface layer stabilises (Walstra et al., 2006). 
 
Generally, LAB are not very lipolytic when compared to other bacteria such as 
Pseudomonas, and moulds (Fox et al., 2000; Walstra et al., 2006). However, when 
lactococci and lactobacilli are present in high numbers over a long period, their 
lipases and esterases become the principal lipolytic agents in Cheddar and Dutch-type 
cheeses (Fox et al., 2000). In cheese types such as Romano, Parmesan and Blue, they 
are characterised by extensive lipolysis. 
 
2.5.4. Flavour development 
The balance of different flavour compounds is essential to the flavour of the cheese. 
The flavour compounds originating from the fat and other milk components are weak 
in the curd. The sweetness from lactose in the curd disappears rapidly due to lactic 
fermentation, which produces an acidic flavour in the cheese (Walstra et al., 2006). 
The flavour compounds (e.g. diacetyl) produced by the starter bacteria are particularly 
important in fresh-type cheeses. In ripened cheeses, salt provides additional flavour, 
and its concentration varies between cheese varieties (Kristensen, 1999). 
 
Degradation of protein 
The flavour of the cheese undergoes major changes during ripening. The degradation 
of protein results in the formation of short-chain peptides and free amino acids, which 
contribute to the flavour profile of the cheese (Boutrou et al., 2006). The conditions of 
ripening can lead to variation in specific tastes and their relative intensities. A bitter 
flavour may be developed if a specific way of protein degradation is encouraged, 
resulting in the accumulation of short-chained hydrophobic peptides (Walstra et al., 
2006). 
 
The conversion of amino acids by deaminases, decarboxylases, transaminases, and 
lyases also play an important role in developing more complex flavours in cheese 
(Figure 2.2) (Walstra et al., 2006). Deaminases and decarboxylases yield amines and 
ammonia from amino acids. Alpha-keto acids are formed from transamination of 
amino acids, which may be subsequently converted into aldehydes by 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  24 
decarboxylation, and then into alcohols or carboxylic acids by dehydrogenation 
(Walstra et al., 2006; Fox et al., 2000). The majority of these volatile compounds are 
odour active. Aromatic amino acids, branched-chain amino acids, and methionine are 
important substrates for the development of flavour in cheese. They can be converted 
into compounds such as p-cresol and indole by lyases, which contribute to off-
flavours. Branched-chain amino acids can be converted into compounds such as 
isobutyrate, isovalerate, and 3-methylbutanal often with the presence of Lactococcus 
lactis strains (Walstra et al., 2006). Sulphur compounds produced from methionine by 
several lyases contribute to the typical flavour of many cheese varieties. Other notable 
compounds formed from methionine are thiaalkanes (e.g. 2,4 dithiapentane), which 
gives a distinctive garlic flavour in Camembert cheese (Fox et al., 2000; Walstra et 
al., 2006). Additionally, cheese consistency and mouthfeel are affected by protein 
degradation, where cheese consistency is correlated with flavour perception (Hersleth 
et al., 2005). 
 
Figure 2.2: Overview of general protein conversion pathways for flavour 
formation in cheese (Walstra et al., 2006). 
 
Degradation of lipids 
Free fatty acids (FFA) produced in the lipolysis of fat are also involved in the flavour 
perception of cheese. Depending on the concentration and perception threshold of the 
volatile fatty acids, they can contribute to the aroma of the cheese, or to a rancid 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  25 
defect (Poveda et al., 2008). The FFA act as precursors of flavour compounds such as 
methylketones, alcohols, lactones, and esters (Figure 2.3). 
 
The effect of LAB on fatty acids conversion is relatively small, but in mould-ripened 
cheeses (e.g. Camembert and Roquefort) the fat-derived flavour compounds play a 
significant role (Figure 2.3) (Walstra et al., 2006). The mould in these cheeses 
produces ?-keto acids through the ?-oxidation of fatty acids, which is subsequently 
decarboxylated to produce methylketone, characterised by its fruity, floral and musty 
notes. Secondary alcohols can be formed by the reduction of methylketones, including 
heptan-2-ol and nonan-2-ol. Together with the methylketones, they represent 20 to 
30% of all aroma compounds in Camembert-type cheeses (Walstra et al., 2006). 
Meanwhile, primary alcohols are formed from the metabolism of lactose and amino 
acids, which is described by its mild and fruity notes. Eight-carbon aroma compounds, 
such as oct-1-en-3-ol, are derived from linoleic and linolenic acids by lipoxygenase 
and hydroperoxide lyase found in moulds, which is responsible for a mushroom note 
in cheese (Walstra et al., 2006). In addition, cheeses consist of a wide range of esters, 
usually described as fruity and floral notes, and are formed by esterases present in 
most microorganisms involved in cheese ripening (Fox et al., 2000; Coolbear et al., 
2008). 
 
Figure 2.3: Formation of flavour compounds from fat in cheese (Walstra et al., 
2006). 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  26 
2.5.5. Texture development 
In the microstructure of a fresh cheese immediately after curd making, the cavities in 
the matrix of the paracasein micelles are mostly filled with fat globules and a small 
amount of whey (Gunasekaran and Ak, 2003). The moisture is still able to move 
freely through the network. However, the matrix alters within a day and becomes 
more homogeneous. Whey is drained from the cavities and it has become increasingly 
difficult for moisture to move freely (Walstra et al., 2006). This can be attributed to 
the dissolution of calcium phosphate and subsequent proteolysis. When the cheese is 
matured for a long period, fat globules become partly fused and aggregated (Walstra 
et al., 2006; Gunasekaran and Ak, 2003). 
 
During ripening, various changes in the cheese composition occur. Proteolysis causes 
the uptake of water and the formation of ionic groups (-COO- and ?NH3
+), production 
of ammonia and subsequently increase in pH. The rate of such change depends on 
several conditions, such as the size and shape of the cheese, amount of retained 
rennet, amount and type of bacteria and enzymes released, pH, surface flora, and 
ripening conditions (Walstra et al., 2006). The changes in consistency over the 
ripening period are dependent on the type of cheese. Cheese with a high water content 
and low initial pH may start off as being short and firm, but as the pH increases over 
time (e.g. 5.4), the consistency becomes rubbery. This may be attributed to proteolysis 
which causes the cheese to develop a soft and plastic, and in some cases, liquid-like 
texture at higher pH levels by degrading the casein responsible for holding the cheese 
structure (Walstra et al., 2006; Gunasekaran and Ak, 2003). 
 
2.6. Analysis of cheese 
2.6.1. Determination of pH 
In cheese manufacture and ripening, pH is useful in predicting the activity of the 
microflora and their enzymes in the milk, curd, and cheese (Gunasekaran and Ak, 
2003). The rate of change in pH is usually of a predictable and consistent nature in 
cheese-making. If, for example, the pH decreases too rapidly before the addition of 
starter or does not decrease during acidification, it is likely that a contamination is 
present in the cheese milk (Fox et al., 2004). During the ripening of Camembert, the 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  27 
cheese usually has an initial pH of about 4.8 which increases to about 7.5 (Fox, 1993). 
This is due to the catabolism of lactic acid and the production of ammonia by 
proteolysis, and pH serves as an indicator for the age of the cheese during maturation 
(Walstra et al., 2006). The measurement of pH is commonly carried out using a pH-
meter with various electrodes (Kosikowski and Mistry, 1997). Additionally, specific 
electrodes designed for the measurement of pH in cheese and dairy products are 
available. 
 
2.6.2. Total solids 
Total solids (TS) in cheese consist of fat, proteins, lactose, minerals and ash (Fox et 
al., 2000). The TS content in cheese is expected to increase during ripening as a result 
of surface evaporation, exchange of volatile compounds, and action of proteolysis 
(Walstra et al., 2006). TS content is determined by the loss of moisture through drying 
of the cheese sample using a hot-air or vacuum oven at 105 ?C (AOAC International, 
2005). 
 
2.6.3. Salt content 
In cheese-making, salt is used to inhibit or slow down the growth of microorgasms 
and the activity of enzymes after curd making. It also reduces moisture content in the 
cheese and contributes significantly to cheese flavour (Walstra et al., 2006). The 
effect of salting on the ripening and subsequent flavour and texture of the cheese 
depends on the cheese variety and the method used (Scott, 1998). The Volhard 
method is used to determine the salt content in cheese (Krik and Sawyer, 1991). The 
salt reacts with standard silver nitrate solution in excessive amounts and the unused 
silver nitrate is titrated against potassium thiocyanate solution. 
 
2.6.4. Fat content 
Fat content contributes significantly to the flavour and texture of cheese. The 
Mojonnier method is commonly used for determining fat in cheese, which is 
described in the AOAC Official Methods (AOAC International, 2005). The modified 
Schmid-Bondzynski-Ratzlaff method is also very similar but requires slight 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  28 
alterations. Generally, the fat in cheese is extracted using ethanol, diethyl ether and 
petroleum ether, where the solvents are subsequently evaporated and the extracted fat 
is weighed (AOAC International, 2005). 
 
2.6.5. Nitrogen content 
The extent of proteolysis in cheese can be determined by the analysis of the N content 
in different fractions of the cheese. This includes different N fractions which enables 
the measurement of different activities in cheese (Ard?, 1999; Butikofer et al., 1993). 
The fractionation scheme described by Ard? (1999) comprises the total nitrogen (TN); 
pH 4.4-SN, which consists of all proteins excluding casein, and non-protein nitrogens 
(NPN) including peptides, amino acids and smaller N compounds soluble at pH 4.4 to 
4.6; and TCA-SN, which consists mainly of NPN including medium sized to small 
peptides, amino acids, and smaller N compounds (Ard?, 1999). Fractionation of the N 
compounds in cheese allows the monitoring of proteolysis throughout cheese 
maturation, as the caseins in cheese are continuously broken down into N compounds 
by enzyme action (Fox et al., 2000). The Kjeldahl method is subsequently used to 
determine the N content in each fraction, which involves extraction, digestion, and 
titration (AOAC International, 2005). 
 
2.6.6. Texture analysis 
Texture is an important quality which determines the identity of a cheese and its 
preference by consumers (Antoniou et al., 2000). Texture measurement techniques 
can be categorised as either subjective or instrumental (Gunasekaran and Ak, 2003). 
Instrumental methods are often used to mechanically mimic the sensory evaluation of 
human assessors, without being subjective and biased. Texture profile analysis (TPA) 
is used to imitate the grading action of the jaw, by inducing two successive 
deformations in the food sample (Gunasekaran and Ak, 2003). It enables the 
explanation of primary mechanical properties such as hardness, cohesiveness, 
viscosity, elasticity, and adhesiveness, and secondary properties such as brittleness, 
chewiness, and gumminess (Gunasekaran and Ak, 2003; Antoniou et al., 2000). 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  29 
In the other texture analyses which do not follow the TPA protocol, uniaxial 
compression test is often used in measuring cheese properties (Gunasekaran and Ak, 
2003). It is simple and measures some of the most important properties in cheese, 
such as firmness and springiness (Fox, 1993). The method usually involves the 
measurement of the force required to compress or deform the cheese sample to a 
certain percentage of its original height (Antoniou et al., 2000; Wium et al., 1997). 
Attributes such as firmness, fracturability, springiness, chewiness and adhesiveness 
can be calculated based on the peak forces and work required (Fox et al., 2004). 
 
2.6.7. Sensory analysis 
The most important sensory properties of cheese ripening are those of which affects 
the consumer in purchasing the product, including flavour, aroma, texture, and 
appearance (Fox et al., 2000). 
 
Cheese grading 
Sensory analysis of cheese has been traditionally performed by cheese graders or 
experienced experts in determining the quality of the cheese at the point of sale. In 
these systems, points are usually deducted for perceived defects with respect to their 
intensity. The characteristics sought after by expert judges are usually desired by 
general consumers, but the scores or grades awarded do not always correlate with 
consumer preference (Meilgaard et al., 2007). The expert grading systems are 
commonly used in the cheese industry to ensure consistent product quality and may be 
used for pricing (Fox et al., 2000). 
 
Discriminative analysis 
The use of discrimination tests, such as the paired comparison, duo-trio, and triangle 
tests enables one to assess whether there are differences between cheeses in terms of 
certain sensory characteristics (Meilgaard et al., 2007). The protocols of these tests are 
very similar, where two samples (paired comparison test), two samples with a 
reference sample (duo-trio test), or three samples (triangle test) are evaluated based on 
specific attributes or overall likings (Meilgaard et al., 2007). This type of sensory 
analysis requires trained panellists and is dependent on the perception threshold of 
individual panellists on specific sensory stimuli (Fox et al., 2004). 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  30 
 
Descriptive analysis 
Descriptive evaluation requires a panel of trained assessors and is used to determine 
the sensory profile of the cheese tested. It involves three main stages: selection of 
potential assessors based on their ability to perceive sensory stimuli; determining a list 
of sensory terminology or ?descriptors? which describe the sensory attributes of the 
cheese; and defining the processes to quantify the intensity of each descriptor (Fox et 
al., 2000). Assessors in the panel are required to differentiate between atrributes, give 
reproducible scores, and agree with other panellists (Meilgaard et al., 2007). 
 
Consumer acceptance 
An acceptance test is used to determine the ?affective status? of a product, usually 
measured by a hedonic line, intensity, or magnitude estimation scales with degrees of 
acceptance or liking. This type of test usually involves untrained consumer groups 
which are highly diverse, and the tests easily understandable. The tests are similar to 
attribute difference tests except that the attribute in this case is acceptance or liking 
(Meilgaard et al., 2007). 
 
2.6.8. Microbiological analysis 
The consumption of contaminated cheese can lead to food illnesses. In 2006, 
contaminated cheese accounted for about 0.4% of the total food outbreaks in Europe 
(Kousta et al., 2010). Common food pathogens associated with food outbreaks include 
Listeria monocytogenes, Samonella sp, Escherchia coli, and Staphylococcus aureus 
(Kousta et al., 2010). Cheeses are food products that are ready for direct consumption 
without any further treatment to ensure their safety. Contamination of cheese by 
pathogens can occur at several stages and it is therefore important that strict 
microbiological analyses are carried out to as part of a food control system. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  31 
2.7. Conclusions 
Membrane technology has become widely used with the invention of the MMV 
process in producing Camembert cheese. Manufacture of cheese using UF has shown 
to exhibit different properties compared to traditional processes, including the 
increase in buffering capacity, incorporation of whey proteins in cheese, and changes 
in acidification profile. The ?Camatic? procedure employed by Alfa-Laval 
demonstrated the success in producing UF Camembert cheese using a highly 
automated and continuous process. The properties of ripened cheeses, especially in 
those of surface mould-ripened soft cheeses, are highly dynamic and are affected by a 
number of factors. Learning these fundamentals is essential in understanding the 
different quality attributes present in UF cheeses. 
 
The factors which affect cheese ripening and its subsequent quality include lactic 
fermentation of the cheese; proteolysis and lipolysis induced by a variety of enzymes; 
the development of flavour and texture. Proteolysis in cheese is facilitated by different 
types of enzymes including milk proteinases, and enzymes of clotting agents, LAB, 
and non-starter microorganisms, which act on the paracasein structure of the cheese 
and produce smaller peptides and N compounds. Lipolysis in cheese liberates volatile 
compounds which contribute greatly to the aroma and flavour of cheese. Some 
secondary floras such as Pseudomonas and moulds have high lipolytic activity which 
gives certain cheese types their typical flavour. The degradation of protein and fat 
globules by the fore-mentioned enzymes lead to the production of smaller peptides, 
amino acids and other N compounds, as well as FFA and other FFA-derivatives which 
contributes to the flavour and aroma of the final cheese. The breakdown of the cheese 
microstructure also causes changes in texture and consistency of the cheese, which are 
of significant importance in consumer preference. 
 
Although production of Camembert using UF has been highly successful, the 
processing parameters are often confidential for commercial purposes and vary 
between processing plants, ingredients, and product requirements. This study aims to 
determine the optimum processing parameters for the production of Camembert 
cheese from bovine milk utilising a hollow-fibre UF system, with respect to 
improving the overall quality and shelf-life of the cheese.
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  32 
3. MATERIALS AND METHODS 
3.1. Experimental design 
In this study, an incomplete block design was used to investigate the effect of various 
variables in the manufacture of Camembert cheese on their impact on the final cheese 
quality. The manufacturing process was carried out at Goodman Fielder NZ Ltd Puhoi 
Valley Cheese plant. The design involved four variables: level of fat (in fat/protein 
ratio), salting method, level of acidification (determined by pH), and moulding type. 
Each variable contained two levels, resulting in sixteen unique treatments for the 
Camembert cheeses. The treatments were divided into four separate batches (no. 1-4, 
5-8, 9-12, and 13-16, respectively) each containing four treatments (Table 3.1) as it 
was the maximum handling capacity per batch. Ideally, all sixteen treatments should 
be randomised across the four batches, but different levels of fat cannot be used 
simultaneously in a single batch and required scheduling of the processing plant to 
coordinate with mainstream products manufacture. The experiment was partially 
duplicated with all low fat treatments being identically repeated. The treatments were 
coded by underlining as follows: Low-fat/High-fat; Brine-salted/Retentate-salted; pH 
5.2/pH 4.9; Tube mould/Small mould (e.g. LB5.2T represents low fat, brine salting, 
pH 5.2, tube mould). The underlined letters and numbers indicate the codes used. 
 
Table 3.1: Incomplete block design of sixteen unique treatments for the 
manufacture of Camembert cheese 
No. Treatment 
Codes 
Milk F/P 
Ratio 
Salting  
Method 
Extent of 
Acidification 
Mould Type 
1 LB5.2T F/P 1.32:1 Brine-salted pH 5.2 Tube mould 
2 LB4.9S F/P 1.32:1 Brine-salted pH 4.9 Small mould 
3 LR5.2S F/P 1.32:1 Retentate-salted pH 5.2 Small mould 
4 LR4.9T F/P 1.32:1 Retentate-salted pH 4.9 Tube mould 
5 LB5.2S F/P 1.32:1 Brine-salted pH 5.2 Small mould 
6 LB4.9T F/P 1.32:1 Brine-salted pH 4.9 Tube mould 
7 LR5.2T F/P 1.32:1 Retentate-salted pH 5.2 Tube mould 
8 LR4.9S F/P 1.32:1 Retentate-salted pH 4.9 Small mould 
9 HB5.2T F/P 1.59:1 Brine-salted pH 5.2 Tube mould 
10 HB4.9S F/P 1.59:1 Brine-salted pH 4.9 Small mould 
11 HR5.2S F/P 1.59:1 Retentate-salted pH 5.2 Small mould 
12 HR4.9T F/P 1.59:1 Retentate-salted pH 4.9 Tube mould 
13 HB5.2S F/P 1.59:1 Brine-salted pH 5.2 Small mould 
14 HB4.9T F/P 1.59:1 Brine-salted pH 4.9 Tube mould 
15 HR5.2T F/P 1.59:1 Retentate-salted pH 5.2 Tube mould 
16 HR4.9S F/P 1.59:1 Retentate-salted pH 4.9 Small mould 
Notes: Milk F/P Ratio = milk fat : protein ratio 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  33 
3.1.1. Description of processing variables 
Level of fat (Milk fat/protein ratio) 
The level of fat in cheese can significantly affect its flavour and texture (Gunasekaran 
and Ak, 2003). Therefore, fat content in the pre-cheese retentate was adjusted at two 
levels. The average F/P ratio in the non-standardised milk used in Camembert cheese 
production was 1.32:1 and this was used as the lower level. The higher level of F/P 
ratio was adjusted to approximately 1.59:1. This was equivalent to the ratio used for 
single cream Camembert cheese produced by the conventional method. Adjusting the 
desired level of milk F/P ratio was achieved by the addition of pasteurised cream to 
the pre-cheese retentate. 
 
Salting method 
The two salting methods used were addition of dry salt to the pre-cheese retentate and 
brine immersion of the cheese after moulding. Retentate-salting involved the addition 
of dry salt into the pre-cheese retentate prior to rennet and formation of cheese blocks. 
Brine salting was done by immersing whole cheese units (3 cm height ? 8 cm 
diameter in brine tanks with about 95% saline solution for 10 minutes after the cheese 
blocks were formed and the desired pH was achieved by acidification. 
 
Extent of acidification 
The acidification stage during cheese production is important as it determines the 
ultimate pH of the cheese, which in turn affects the growth of starter cultures and 
formation of the casein network (Fox et al., 2000). Pre-cheese retentate was placed in 
the acidification chamber at warm temperatures where they were allowed to form 
blocks of cheese. The amount of time the cheese was left in the chamber determines 
the ultimate pH, being either pH 5.20 ? 0.05 or 4.90 ? 0.05. 
 
Mould Type 
Two types of moulds were used: a cylindrical plastic tube mould measuring 35 cm 
height ? 8 cm diameter (tube moulds); and silicone tray mould measuring 3 cm height 
? 8 cm diameter (small moulds). The pre-cheese retentate was poured into the 
cylindrical plastic tube moulds lined with plastic film. When the cheese was formed, 
they were removed from the moulds and plastic film. Cheese formed in tube moulds 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  34 
were sliced with a wire cutter, and then loaded onto metal racks for further 
processing. Meanwhile, cheese formed in small moulds were removed from the 
moulds using a blunt knife, and they were loaded onto metal racks. 
 
3.2. Production of Camembert 
Pilot-scale production of the cheeses was conducted at Goodman Fielder (NZ) Ltd. 
Puhoi Valley Cheese plant. Seventy-two blocks of Camembert cheeses (125 g) were 
produced for each treatment. 
 
Preparation of pre-cheese retentate  
As part of the quality control system, cheese milk was analysed at the Puhoi Valley 
Cheese laboratory to ensure that the chemical and microbiological quality were within 
acceptable limits before being processed (Figure 3.9). The milk was pasteurised by a 
heat exchanger at 78 ? 2 oC for 15 seconds. A sample of the pasteurised milk was 
analysed for coliform count to ensure it was less than 10 cfu ml-1. The milk was then 
pumped through a hollow fibre ultrafiltration unit and re-circulated until a TS content 
of 41.5 ? 1.0 % was reached, which was determined by the Lactoscope FTIR 
Advanced (Delta Instruments BV, Drachten, Netherlands) and its software (FTIR 
Scope Advanced). The retentate was homogenised at 19.31 bars, re-pasteurised at 78 
? 2 oC for 15 seconds, and cooled to 38 ? 1 oC in the cheese vats. A sample of the 
retentate was tested for coliforms level. 
 
Addition of culture 
Commercial freeze dried starter cultures (DELVO-ADD? 100-X DSF; 170 g 100 L-1 
retentate, and DELVO-TEC? T30A DSF; 186 g 100 L-1 retentate, DSM, Heerlen, 
Netherlands) containing a mixture of Lactococcus lactis ssp., Leuconostoc sp., and 
Streptococcus thermophilus respectively was rehydrated in ultra-high temperature 
(UHT) treated milk at 38 ? 1 oC for 60 minutes. A commercial freeze dried mould 
culture (CHOOZIT? PC12; 5 doses 100 L-1 retentate, Danisco, Copenhagen, 
Denmark) containing Penicillium camemberti was rehydrated in UHT milk at 4 ? 1 oC 
for 24 hours. Both rehydrated starter and mould cultures were added to the retentate 
and incubated for 30 minutes and the pH decreased to below 6.50. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  35 
Introduction of process variables 
A trial fraction of the retentate (80 L) was obtained at this stage to conduct the 
experiment. The portion of retentate was divided into four treatments (20 L) for every 
batch (Figure 3.2), following the experimental design in Figure 3.1. In treatments 
involving lower fat contents, the retentate was not adjusted. Treatments involving 
higher fat content were subjected to the adjustment of fat/protein ratios of 
approximately 1.59:1 by adding 40% cream. The F/P ratio was analysed with the 
Lactoscope FTIR Advanced. Consequently, in treatments that involved salting in the 
retentate, 300 grams of dry salt was added per treatment of retentate (20 L). Salt was 
not added to treatments intended for brine salting at this stage. Fungal protease 
(Fromase?, DSM, Heerlen, Netherlands) was added to the retentate. All mixing of 
ingredients were done in a pour tank (Figure 3.3) and the mixture was then pumped 
into either silicone tray small moulds (Figure 3.4) or plastic tube moulds (Figure 3.5). 
A sample of the retentate was analysed for coliform level. The moulded retentate were 
left to set in a temperature controlled chamber at 30 ? 1 oC for 6 hours, 27 ? 1 oC for 
18 hours, and then 22 ? 1 oC for 24 hours. During the 48-hour cheese setting period, 
treatments with lower acidification were removed from the chamber at pH 5.20 ? 
0.05. Treatments with higher acidification were removed at pH 4.90 ? 0.05. Plastic 
tubes were removed from the young cheese logs and they were sliced by a wire cutter 
(Figure 3.6) to approximately 125 g units. Young cheeses in the small moulds were 
removed from the silicone trays. All young cheeses were loaded onto metal racks 
(Figure 3.7) and treatments which require brine salting were immersed in brine tanks 
(Figure 3.8) with about 95% salinity for 10 minutes before being placed in the 
ripening room together with other treatments. 
 
Ripening, packing and storage 
All the young cheeses were ripened at 18 ? 1 oC for two days and then 13 ? 1 oC for 
six days in 95% humidity. Each cheese was turned every two days. At the end of the 
ripening period, the cheeses were cooled to 7 ? 1 oC for 12 hours before they were 
packed (Figure 3.10). The cheeses were stored in 4 ? 1 oC chilling chambers at 
Massey University for seven weeks. They were sampled on a every 7 days and 
monitored for pH, TS, salt content, fat content, nitrogen fractions, texture and sensory 
attributes. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  36 
Retentate 
 
 
No salt 
 
Salted 
 
Small Mould 
 
 
pH 5.20 
 
 
 
 
 
Brining 
Tube Mould 
 
 
pH 4.90 
 
 
Cutting 
 
 
Brining 
Tube Mould 
 
 
pH 5.20 
 
 
Cutting 
Small Mould 
 
 
pH 4.90 
 
Figure 3.1: Set-up of the four processing variables applied to four treatments per 
block of Camembert cheese production. 
 
 
Figure 3.2: A small batch of retentate (80 L) was separated from the main batch 
to carry out trial experiment with four treatments (20L each). 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  37 
 
Figure 3.3: A pour tank with pump which circulates and allows the mixing of 
retentate with other ingredients. The pre-cheese retentate was then pumped into 
different moulds. 
 
 
Figure 3.4: Small silicone tray moulds containing pre-cheese retentate. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  38 
 
Figure 3.5: Tube moulds containing pre-cheese retentate in plastic films. 
 
 
Figure 3.6: Wire cutter used for slicing young cheese logs removed from tube 
moulds into approximately 125 g units. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  39 
 
Figure 3.7: Individual young Camembert cheeses loaded onto metal racks. 
 
 
Figure 3.8: Brine tank used to immerse young cheeses. Stirring of the excess salt 
at the bottom of the tank was done prior to the immersion of the young cheeses, 
which ensures the salinity is above 95%. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  40 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 3.9: Process flow chart of Camembert cheese production using 
ultrafiltration, from milk reception to culture addition. 
 
Ultrafiltration 
Storage 
Pasteurisation 
 Incubation 
Incubation time   30 min 
Incubation temp.    38
o
C 
38
o
C 
Coliform <10 cfu ml
-1
 
Mould 
Danisco PC12 
Penicillium camemberti 
5 doses 100 L
-1
 retentate 
 
LAB Starter 
DELVO-ADD 100-X DSF 
Lactococcus lactis ssp. 
Leuconostoc sp. 
170 g 100 L
-1
 retentate 
 
DELVO-TEC T30A DSF 
Streptococcus thermophilus 
186 g 100 L
-1
 retentate 
UHT Milk 
Setting time   60 min 
Setting temp. 38.5 
o
C Culture UHT Milk 
Rehydration time   24 h 
Rehydration temp  4
o
C 
Permeate 
Pasteurisation temp.  78.0 ? 2
o
C 
Holding time       15 sec 
Outlet temp.       45.0 ? 1
o
C 
Recirculation temp.    24.5
o
C 
Flow rate       9 m
3
h
-1
 
Configuration   Hollow fibre 
Membrane type   Polysulfone 
Flux    >260 m
3
h
-1
 
 
Target TS   41.5% 
Temperature   45-55
o
C 
Boost pressure   2.10 bar 
Feed pressure   1.10 bar 
Permeate flow   2.0-2.6 m
3
h
-1
 
End permeate flow    >100 m
3
h
-1
 
Homogenisation & 
Repasteurisation 
Homogenisation    19.31 bar 
 
Cooking temp.   78.0 ? 2
o
C 
Holding time   15 seconds 
Cooling temp.   38.0 ? 1
o
C 
Retentate 
38
o
C 
Coliform         <10 cfu ml
-1
 
Phosphatase <7 Lovibond 
Temperature       ?5
o
C 
Storage time       ?48 hours 
Temperature  ?12
o
C 
Age  <2 days 
Organoleptic Fresh 
Foreign matter Absent 
 
 
 
 
Laboratory Analysis 
Fat      3.5-5.5% 
Protein      3.5-4.5% 
Total solids     12-16% 
Titratable acidity     <0.15% 
pH       6.50-6.90 
Phosphatase      <10 
Lovibond 
 
Somatic cell count  ?400,000 ml
-1
 
Coliforms    <500 g
-1
 
Aerobic plate count <100,000 g
-1
 
Thermoduric     <1500 g
-1
 
 
Milk Reception 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  41 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 3.10: Process flow chart of Camembert cheese production using 
ultrafiltration, continuing from culture addition to the finished product. (*) 
represents process variables which were manipulated. 
?7
o
C 
Packing 
8 days 
Total solids    40-45% Camembert Cheese 
Coliform <50 cfu g
-1
 
Escherichia coli <10 cfu g
-1
 
Staphylococcus aureus <10 cfu g
-1
 
pH 4.90-5.20 
Pouring/Moulding 
Acidification 
38
o
C 
Coliform <10 cfu ml
-1
 
Stage 1 ? 30
o
C for 6 hours 
Stage 2 ? 27
o
C for 18 hours 
Stage 3 ? 22
o
C for 24 hours 
Ripening 
Ripening time   10 days 
Relative humidity   95% 
Stage 1 ? 18
o
C for 2 days 
Stage 2 ? 13
o
C for 6 days 
Turning on days 2,4,6 
pH ?5.20 
Young Cheese 
*Type of mould used 
*Extent of acidication determined 
*Brine Salting 
pH ?6.50 
Renneting Fromase 
4% Fromase XLG solution 
100 mL 100 L
-1
 retentate 
Pure dried vaccuum salt 
1500 g 100 L
-1
 retentate 
*Fat content adjustment 
*Retentate-Salting 
40% Cream to F/P 1.59:1 
Brine salinity      ?95% 
H2O2 level          ?85 ppm 
Immersion time  10 min 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  42 
3.3. Analysis of nitrogen fractions 
Three types of nitrogen content were analysed in order to monitor the release of 
soluble nitrogen fractions during the maturation of the cheese. These include total 
nitrogen (TN), soluble nitrogen (SN) at pH 4.4, and non-protein nitrogen (NPN) in 
24% trichloroacetic acid (TCA). The method used for separating various nitrogen 
fractions in cheese has been described by Ard? (1999). The ratios of SN or NPN to 
TN serve as useful indicators for the extent of proteolysis in cheese. 
 
Fractionation of nitrogen contents 
Duplicate cheese from each treatment and one sample from each cheese were 
analysed at weeks one, three, five, and seven to determine the degree of proteolysis 
during maturation. A dispersion of the cheese was obtained by adding 80 mL warm 
0.5 M tri-sodium citrate solution (AR, LabServ Pronalys) to 17 ? 19 grams of grated 
cheese without rind and stirred for 60 minutes using a magnetic stirrer and hotplate at 
80 oC. The dispersion was cooled to room temperature (20 oC) and made up to 250 mL 
with distilled water. A 20 mL sample of the dispersion was used for TN analysis. The 
pH of the remaining cheese dispersion was adjusted to pH 4.4 using 1.0 M 
hydrochloric acid (AR, LabServ Pronalys). The dispersion was stirred as the pH was 
being adjusted, and the pH was constantly monitored by a pH meter (PB-20, 
Sartorious) with a cheese electrode (model L8880, Schott) attachment. Two hundred 
millilitres of the pH adjusted dispersion was made up to 250 mL using distilled water. 
Of the mixture, 80 mL was filtered using quantitative filter paper (model 
LBS0040.110, LabServ Filtration). A 25 mL sample of the filtrate was used for SN 
analysis. Then, 50 mL of the unfiltered pH 4.4 dispersion was made up to 100 mL 
using 24% (w/v) trichloroacetic acid solution (AR, Normapur). The solution was kept 
overnight at 4 oC and filtered using quantitative filter paper (model LBS0040.110, 
LabServ Filtration). A 25 mL sample of the filtrate was used for NPN analysis. 
 
Digestion 
The AOAC (2005) method 933.20 was used to determine the nitrogen content of TN, 
SN, and NPN. All samples to be digested were placed in Kjeldahl tubes and at least 
one nitrogen recovery test (0.20 g phenylalanine) was included per batch (20 tubes). 
Two Kjeltabs (Foss?) and 20 mL concentrated sulphuric acid (LabServ Pronalys, 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  43 
Reagent Grade) were added to each digestion tube. Digestion was carried on a 
digestion unit (model 2020, Tecator, Sweden) using a temperature scheme of 200 oC 
for 60 minutes, 270 oC for 45 minutes, 330 oC for 45 minutes, 405 oC until the samples 
turned clear and colourless. The samples were cooled after digestion. Subsequently 70 
mL of distilled water was added to each tube and they were further allowed to cool 
prior to distillation. 
 
Distillation 
The digestion tubes were placed in an automatic distillation unit (model 1026, Tecator 
Kjeltec System, Denmark). The settings of the automatic distillation included three 
injections of 20 mL aliquots of 40% (w/v) sodium hydroxide solution (LabServ, 
Reagent Grade) and a steam-on time of 3.2 minutes. Samples were distilled against 50 
mL of 4% (w/v) boric acid solution (LabServ, Reagent Grade) containing 1% 
bromocresol green (0.1% w/v dissolved in 95% ethanol) and 0.7% methyl red (0.1% 
w/v dissolved in 95% ethanol) in a receiver flask. The presence of nitrogen would 
induce a colour change from red to green in the boric acid during distillation. The 
nitrogen content was determined by titrating the samples against standardised 0.1 M 
hydrochloric acid solution to a grey-mauve endpoint. 
 
3.4. Measurement of pH 
Determination of pH was carried out on both the centre and outer portions with rind of 
the cheese samples. Duplicate cheeses from each treatment and duplicate samples 
from each cheese were tested every week for six weeks to monitor pH changes. A 
calibrated pH meter (PB-20, Sartorious) with cheese electrode (model L8880, Schott) 
attachment was used. Cheese (10 g) from the both centre and outer portions were 
added with 5 mL of distilled water and minced in small sample cups. The temperature 
of the samples was measured using a thermometer. The pH was measured at 20 ? 1 oC 
by inserting the pH electrode into the slurry with gentle mixing; two measurements 
were recorded for each sample. The pH electrode was rinsed with distilled water after 
each measurement. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  44 
3.5. Determination of total solids 
Total solids (TS) content was determined using the hot air oven method (AOAC 
International, 2005). Duplicate cheeses from each treatment and one sample from 
each cheese were tested every week for six weeks. A dry aluminium moisture dish 
was placed in a 105 oC air oven for 60 minutes to ensure all moisture was removed 
and then cooled to room temperature (20 oC) in a desiccator. The weight of the cooled 
dry moisture dish was recorded and 10 grams of prepared cheese sample without rind 
were spread uniformly in the dish. The weight of the sample was recorded. The dish 
was placed into 105 oC air oven for 8 hours and cooled to room temperature in a 
desiccator. The weight of the dish and its contents were recorded, then returned into 
the 105 oC air oven for 60 minutes and cooled to room temperature in a desiccator. 
The analysis was performed to within a standardising ? 1 mg. 
 
3.6. Determination of fat content 
The AOAC (2005) method was used in the determination of fat in cheese (933.05). 
Duplicate cheeses from each treatment and one sample from each cheese were tested 
in the first week after packing. Two to three grams of prepared cheese sample without 
rind were measured into Mojonnier tubes and 10 mL 25 % hydrochloric acid (w/v) 
was added. The tubes were placed in 60 oC water bath for 30 minutes with occasional 
swirling and cooled for five minutes. 
 
Fat Extraction 
Ten mL 95% ethanol (LabServ, Reagent Grade), 25 mL diethyl ether (LabServ 
Pronalys, Reagent Grade), and 25 mL petroleum ether (boiling range 40 oC ? 60 oC, 
LabServ Pronalys, Reagent Grade) were added to the samples successively and mixed 
thoroughly after adding each reagent. The samples were centrifuged for two minutes 
at 20 g and the organic solvent layer was decanted carefully into the pre-weighed 
aluminium dish. The solvents in aluminium dishes were evaporated on hot plates at 
mild temperatures below 40 oC. 
 
A second extraction was repeated using 5 mL ethanol, 15 mL diethyl ether, 15 mL 
petroleum ether. The samples were centrifuged for another 2 minutes at 20 g and the 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  45 
organic solvent layer was decanted into their respective aluminium dishes. After 
evaporating all solvents, the dry aluminium dishes were placed in the 105 oC air oven 
for 30 minutes. The dishes were allowed to cool to room temperature (20 oC) in a 
desiccator. The fat content was determined by weighing the fat in the dry dish. 
 
3.7. Determination of salt content 
The determination of the salt content is described in the AOAC (2005) methods 
935.43 and Kirk and Sawyer (1991). Duplicate cheeses from each treatment and 
duplicate samples from each treatment were tested in the first week after packing. 
Three grams of prepared cheese sample without rind were weighed into a 200 mL 
conical flask and 10 mL distilled water, 0.05 M silver nitrate solution (Ajax 
Chemicals, Analytical Reagent Grade, Univar), and 10 mL concentrated nitric acid 
(Ajax Chemicals, Analytical Reagent Grade, Univar, 70% w/w) were added. The 
cheese was dispersed by swirling and boiled gently until granular precipitation was 
present. To the contents of the flask, 50 mL distilled water were added and cooled to 
room temperature (20 oC). The excess silver nitrate was titrated with 0.05 M 
potassium thiocyanate solution (Sigma, ReagentPlus? Grade, 99.0%) using saturated 
ferric alum solution (LabServ, Reagent Grade) as an indicator to a reddish-brown 
endpoint. 
 
3.8. Instrumental analysis 
Instrumental texture analysis was performed using a TA-XT2 Texture Analyzer (SMS 
Stable Micro Systems Ltd, United Kingdom), operating in the compression mode. The 
test measurements were analysed using the Texture Analysis TE32 software. 
Duplicate cheeses from each treatment were analysed weekly for six weeks. Duplicate 
samples from each cheese were prepared by cutting out test pieces (15 mm in 
diameter) 10 mm from the centre of the cheese unit. The height of the test pieces were 
approximately 30 to 35 mm with the rind intact. Samples were cut at 4 oC and left at 
room temperature (20 oC) for at least 30 minutes before testing. The samples were 
analysed with compression up to 50% of initial height at a rate of 10 mm/min using an 
automatic trigger force of 5.0 gram. The force (N) was recorded for analysis. The 
measurements were carried out using a 50 mm diameter flat plate probe. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  46 
3.9. Cheese grading  
Grading system 
The cheese grading method used was developed exclusively by Puhoi Valley Cheese 
to periodically assess their cheese samples (Appendix 2). A list of undesirable 
attributes (e.g. brown edges, under ripe, too acidic etc.) was divided into four 
subgroups comprising appearance, cut, texture, and flavour (Table 3.2). The cheeses 
were given a total score of 40 points, and for each presences of undesired attribute, a 
point was deducted. Only cheeses with distinctively outstanding attributes were given 
extra points above 40 to the maximum of 50 points. Additional comments were also 
recorded on points being deducted or given, where these were extensively used during 
the grading of UF Camembert. 
 
Grading procedure 
Eight cheeses of each treatment type were graded by 3 ? 4 experienced cheese graders 
at Puhoi Valley Cheese, and a consensus of the grading result was recorded. Each 
cheese was first graded on the exterior appearance. The cheese was then cut across the 
centre using a knife and the appearance of the cross section (cut) was graded. The 
cheese was cut up into wedges and evaluated by the graders, and the texture and 
flavour of the cheese were then graded. 
 
Analysis of cheese grading results 
A selective list of attributes as well as commonly used comments was grouped into 
nine major sensory defects. The defects included rind discolouration (tanning or 
bolding of the surface mould flora), rind deformation (unevenness and misshaping of 
the rind), thick rind, core unevenness (a firm core texture which gradually gets softer 
towards the rind), excessive softness, saltiness, sourness, bitterness, blandness. The 
percentage of occurrence was calculated based on the defect detected (based on 
consensus and a yes/no answer) out of eight graded cheese in each treatment type (e.g. 
4 out of 8 low-fat cheese had thick rind = 50% occurrence). 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  47 
Table 3.2: A list of attributes used in cheese grading and the total scores given in 
each subgroup. 
 
Subgroups Attributes Standard Score 
Total 
Maximum Score 
Total 
Appearance 
 
 
Crooked or lopsided 
Oversized or 
undersized 
Bare mould & uneven 
Too moist or wet 
Uneven colour 
Brown edges 
8 
 
 
10 
 
 
Cut 
 
 
Dull 
Presence of holes 
Under ripe 
Over ripe 
Undesired rind 
4 
 
 
5 
 
 
Texture 
 
Too soft 
Too hard or dry 
Chalky 
12 
 
15 
 
Flavour 
 
 
 
Bitter 
Too acidic 
Ammonia 
Lacking flavour 
Chemical soapiness 
Too salty 
Metallic 
Rancid 
16 
 
 
 
20 
 
 
 
 
3.10. Consumer acceptance 
Panellists were recruited across Massey University Oteha Rohe campus which 
consisted of students, staff and guests. The panel sessions following the incomplete 
block design and the sixteen treatments of cheeses were divided into four separate 
sessions, with two extra sessions serving as duplicates. Approximately 30 panellists 
participated in each consumer acceptance test. 
 
Cheese samples at four weeks of maturation were prepared by cutting into pieces of 
eighths, coded with random numbers, and served at 18 ? 2 oC in sensory booths. Each 
panellist was presented with five cheese samples which comprised four treatments and 
a duplicate of one of the treatments, which tested for consistency within the repeats. 
Although it may not be ideal to present all samples simultaneously, this was done to 
manage the sensory sessions, thereby reducing possible errors.  
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  48 
The panellists were given questionnaires, where they were asked to taste each sample 
and rate, using a 9-point hedonic scale ranging from ?dislike extremely? to ?like 
extremely?, for the attributes of appearance, aroma, flavour, texture, and overall 
acceptance independently. The panellists were also asked to thoroughly rinse their 
mouth with the water provided between samples and to rate each sample without 
comparing between them. 
 
3.11. Microbiological analysis 
Microbiological analysis was carried out by Goodman Fielder (NZ) Ltd. Puhoi Valley 
Cheese plant. One sample of cheese from each treatment was tested for coliforms, E. 
coli, and Staphylococcus aureus counts on the first day of ripening. For the cheese to 
be considered safe for human consumption, the counts for each of the above species 
must be <1 cfu/g (FSANZ, 2009). 
 
3.12. Statistical analysis of data 
The data were analysed using the Minitab 15 (Minitab Ltd., Coventry, U.K.). One-
way Analysis of Variance (ANOVA) was used to determine significant differences 
between treatments with significant level set at 5%. Graphical presentations were 
produced using Microsoft? Office Excel 2003 (Microsoft Corporation, U.S.A.) with 
means and standard error of the mean (SEM), n values were calculated by number of 
treatments ? number of cheese in treatment ? number of replicates per cheese in 
analysis (e.g. n = 16 low fat treatments ? duplicate cheese per treatment ? duplicate 
pH measurements per cheese = 64). 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  49 
4. RESULTS 
Camembert cheese from the sixteen treatments, during storage at 4 ? 1 ?C, were 
monitored weekly for changes in TS, pH, and textural properties using the texture 
analyser (TA); fortnightly for changes in non-protein nitrogen (NPN) and soluble 
nitrogen (SN). The salt and fat content were analysed a week after packaging to assess 
the effects of the treatments. The organoleptic qualities of the cheese samples were 
also evaluated by consumer sensory panels, and experienced cheese tasters in cheese 
grading. The results are illustrated graphically in Figures 4.1 to 4.50. 
 
For quick references, treatments of the cheese samples were abbreviated in the 
following manner, explained by the underlined letter as: Low-fat/High-fat; Brine-
salted/Retentate-salted; pH 5.2/pH 4.9; Tube mould/Small mould. 
 
4.1. Change in total solids 
Table 4.1: p-values (p ? 0.05) for total solids within each type of treatment. 
 
  
 p-values for weeks 1- 6 
Cheese treatments 1 2 3 4 5 6 
Level of fat 0.791 0.994 0.740 0.954 0.039 0.008 
Type of salting 0.002 0.301 0.085 0.004 0.986 0.963 
Final acidification pH 0.000 0.000 0.000 0.000 0.000 0.000 
Mould type 0.226 0.191 0.772 0.451 0.933 0.959 
Note: The significant level was set at 5%. 
 
Low-fat versus high-fat 
The results obtained for the low-fat cheese samples were combined with the results 
for the high-fat cheese samples to ascertain if fat content had any impact on TS. In the 
first four weeks of their shelf life there were no significant differences in the TS 
content between low-fat and high-fat cheeses (Figure 4.1). From week five onwards 
the high-fat cheese samples had significantly higher mean TS content than that of the 
low-fat treatments. 
 
Brine-salted versus retentate-salted 
The results obtained for the brine-salted cheese samples were combined with the 
results for the retentate-salted cheese samples to ascertain if the salting method had 
any significant impact on TS. The mean TS in brine-salted cheeses were significantly 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  50 
higher than retentate-salted cheeses at weeks 1 and 4, but at weeks 2, 3, 5 and 6 there 
was no significant differences between the TS in all of the cheese samples (Figure 
4.2). The effects of the two salting methods on cheese TS was only significant during 
early stages of maturation. 
 
pH 5.2 versus pH 4.9 
The results obtained for the cheese samples acidified to pH 5.2 were combined with 
the results for the cheese samples acidified to pH 4.9 to ascertain if the final 
acidification pH had any impact on TS. Cheese produced with higher final 
acidification pH of 5.2 had significantly higher mean TS than the lower acidification 
final pH of 4.9 cheeses during all six weeks (Figure 4.3). The cheese samples made 
with a higher final acidification pH exhibited higher moisture loss. 
 
Tube mould versus small mould 
The results obtained for the two mould types were combined for each mould type to 
determine if the mould type had any impact on TS. The TS in the cheese samples 
from the two mould types were not significantly different from each other (Figure 
4.4). 
 
The mean TS of each cheese type during the maturation period (six weeks) are shown 
in Figures 4.5 to 4.8. Generally, cheese samples acidified to pH 5.2 (LB5.2T, LB5.2S, 
LR5.2T, LR5.2S, HB5.2T, HB5.2S, HR5.2T, HR5.2S) had higher mean TS than 
cheese acidified to pH 4.9 (LB4.9T, LB4.9S, LR4.9T, LR4.9S, HB4.9T, HB4.9S, 
HR4.9T, HR4.9S). When comparing the types of mould used, the cheeses made using 
the tube moulds had higher mean TS than those made using the small moulds, but 
only in those acidified to pH 5.2 and not those at pH 4.9. This is not shown in the 
combined mean TS of mould types shown in Figure 4.4. 
 
In general, the two final acidification pH levels (pH 5.2 and 4.9) produced cheese with 
significantly different mean TS (Figure 4.2). Samples of cheese at pH 5.2 had higher 
mean TS during the storage period than those produced at pH 4.9. Treatments with 
different levels of fat (Figure 4.1) and salting methods (Figure 4.2) had significant 
differences in the mean TS but only in the early stages of maturation post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  51 
 
Figure 4.1: Changes in total solids (TS) (mean ? SEM) of low-fat (n = 32) and 
high-fat (n = 16) Camembert cheese stored at 4 ? 1 ?C for six weeks post-
packaging. 
 
 
Figure 4.2: Changes in total solids (TS) (mean ? SEM) of brine-salted (n = 24) 
and retentate-salted (n = 16) Camembert cheese stored at 4 ? 1 ?C for six weeks 
post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  52 
 
Figure 4.3: Changes in total solids (TS) (mean ? SEM) of Camembert cheese 
acidified to pH 5.2 (n = 24) and pH 4.9 (n = 24). The samples were stored at 4 ? 1 
?C for six weeks post-packaging. 
 
 
Figure 4.4: Changes in total solids (TS) (mean ? SEM) of Camembert cheese 
made using tube moulds (n = 24) and small moulds (n = 24). The samples were 
stored at 4 ? 1 ?C for six weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  53 
 
Figure 4.5: Changes in total solids (TS) (mean ? SEM) of Camembert cheese for 
low-fat brine-salted samples (n = 4 in each treatment). The samples were stored 
at 4 ? 1 ?C for six weeks post-packaging. 
 
 
Figure 4.6: Changes in total solids (TS) (mean ? SEM) of Camembert cheese for 
low-fat retentate-salted samples (n = 4 in each treatment). The samples were 
stored at 4 ? 1 ?C for six weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  54 
 
Figure 4.7: Changes in total solids (TS) (mean ? SEM) of Camembert cheese for 
high-fat brine-salted samples (n = 4 in each treatment). The samples were stored 
at 4 ? 1 ?C for six weeks post-packaging. 
 
 
Figure 4.8: Changes in total solids (TS) (mean ? SEM) of Camembert cheese for 
high-fat retentate-salted samples (n = 4 in each treatment). The samples were 
stored at 4 ? 1 ?C for six weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  55 
4.2. Salt content 
The mean salt content for all cheese samples, consisting of the four variables (level of 
fat, salting method, final acidification pH, and mould type) and their two levels are 
shown in Figure 4.9. The low-fat cheese samples had significantly higher mean salt 
content (1.74 ? 0.02 % w/w) than the high-fat samples (1.52 ? 0.04% 
w/w). This was 
shown in Figure 4.10, where five of the high-fat cheese samples (HB4.9S, HR4.9T, 
HB4.9T, HR5.2T, and HR4.9S) had noticeably lower salt content when compared to 
the rest of the cheese samples. The mean salt contents of the cheese samples made 
using brine-salting and retentate-salting were not significantly different from each 
other, as shown in Figure 4.9. In the cheese samples ripened to different final 
acidification pH, the pH 5.2 samples (1.74 ? 0.03 % w/w) had higher mean salt content 
than the pH 4.9 samples (1.52 ? 0.04 % w/w). This was observed in Figure 4.10 where 
almost all the pH 5.2 samples had higher salt content than the pH 4.9 samples made 
using the same salting method. Finally, cheese samples made using the tube mould 
were not significantly different to samples made using the small mould (Figure 4.9). 
 
 
Figure 4.9: Combined mean of salt content (mean ? SEM) in Camembert cheese 
samples for each treatment: high-fat (n = 32) and low-fat (n = 32); brine-salted (n 
= 32) and retentate-salted (n = 32); final acidification pH of 5.2 (n = 32) and pH 
4.9 (n = 32); and tube mould (n = 32) and small mould (n = 32). Samples were 
analysed in the first week post-packaging. 
 
Level of Fat  Salting method Final acidification pH Mould type 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  56 
1.0
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
1.9
2.0
L
B
5
.
2
T
L
B
4
.
9
S
L
R
5
.
2
S
L
R
4
.
9
T
L
B
5
.
2
S
L
B
4
.
9
T
L
R
5
.
2
T
L
R
4
.
9
S
H
B
5
.
2
T
H
B
4
.
9
S
H
R
5
.
2
S
H
R
4
.
9
T
H
B
5
.
2
S
H
B
4
.
9
T
H
R
5
.
2
T
H
R
4
.
9
S
M
e
a
n
 
S
a
l
t
 
C
o
n
t
e
n
t
 
(
%
)
 
Figure 4.10: Salt content (mean ? SEM) in Camembert cheese samples of sixteen 
treatments (n = 4 in each treatment) analysed in the first week post-packaging. 
 
Table 4.2: p-values (p ? 0.05) for salt and fat levels within each type of treatment. 
 
Cheese treatments Salt content Fat content 
Level of fat 0.000 0.000 
Type of salting 0.455 0.689 
Final acidification pH 0.000 0.575 
Mould type 0.311 0.762 
Note: The significant level was set at 5%. 
 
4.3. Fat content 
The combined mean fat content of each variable and their two levels are shown in 
Figure 4.11. The fat content of low-fat cheese samples (24.96 ? 0.23 % w/w) was 
significantly lower than in the high-fat samples (28.03 ? 0.14 % w/w). This is also 
observed in Figure 4.12 where the eight low-fat samples had significantly lower fat 
contents than the eight high-fat samples. Note that LB5.2S had a slightly higher fat 
content (26.57 ? 0.14 % w/w) than the other low-fat cheeses, but the fat content was 
still significantly lower than that of high-fat samples (above 27 %). This may be 
attributed to error in blending of the cheese samples. 
 
No significant differences were observed between the levels of the three other 
variables: salting method, final acidification pH, and mould type. 
Treatments 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  57 
 
 
Figure 4.11: Combined mean of fat content (mean ? SEM) in Camembert cheese 
samples for each treatment: low-fat (n = 16) and high-fat (n = 16); brine-salted (n 
= 16) and retentate-salted (n = 16); final acidification pH of 5.2 (n = 16) and pH 
4.9 (n = 16); and tube mould (n = 16) and small mould (n = 16). Samples were 
analysed in the first week post-packaging. 
 
23
24
25
26
27
28
29
L
B
5
.
2
T
L
B
4
.
9
S
L
R
5
.
2
S
L
R
4
.
9
T
L
B
5
.
2
S
L
B
4
.
9
T
L
R
5
.
2
T
L
R
4
.
9
S
H
B
5
.
2
T
H
B
4
.
9
S
H
R
5
.
2
S
H
R
4
.
9
T
H
B
5
.
2
S
H
B
4
.
9
T
H
R
5
.
2
T
H
R
4
.
9
S
M
e
a
n
 
F
a
t
 
C
o
n
t
e
n
t
 
(
%
)
 
Figure 4.12: Fat content (mean ? SEM) in Camembert cheese samples of sixteen 
treatments (n = 4 in each treatment) analysed in the first week post-packaging. 
 
Level of fat Salting method Final acidification pH Mould type 
Treatments 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  58 
4.4. Changes in pH levels 
The pH levels of the cheese samples were determined at the core (centre) and on the 
rind surface. The mean core pH levels were significantly higher in high-fat samples at 
weeks 5 and 6 (Figures 4.13), as was in brine-salted samples at weeks 4 and 6 (Figure 
4.14), in samples acidified to pH 5.2 at weeks 1, 3 and 5 (Figure 4.15), and in small 
mould samples at weeks 2 and 5 (Figure 4.16). The rind pH exhibited similar trends as 
the core pH but only in acidification pH and mould treatments (Figures 4.13 to 4.16). 
The core pH levels of all treatments had similar trends including the rate of increase 
(mean = 0.4744, SEM = 0.1677) starting at the mean pH of 5.28 ? 0.007 up to pH 7.57 
? 0.011 at week 6. Also, the rind pH levels of all treatments had the same trends with 
similar rates of increase (mean = 0.2497 SEM = 0.0883) starting at mean pH of 6.62 ? 
0.020 up to pH 7.89 ? 0.004) at week 6. 
 
Figures 4.17 to 4.20 show the changes in core pH levels of the 16 cheese samples for 
each different treatment. Low-fat brine-salted samples acidified to a higher final pH, 
(LB5.2T and LB5.2S) generally had higher core pH than those acidified to the lower 
final pH cheeses (LB4.9T and LB4.9S) (Figure 4.17), this was observated when the 
data was combined with other high-fat or retentate-salted samples to generate mean 
values in Figure 4.15, suggesting that pH adjustment was somewhat effective. For 
low-fat retentate-salted samples (Figure 4.18), no conclusive trends were observed in 
core pH levels between samples acidified to pH 5.2 or 4.9 or in the different mould 
types. However, from week three to week six, LR5.2T and LR4.9S samples had 
higher mean core pH values than LR5.2S and LR4.9T. With respect to the rind pH, no 
significant trends were observed between the low-fat samples (Figures 4.17 and 4.18). 
 
The changes in mean core pH of high-fat brine-salted cheese samples are shown in 
Figure 4.19. At weeks one and two, the samples acidified to pH 5.2 (HB5.2T and 
HB5.2S) had higher mean core pH levels than those acidified to pH 4.9 (HB4.9T and 
HB4.9S). This suggested that a relatively higher pH was maintained after one week of 
on-site ripening. At weeks three and four, HB5.2S had higher mean core pH levels 
than HB4.9T, followed by HB4.9S and then HB5.2T. Finally, at week five and six, 
HB4.9S samples had a noticeably significant increase in core pH. This particular 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  59 
cheese resulted with the highest pH levels compared to the four high-fat samples after 
six weeks.  
 
The changes in core mean pH levels of high-fat retentate-salted samples are shown in 
Figure 4.20. At week one, samples acidified to pH 5.2 (HB5.2T and HB5.2S) had 
higher core mean pH levels than those acidified to pH 4.9 (HB4.9T and HB4.9S). No 
obvious trends could be observed from the four samples, but at week six, cheese 
samples with lower final acidification pH (HB4.9T and HB4.9S) also had higher 
mean core pH levels than those with higher final acidification pH levels (HB5.2T and 
HB5.2S). These observations are however lost, when the data are combined with other 
low-fat and brine-salted samples as shown in Figure 4.15. In terms of the rind pH 
levels, no significant trends were found between the high-fat samples (Figures 4.19 
and 4.20). This suggests that the growth of surface-mould flora altered the rind pH of 
the cheese in a consistent manner (Fox et al., 2000), regardless of the four variables 
implemented (level of fat, salting method, final acidification pH, and mould type). 
 
Table 4.3: p-values (p ? 0.05) for core pH within each type of treatment. 
 
  p-values for weeks 1-6 
Cheese treatments 1 2 3 4 5 6 
Level of fat 0.860 0.828 0.323 0.923 0.019 0.118 
Type of salting 0.632 0.271 0.849 0.005 0.170 0.004 
Final acidification pH 0.005 0.470 0.001 0.044 0.018 0.073 
Mould type 0.056 0.002 0.369 0.435 0.003 0.706 
Note: The significant level was set at 5%. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  60 
 
Figure 4.13: Changes in the core and rind pH (mean ? SEM) of low-fat (n = 64) 
and high-fat (n = 32) Camembert cheese stored at 4 ? 1 ?C for six weeks post-
packaging. 
 
 
Figure 4.14: Changes in the core and rind pH (mean ? SEM) of brine-salted (n = 
48) and retentate-salted (n = 48) Camembert cheese stored at 4 ? 1 ?C for six 
weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  61 
 
Figure 4.15: Changes in the core and rind pH (mean ? SEM) of Camembert 
cheese acidified to pH 5.2 (n = 48) and pH 4.9 (n = 48). The samples were stored 
at 4 ? 1 ?C for six weeks post-packaging. 
 
 
Figure 4.16: Changes in the core and rind pH (mean ? SEM) of Camembert 
cheese made using tube moulds (n = 48) and small moulds (n = 48). The samples 
were stored at 4 ? 1 ?C for six weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  62 
 
Figure 4.17: Changes in the core and rind pH (mean ? SEM) of Camembert 
cheese for low-fat brine-salted samples (n = 8 in each treatment). The samples 
were stored at 4 ? 1 ?C for six weeks post-packaging. 
 
 
Figure 4.18: Changes in the core and rind pH (mean ? SEM) of Camembert 
cheese for low-fat retentate-salted samples (n = 8 in each treatment). The 
samples were stored at 4 ? 1 ?C for six weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  63 
 
Figure 4.19: Changes in the core and rind pH (mean ? SEM) of Camembert 
cheese for high-fat brine-salted samples (n = 8 in each treatment). The samples 
were stored at 4 ? 1 ?C for six weeks post-packaging. 
 
 
Figure 4.20: Changes in the core and rind pH (mean ? SEM) of Camembert 
cheese for high-fat retentate-salted samples (n = 8 in each treatment). The 
samples were stored at 4 ? 1 ?C for six weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  64 
4.5. Level of proteolysis 
4.5.1. Non-protein nitrogen (NPN) 
Total nitrogen (TN) values and non-protein nitrogen (NPN) fractions of the cheese 
samples were determined by the Kjeldahl method. The portion of NPN in TN serves 
as an indicator for the level of proteolysis at different stages during the maturation of 
the cheese. 
 
Table 4.4: p-values (p ? 0.05) for NPN/TN ratios within each type of treatment. 
 
  p-values for selected periods (weeks) 
Cheese treatments 1 3 5 7 
Level of fat 0.030 0.000 0.502 0.000 
Type of salting 0.003 0.000 0.000 0.000 
Final acidification pH 0.046 0.041 0.034 0.009 
Mould type 0.138 0.372 0.193 0.682 
Note: The significant level was set at 5%. 
 
High-fat versus low-fat 
In Figure 4.21, the mean NPN/TN ratios for high-fat cheese samples were 
significantly higher than those of low-fat samples at weeks one, three and seven, 
suggesting that in general, high-fat samples had a higher level of proteolysis than low-
fat samples. 
 
Brine-salted versus retentate-salted 
In Figure 4.22, the mean NPN/TN ratios for brine-salted samples were significantly 
higher than those of retentate-salted samples for all four measurements at weeks one 
to seven, suggesting a consistently higher level of proteolysis in brine-salted cheese.  
 
pH 5.2 versus pH 4.9 
NPN/TN ratios for cheese samples produced with the two final acidification pH levels 
are shown in Figure 4.23. Samples acidified to pH 4.9 had significantly higher 
NPN/TN ratios than samples acidified to pH 5.2 from week one to seven. This 
suggested that the levels of proteolysis were higher in cheese acidified to pH 4.9 than 
those acidified to pH 5.2 in a consistent manner for seven weeks during maturation. 
 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  65 
Tube mould versus small mould 
The differences of the mean NPN/TN ratios in cheese samples made using tube 
moulds and small moulds were not significant (Figure 4.24). 
 
The mean NPN/TN ratios for each of the sixteen treatments are presented in Figures 
4.25 to 4.28. When comparing low-fat (Figures 4.25 and 4.26) with high-fat cheese 
samples (Figures 4.27 and 4.28), the high-fat samples generally had higher mean 
NPN/TN ratios than the low-fat samples with the exception of LB4.9S in Figure 4.25. 
The mean NPN/TN ratios of the low-fat retentate-salted treatments shown in Figure 
4.26 were generally lower than other treatments shown in Figures 4.25, 4.27 and 4.28, 
which probably gave rise to lower overall means of NPN/TN ratio for retentate-salted 
cheeses.  
 
In the case of mean NPN/TN ratios for the high-fat cheese samples acidified to the 
two final acidification pH levels (pH 5.2 and pH 4.9), results in Figure 4.27 show that 
the pH 4.9 samples (HB4.9T and HB4.9S) were clearly above the pH 5.2 samples 
(HB5.2T and HB5.2S) for all seven weeks. Low-fat cheese samples, LB4.9S (Figure 
4.25) and LR4.9S (Figure 4.26) had slightly higher mean NPN/TN ratios than other 
treatments in Figures 4.25 and 4.26, respectively. Regarding the two types of moulds 
used in terms of mean NPN/TN ratios, no apparent trends were observed. Therefore, 
there were no significant differences between the two types of moulds used with 
respect to the NPN/TN ratios in the cheese samples (Figure 4.24). 
 
 
 
 
 
 
 
 
 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  66 
 
Figure 4.21: Changes in the NPN/TN ratios (mean ? SEM) of low-fat (n = 32) and 
high-fat (n = 16) Camembert cheese stored at 4 ? 1 ?C for seven weeks post-
packaging. 
 
 
Figure 4.22: Changes in the NPN/TN ratio (mean ? SEM) of brine-salted (n = 24) 
and retentate-salted (n = 24) Camembert cheese stored at 4 ? 1 ?C for seven 
weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  67 
 
Figure 4.23: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese 
acidified to pH 5.2 (n = 24) and pH 4.9 (n = 24). The samples were stored at 4 ? 1 
?C for seven weeks post-packaging. 
 
 
Figure 4.24: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese 
made using tube moulds (n = 24) and small moulds (n = 24). The samples were 
stored at 4 ? 1 ?C for seven weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  68 
 
Figure 4.25: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese 
for low-fat brine-salted samples (n = 4 in each treatment). The samples were 
stored at 4 ? 1 ?C for seven weeks post-packaging. 
 
 
Figure 4.26: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese 
for low-fat retentate-salted samples (n = 4 in each treatment). The samples were 
stored at 4 ? 1 ?C for seven weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  69 
 
Figure 4.27: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese 
for high-fat brine-salted samples (n = 4 in each treatment). The samples were 
stored at 4 ? 1 ?C for seven weeks post-packaging. 
 
 
Figure 4.28: Changes in the NPN/TN ratio (mean ? SEM) of Camembert cheese 
for high-fat retentate-salted samples (n = 4 in each treatment). The samples were 
stored at 4 ? 1 ?C for seven weeks post-packaging. 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  70 
4.5.2. pH 4.4 soluble nitrogen (SN) 
The mean soluble nitrogen/total nitrogen SN/TN ratios were determined by the 
Kjeldahl method, which also indicate the level of proteolysis at different stages of 
cheese ripening. The SN includes protein fractions which are soluble in pH 4.4 as well 
as NPN. 
 
Table 4.5: p-values (p ? 0.05) for SN/TN ratios within each type of treatment. 
 
   p-values for selected periods (weeks)  
Cheese treatments 1 3 5 7 
Level of fat 0.709 0.176 0.351 0.000 
Type of salting 0.639 0.916 0.001 0.912 
Final acidification pH 0.529 0.097 0.050 0.000 
Mould type 0.329 0.439 0.450 0.695 
Note: The significant level was set at 5%. 
 
The mean SN/TN ratios of high-fat cheese samples were significantly higher than 
those of the low-fat samples at week seven of maturation only as shown in Figure 
4.29. There were also no significant differences between brine-salted cheeses and 
retentate-salted cheeses in terms of the SN/TN ratio, except at week five, where the 
brine-salted samples had a higher mean SN/TN ratio than the retentate-salted samples 
(Figure 4.30). 
 
Cheese samples acidified to pH 4.9 had significantly higher mean SN/TN ratios than 
samples acidified to pH 5.2 only significant at week 7 of storage (Figure 4.31). The 
mean SN/TN ratios of cheese samples made using tube moulds and small moulds are 
shown in Figure 4.32, where both treatments were not significantly different from 
each other. 
 
In comparing low-fat (Figures 4.33 and 4.34) and high-fat (Figure 4.35 and 4.36) 
treatments, generally higher mean SN/TN ratio was observed in high-fat treatments at 
week seven. No trends were observed between brine-salted (Figures 4.33 and figure 
4.35) and retentate-salted treatments (Figures 4.34 and 4.36). In the case of cheese 
samples made at two final acidification pH levels, no significant observations could 
be drawn for the first five weeks of maturation for the SN/TN ratios (Figures 4.33 to 
4.36). However, at week seven, the majority of the samples acidified to pH 4.9 had 
significantly higher mean SN/TN ratios than their pH 5.2 equivalent treatments with 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  71 
other variables remaining constant. The exception was HB4.9T, the SN/TN ratio for 
this treatment was not significantly higher than its equivalent HB5.2T (Figure 4.35). 
No apparent trends could be observed in tube mould and small mould treatments 
shown in Figures 4.33 to 4.36, suggesting the use of different moulds had no apparent 
impact in cheese proteolysis. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  72 
 
Figure 4.29: Changes in the SN/TN ratio (mean ? SEM) of low-fat (n = 32) and 
high-fat (n = 16) Camembert cheese stored at 4 ? 1 ?C for seven weeks post-
packaging. 
 
  
Figure 4.30: Changes in the SN/TN ratio (mean ? SEM) of brine-salted (n = 24) 
and retentate-salted (n = 24) Camembert cheese stored at 4 ? 1 ?C for seven 
weeks post-packaging. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  73 
 
Figure 4.31: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese 
acidified pH 5.2 (n = 24) and pH 4.9 (n = 24). The samples were stored at 4 ? 1 ?C 
for seven weeks post-packaging. 
 
 
Figure 4.32: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese 
made using tube moulds (n = 24) and small moulds (n = 24). The samples were 
stored at 4 ? 1 ?C for seven weeks post-packaging. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  74 
 
Figure 4.33: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese for 
low-fat brine-salted samples (n = 4 in each treatment). The samples were stored 
at 4 ? 1 ?C for seven weeks post-packaging. 
 
 
Figure 4.34: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese for 
low-fat retentate-salted samples (n = 4 in each treatment). The samples were 
stored at 4 ? 1 ?C for seven weeks post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  75 
 
 
Figure 4.35: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese for 
high-fat brine-salted samples (n = 4 in each treatment). The samples were stored 
at 4 ? 1 ?C for seven weeks post-packaging. 
 
 
Figure 4.36: Changes in the SN/TN ratio (mean ? SEM) of Camembert cheese for 
high-fat retentate-salted samples (n = 4 in each treatment). The samples were 
stored at 4 ? 1 ?C for seven weeks post-packaging. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  76 
4.6. Texture analysis 
Instrumental texture analysis was used to measure the force required to compress a 
cheese sample to 50% of its original height. A higher deformation force indicated the 
cheese sample was relatively firmer and a lower deformation force indicated the 
sample was softer. Figures 4.37 to 4.40 show the deformation force for the sixteen 
treatments analysed. In Figure 4.37, samples LB5.2T and LB5.2S were significantly 
firmer (higher deformation forces) than LB4.9T and LB4.9S at weeks one and two. At 
weeks five and six, LB5.2T was significantly firmer than the other three samples. 
However, this does not suggest that samples acidified to pH 5.2 were necessarily 
firmer than those acidified to pH 4.9 as LB5.2S had similar low deformation forces 
similar to pH 4.9 samples at week three to six. In Figure 4.38, no meaningful trends 
can be observed, with the exception of LR5.2T, which was significantly firmer at 
week two to four than other samples. In comparing the brine-salted samples in Figure 
4.37 and retentate-salted samples in Figure 4.38, the two salt treatments did not result 
in any significant differences in the case of texture analysis measurements. 
 
Texture analysis results for the high-fat cheeses are shown in Figure 4.39 and Figure 
4.40. In Figure 4.39, samples acidified to pH 5.2, HB5.2T and HB5.2S were 
significantly firmer than HB4.9T and HB4.9S at weeks one and two. At week four to 
six, HB5.2T was significantly firmer than the rest of the samples, which was similar 
to LB5.2T in Figure 4.37. This suggests that the combination of high pH and tube 
mould treatment, regardless of fat content, contribute to a firmer cheese during its 
maturation period. In Figure 4.40, HR4.9T was significantly firmer than the samples 
acidified to pH 5.2 at week one and also in weeks five and six. This is apparent as the 
samples acidified to pH 4.9 were softer (lower deformation forces) than those 
acidified to pH 5.2 in the first two weeks of the brine-salted samples (Figures 4.37 and 
4.39), suggesting the influence of salt in cheese firmness. Sample HR5.2T was firmer 
than the other samples at week two to four (Figure 4.40). This was an interesting 
observation as pH 5.2-tube mould samples were always firmer at week two in all 
texture analysis results. Sample HR4.9S was generally softer than all samples and was 
consistent in all texture analysis measurements for being softer than other samples in 
the first two weeks. No trends could be observed when comparing between samples of 
the two salt treatments for high-fat cheese samples in Figures 4.39 and 4.40. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  77 
 
Figure 4.37: Change in force (N) (mean ? SEM) at 50% deformation for low-fat 
brine-salted cheese samples (n = 8 in each treatment) stored at 4 ? 1 ?C for six 
weeks post-packaging. 
 
 
Figure 4.38: Change in force (N) (mean ? SEM) at 50% deformation for low-fat 
retentate-salted cheese samples (n = 8 in each treatment) stored at 4 ? 1 ?C for six 
weeks post-packaging. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  78 
 
Figure 4.39: Change in force (N) (mean ? SEM) at 50% deformation for high-fat 
brine-salted cheese samples (n = 8 in each treatment) stored at 4 ? 1 ?C for six 
weeks post-packaging. 
 
 
Figure 4.40: Change in force (N) (mean ? SEM) at 50% deformation for high-fat 
retentate-salted cheese samples (n = 8 in each treatment) stored at 4 ? 1 ?C for six 
weeks post-packaging. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  79 
4.7. Cheese grading 
The week four cheese grading data are presented in Figures 4.41 to 4.44. Nine major 
sensory defects were plotted on a graph against the percentage of occurrence when 
each sample from the 16 treatments was graded. The defects included rind 
discolouration (tanning or bolding of the surface mould flora), rind deformation 
(unevenness and misshaping of the rind), thick rind, core unevenness (a firm core 
texture which gradually gets softer towards the rind), excessive softness, saltiness, 
sourness, bitterness, and blandness. The percentage of occurrence indicated the 
frequency of detecting certain detects within the same group of cheese samples at the 
age of four and six weeks as determined by cheese judges. 
 
Low-fat versus high-fat 
The occurrence of rind discolouration was more prominent among low-fat cheese 
samples with 100% occurrence compared to 50% in the high-fat samples (Figure 
4.41). Twenty-five percent of the low-fat samples graded were affected by core 
unevenness whereas the high-fat samples did not have this defect. Twenty-five 
percent of the low-fat samples and 12.5% of the high-fat samples were judged to have 
excessive sourness. Both treatments were often affected by thick rind, with 87.5% and 
100% occurrence in the high-fat and low-fat samples, respectively. The majority of 
cheese samples from both treatments were also affected by excessive softness, with 
87.5% and 75% occurrence in the high-fat and low-fat samples, respectively. Rind 
deformation at week four was apparent in the high-fat samples, occurring in 37.5% 
compared to 12.5% in the low-fat samples. The high-fat samples also had excessive 
saltiness and bitterness with 75% and 50% occurrence, respectively, whereas in the 
low-fat samples their occurrences were only 50% and 25%, respectively. Forty 
percent of both the high-fat and low-fat samples had excessive blandness. 
 
Brine-salted versus retentate-salted 
The retentate-salted cheese samples had higher occurrences in rind discolouration, 
rind deformation, and excessive saltiness than the brine-salted samples (Figure 4.42). 
The occurrence of rind discolouration in the retentate-salted samples was 87.5% 
compared to 62.5% in the brine-salted samples. Thirty-eight percent of the retentate-
salted cheese samples exhibited rind deformation, whereas only 12.5% of the brine-
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  80 
salted samples were affected. The majority of the retentate-salted samples had 
excessive saltiness with 87.5% occurrence and only 37.5% occurrence in the brine-
salted samples. In contrast, defects such as thick rind, excessive softness, sourness, 
bitterness and blandness had higher occurrence in the brine-salted samples compared 
to the retentate-salted samples. All the graded brine-salted samples (100%) were 
found to have thick rinds and 87.5% had excessive softness, whereas only 87.5% of 
the retentate-salted samples had thick rind and 75% had excessive softness. In terms 
of flavour defects, 25% of the brine-salted samples had excessive sourness, 62.5% had 
excessive bitterness, and 50% had excessive blandness, compared to 12.5%, 12.5% 
and 25% of the retentate-salted samples, respectively. 
 
pH 5.2 versus pH 4.9 
A comparison of grading defects in cheese samples acidified to pH 5.2 or pH 4.9 
showed that those acidified to pH 4.9 were more often affected by rind discolouration, 
thick rind, excessive softness, sourness, and bitterness (Figure 4.43). Rind 
discolouration occurrence in the pH 4.9 samples was 87.5%, while it was only 62.5% 
in the pH 5.2 samples. Both pH treatments of the cheese samples had high 
occurrences of the thick rind defect, with 100% in the pH 4.9 samples and 87.5% in 
the pH 5.2 samples. All the pH 4.9 samples (100%) had excessive softness during 
grading, whereas only 62.5% of the pH 5.2 samples had excessive softness after four 
weeks maturation. In terms of flavour, 37.5% of the pH 4.9 samples had excessive 
sourness and 50% had excessive bitterness, compared to no occurrence in excessive 
sourness and only 25% had excessive bitterness in the pH 5.2 samples. However, 75% 
of the pH 5.2 samples had excessive saltiness compared to 50% of the pH 4.9 
samples. A small portion of both pH 5.2 and pH 4.9 samples were affected by rind 
deformation (25%) and core unevenness (12.5%). Thirty-eight percent of samples 
from both treatments were considered to be too bland (Figure 4.49). 
 
Tube mould versus small mould 
Rind discolouration in cheese samples made using tube moulds occurred more often 
than those made using small moulds, with 87.5% and 62.5% occurrence, respectively 
(Figure 4.44). All the graded tube mould samples (100%) were affected with thick 
rind, as a high number of the small mould samples also suffered from the same defect 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  81 
at 87.5% occurrence. In contrast, the small mould samples had higher occurrence in 
defects including rind deformation, core unevenness, excessive softness, sourness, 
bitterness and blandness. In the samples made using small moulds, 50% were affected 
by rind deformation and 25% had core unevenness, whereas there were no 
occurrences of these two defects in those made using tube moulds. All of the small 
mould samples (100%) graded had excessive softness at four weeks post-packaging, 
compared to 62.5% of the tube mould samples. For flavour defects, 62.5 % of the 
samples from both treatments were described to have excessive saltiness, and the 
small mould samples were more often perceived to have excessive sourness at 25%, 
excessive bitterness at 50%, and excessive blandness at 62.5% occurrence compared 
to 12.5%, 25%, and 12.5% in the tube mould samples, respectively. 
 
 
 
 
 
 
 
 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  82 
0
10
20
30
40
50
60
70
80
90
100
R
i
n
d
 
d
i
s
c
o
l
o
u
r
a
t
i
o
n
R
i
n
d
 
d
e
f
o
r
m
a
t
i
o
n
T
h
i
c
k
 
r
i
n
d
C
o
r
e
 
u
n
e
v
e
n
n
e
s
s
S
o
f
t
n
e
s
s
S
a
l
t
i
n
e
s
s
S
o
u
r
n
e
s
s
B
i
t
t
e
r
n
e
s
s
B
l
a
n
d
n
e
s
s
Cheese defects (week four post-packaging)
O
c
c
u
r
r
e
n
c
e
 
(
%
)
High-Fat
Low-Fat
 
Figure 4.41: Percentage of occurrence in nine major sensory defects for high-fat 
(n = 8) and low-fat (n = 8) Camembert cheese at the age of four weeks. The 
samples were stored at 4 ? 1 ?C post-packaging. 
 
0
10
20
30
40
50
60
70
80
90
100
R
i
n
d
 
d
i
s
c
o
l
o
u
r
a
t
i
o
n
R
i
n
d
 
d
e
f
o
r
m
a
t
i
o
n
T
h
i
c
k
 
r
i
n
d
C
o
r
e
 
u
n
e
v
e
n
n
e
s
s
S
o
f
t
n
e
s
s
S
a
l
t
i
n
e
s
s
S
o
u
r
n
e
s
s
B
i
t
t
e
r
n
e
s
s
B
l
a
n
d
n
e
s
s
Cheese defects (week four post-packaging)
O
c
c
u
r
r
e
n
c
e
 
(
%
)
Brine-Salted
Retentate-Salted
 
Figure 4.42: Percentage of occurrence in nine major sensory defects for brine-
salted (n = 8) and retentate-salted (n = 8) Camembert cheese at the age of four 
weeks. The samples were stored at 4 ? 1 ?C post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  83 
0
10
20
30
40
50
60
70
80
90
100
R
i
n
d
 
d
i
s
c
o
l
o
u
r
a
t
i
o
n
R
i
n
d
 
d
e
f
o
r
m
a
t
i
o
n
T
h
i
c
k
 
r
i
n
d
C
o
r
e
 
u
n
e
v
e
n
n
e
s
s
S
o
f
t
n
e
s
s
S
a
l
t
i
n
e
s
s
S
o
u
r
n
e
s
s
B
i
t
t
e
r
n
e
s
s
B
l
a
n
d
n
e
s
s
Cheese defects (week four post-packaging)
O
c
c
u
r
r
e
n
c
e
 
(
%
)
pH 5.2
pH 4.9
 
Figure 4.43: Percentage of occurrence in nine major sensory defects for 
Camembert cheese acidified to pH 5.2 (n = 8) and pH 4.9 (n = 8) at the age of 
four weeks. The samples were stored at 4 ? 1 ?C post-packaging. 
 
0
10
20
30
40
50
60
70
80
90
100
R
i
n
d
 
d
i
s
c
o
l
o
u
r
a
t
i
o
n
R
i
n
d
 
d
e
f
o
r
m
a
t
i
o
n
T
h
i
c
k
 
r
i
n
d
C
o
r
e
 
u
n
e
v
e
n
n
e
s
s
S
o
f
t
n
e
s
s
S
a
l
t
i
n
e
s
s
S
o
u
r
n
e
s
s
B
i
t
t
e
r
n
e
s
s
B
l
a
n
d
n
e
s
s
Cheese defects (week four post-packaging)
O
c
c
u
r
r
e
n
c
e
 
(
%
)
Tube Mould
Small Mould
 
Figure 4.44: Percentage of occurrence in nine major sensory defects for 
Camembert cheese made using tube moulds (n = 8) and small moulds (n = 8) at 
the age of four weeks. The samples were stored at 4 ? 1 ?C post-packaging. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  84 
4.8. Consumer acceptance 
Cheese samples were stored at 4 ? 1 ?C and after four weeks they were evaluated for 
consumer acceptability with consumer sensory panels. Sixteen cheese treatments 
inclusive of the four variables (level of fat, salting method, final acidification pH, and 
mould type) were evaluated by consumer sensory panellists, which consisted of four 
sensory sessions. Two of the four sensory sessions were done in duplicate, resulting in 
a total of six sessions. Each session included four different cheese samples at the age 
of four weeks with 21 to 43 panellists varying in each session. The panellists 
consisted of students, guests and staff of Massey University, Albany campus. 
 
Table 4.6 shows the mean and standard deviations in the sensory scores of the four 
treatments (level of fat, salting method, final acidification pH, and mould type) in 
terms of appearance, odour, flavour, texture, and overall acceptance, and Table 4.7 
shows the significant levels (p-values) between the treatments. The tables shows that 
the low-fat cheese samples were significantly more preferred in terms of appearance, 
texture, and overall acceptance; the brine-salted samples were significantly more 
preferred in appearance; the retentate-salted samples were significantly more 
preferred in flavour; and the tube mould samples were significantly more preferred in 
appearance and odour. 
 
Figures 4.45 to 4.48 show the combined overall sensory acceptance scores of the four 
treatments (level of fat, salting method, final acidification pH, and mould type) and 
their two levels. Apart from the level of fat, no significant differences were found 
between the consumer panels for the two levels of each treatment. This may be due to 
the high variation in consumers? preferences of the Camembert cheese, resulting in a 
wide spread of results (Appendix 1). 
  
The overall consumer acceptance of the sixteen cheese samples are shown in Figure 
4.49, providing more information about the consumer acceptance results for the 
overall acceptance of each treatment. All samples with the exception of HR5.2T 
scored above a score of five (?neither like nor dislike?), suggesting that most samples 
were acceptable to the sensory panellists (Figure 4.49). The cheese samples that were 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  85 
?slightly liked? (scored 6 -7) by the panellists were HB5.2S, HR4.9T, HR5.2S, 
LB4.9S, LB5.2T, LR4.9T and LR5.2T. 
 
It should be noted that from the results shown in Figure 4.49, the scores for most 
treatments covered a wide range, with upper quartiles of consumer acceptance scores 
for most cheese samples lying between six and eight, and lower quartiles of scores 
lying between four and six. It was also observed that LR4.9T had a large number of 
outliers as a cluster of the data were plotted outside the quartile range, suggesting two 
distinctive consumer groups were present during these sensory sessions. 
 
Table 4.6: Mean scores and standard deviations of the combined of the five 
sensory attributes. 
 
Cheese 
treatments n Appearance Odour Flavour Texture 
Overall 
Acceptance 
Low-fat 620 5.96 ? 1.71 5.56 ? 1.40 6.11 ? 1.68 6.08 ? 1.74 6.00 ? 1.60 
High-fat 225 5.24 ? 2.17 5.76 ? 1.57 6.02 ? 1.90 5.45 ? 2.20 5.61 ? 1.83 
Brine-salted 420 5.97 ? 1.75 5.67 ? 1.44 5.96 ? 1.78 5.94 ? 1.79 5.92 ? 1.64 
Retentate-
salted 425 5.57 ? 1.96 5.56 ? 1.46 6.21 ? 1.70 5.88 ? 1.99 5.88 ? 1.71 
pH 5.2 413 5.85 ? 1.84 5.60 ? 1.48 6.10 ? 1.71 6.00 ? 1.87 5.91 ? 1.71 
pH 4.9 432 5.69 ? 1.90 5.63 ? 1.41 6.07 ? 1.78 5.82 ? 1.90 5.89 ? 1.63 
Tube Mould 431 5.90 ? 1.79 5.72 ? 1.45 6.09 ? 1.72 6.01 ? 1.86 5.97 ? 1.64 
Small Mould 414 5.63 ? 1.94 5.50 ? 1.44 6.08 ? 1.77 5.81 ? 1.92 5.82 ? 1.71 
 
Table 4.7: p-values (p ? 0.05) of the four processing variables of the five sensory 
attributes. 
 
Variables Appearance Odour Flavour Texture Overall Acceptance 
Level of Fat 0.000 0.069 0.506 0.000 0.003 
Salting Method 0.002 0.253 0.035 0.616 0.719 
Final Acidification pH 0.227 0.716 0.774 0.176 0.869 
Mould Type 0.039 0.028 0.913 0.138 0.204 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  86 
Low-FatHigh-Fat
9
8
7
6
5
4
3
2
1
Level of Fat
O
v
e
r
a
l
l
 
A
c
c
e
p
t
a
n
c
e
 
Figure 4.45: Consumer overall acceptance of high-fat (n = 225) and low-fat (n = 
620) Camembert cheese at the age of four weeks. The cheese samples were stored 
at 4 ? 1 ?C post-packaging. ? represents the mean; horizontal lines represent 
upper quartile, median, and lower quartile respectively from top to bottom; * 
represents outliers. 
Retentate-SaltedBrine-Salted
9
8
7
6
5
4
3
2
1
Salting Method
O
v
e
r
a
l
l
 
A
c
c
e
p
t
a
n
c
e
 
Figure 4.46: Consumer overall acceptance of brine-salted (n = 420) and 
retentate-salted (n = 425) Camembert cheese at the age of four weeks. The cheese 
samples were stored at 4 ? 1 ?C post-packaging. ? represents the mean; 
horizontal lines represent upper quartile, median, and lower quartile 
respectively from top to bottom; * represents outliers. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  87 
pH 4.9pH 5.2
9
8
7
6
5
4
3
2
1
Final Acidification pH
O
v
e
r
a
l
l
 
A
c
c
e
p
t
a
n
c
e
 
Figure 4.47: Consumer overall acceptance of Camembert cheese acidified to pH 
5.2 (n = 413) and pH 4.9 (n = 432) at the age of four weeks. The cheese samples 
were stored at 4 ? 1 ?C post-packaging. ? represents the mean; horizontal lines 
represent upper quartile, median, and lower quartile respectively from top to 
bottom; * represents outliers. 
Small MouldTube Mould
9
8
7
6
5
4
3
2
1
Mould Type
O
v
e
r
a
l
l
 
A
c
c
e
p
t
a
n
c
e
 
Figure 4.48: Consumer overall acceptance of Camembert cheese made using 
small mould (n = 414) and tube mould (n = 431) at the age of four weeks. The 
cheese samples were stored at 4 ? 1 ?C post-packaging. ? represents the mean; 
horizontal lines represent upper quartile, median, and lower quartile 
respectively from top to bottom; * represents outliers. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  88 
L
R
5
.
2
T
L
R
5
.
2
S
L
R
4
.
9
T
L
R
4
.
9
S
L
B
5
.
2
T
L
B
5
.
2
S
L
B
4
.
9
T
L
B
4
.
9
S
H
R
5
.
2
T
H
R
5
.
2
S
H
R
4
.
9
T
H
R
4
.
9
S
H
B
5
.
2
T
H
B
5
.
2
S
H
B
4
.
9
T
H
B
4
.
9
S
9
8
7
6
5
4
3
2
1
Treatments
O
v
e
r
a
l
l
 
A
c
c
e
p
t
a
n
c
e
 
Figure 4.49: Consumer overall acceptance of sixteen treatments of Camembert 
cheese at the age of four weeks. The cheese samples were stored at 4 ? 1 ?C post-
packaging. The sample size of consumer panellists varied between 21 and 43 
panellists per treatment. ? represents the mean; horizontal lines represent 
upper quartile, median, and lower quartile respectively from top to bottom; * 
represents outliers. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  89 
5. DISCUSSION 
The impact of four processing variables (level of fat, salting method, final 
acidification pH, and mould type) on the quality and ripening characteristics of UF 
Camembert cheese were investigated in this study. It should be noted that factors 
affecting the cheese quality are not limited to the variables studied as other processing 
parameters play significant roles in cheese production. 
 
5.1. Moisture retention 
The moisture content and water-binding characteristics in cheese during ripening have 
significant impact on the rheology and texture of the cheese (Fox et al., 2000). High 
moisture retention by sorption of water in protein reduces firmness and contributes to 
a more elastic behaviour in cheese texture (Kneifel et al., 1991). During cheese 
ripening, dehydration of the cheese occurs as water evaporates from the cheese 
surface to the surrounding atmosphere (Riahi et al., 2007; Schlesser et al., 1992). 
Although Camembert is categorised as a type of cheese with high moisture content 
(Scott, 1998), the rate of dehydration during ripening is also attributed to the state of 
water and water-binding capacity of the protein network (Geurts et al., 1974; Kneifel 
et al., 1991). The incorporation of whey proteins in the cheese by UF increases the 
water-binding capacity, and influences the ripening of cheese as they act as inert filler 
within the paracasein network (Hinrichs, 2001; Harper et al., 1989; FIL/IDF, 1991). 
Analysis of TS reveals the changes in cheese moisture during the maturation period. 
Measurements of pH and proteolysis also act as useful indicators for the level of 
moisture retention, and the two factors are valuable in determining the quality of the 
cheese. 
 
Results from this study did not show significant increase in the mean TS for the 
cheese treatments from the four processing variables. This can be partly attributed to 
the moisture loss being located mostly in the rind and outer portion of the cheese, 
where the cheeses were sampled without the rind during TS determination. The 
absence of TS increase can also be attributed to the increased water-binding capacity 
in cheese due to proteolysis (Kneifel et al., 1991). In a study by Schlesser et al. 
(1992), TS in Camembert cheese increased during the first 15 d of maturation due to 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  90 
surface evaporation. The values of TS varied only slightly after 15 d, suggesting that 
the water-binding properties of the cheese changed as the cheese matured (Schlesser 
et al., 1992). 
 
As the paracasein is hydrolysed during ripening, more water-binding sites become 
available from the carboxylic acid and amino groups formed (Tarakci and Tuncturk, 
2008), thereby increasing water sorption (Schlesser et al., 1992). Studies on 
Emmentaler (Malthlouthi et al., 1981) and Gruy?re (Ruegg and Blanc, 1976) cheeses 
have shown that water sorption increases with a higher level of proteolysis. Hence, the 
lack of increase in TS can also be attributed to the general increase in NPN/TN and 
SN/TN ratios of the cheese treatments studied, which enables better water-binding 
capacity due to increased proteolysis. Additionally, the depletion of lactic acid and 
subsequent increase in pH from the isoelectric point of paracasein can also contribute 
to the increase in water sorption (Ruegg and Blanc, 1976). 
 
Effect of the level of fat 
The relatively lower TS in low-fat cheese samples from week five may be due to 
measuring error. However, it has been indicated that increase in cheese fat content 
leads to the replacement of water-binding proteins by fat globules. Subsequently, 
moisture is lost more readily by evaporation (Kneifel et al., 1991). 
 
Effect of the Salting Method 
Results from the two salting methods did not significantly differ in the cheese TS. 
Retentate-salting of the cheese was expected to have significant impact on the 
properties of cheese, as it affects starter growth, the duration of curd formation, and 
the subsequent texture of the formed curd (Walstra et al., 2006). When observing the 
level of proteolysis between the brine-salted and retentate-salted cheese samples, 
NPN/TN and SN/TN ratios were higher in those samples which were brine-salted, 
suggesting that they had better water-binding capacity. However, it should be noted 
that retentate-salted cheese samples had significantly lower TS at some point (weeks 1 
and 4) than those which were brine-salted during maturation. This is perhaps due to 
moisture entrapped within the cheese curd during coagulation of the salted retentate, 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  91 
as drainage of moisture would be minimal compared to those seen in brine-salting 
(Walstra et al., 2006). 
 
Effect of the final acidification pH 
Published data suggest that the water-binding capacity of casein increases as the pH is 
increased from the isoelectric point (Schlesser et al., 1992). This may be due to the net 
charge of the casein becoming more negative, which allows two to three times more 
water sorption than at the isoelectric point (Schlesser et al., 1992; Marchesseau et al., 
1997). The cheese samples that were acidified to pH 4.9 had significantly lower TS 
than samples acidified to pH 5.2, which suggested a higher moisture loss. It should be 
noted that these pH values were only used as end-points for the acidification stage 
during cheese manufacture. Subsequently, the cheeses were subjected to brine-salting 
or directly transferred to ripening, packaging, and some storage post-manufacture 
before the analysis were conducted. The pH levels of the cheeses were expected to 
considerably change during this period. Therefore, the lower TS in cheese samples 
acidified to pH 4.9 can be explained by a high level of proteolysis, indicated by higher 
NPN/TN and SN/TN ratios, which resulted in increased water-binding capacities of 
the proteins. 
 
Effect of the mould type 
The type of mould did not have a significant impact on cheese maturation in terms of 
TS content, NPN and SN indicators. Both tube and small moulds had similar 
dimensions, which enabled comparable surface evaporation (Riahi et al., 2007). 
Further, their effects on the maturation of cheese, particularly in proteolysis, are 
considered minor. In the cheese samples acidified to pH 5.2, relatively lower TS were 
shown in those samples made using small moulds. This may be explained by the 
increased risk of contamination by the non-starter bacteria, which become apparent 
only when the initial level of proteolysis is relatively low (Kousta et al., 2010). 
 
 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  92 
5.2. Salt uptake 
The uptake of salt is fundamentally based on the concentration gradient between the 
cheese moisture and the source of salting, conventionally the brine solution or dry salt 
crystals. The model for salt uptake is primarily based on brine-salting, but it also 
implies other salting methods (McSweeney, 2007). The rate of salt penetration 
depends on several factors, including the porosity of the cheese (Geurts et al., 1980), 
tortuosity of the routes of water within the cheese in bypassing obstructing particles 
(Walstra et al., 2006), the proportion of the total water bound (Melilli et al., 2006), the 
viscosity of the free water phase allowing for transport (Guo et al., 1997; Payne and 
Morison, 1999), and the interaction of sodium with the protein network (Payne and 
Morison, 1999). 
 
Effect of the level of fat 
The high-fat cheese samples were shown to have significantly lower salt uptake. Fat 
globules act as obstacles in the water channels within the protein network (Melilli et 
al., 2006). This decreases the effective diffusion coefficient by reducing the effective 
distance covered by the diffusing ions (Walstra et al., 2006). The effect of this seems 
to be milder in the high-fat samples combined with higher final acidification pH. 
 
Effect of the salting method 
The manufacturer produces two types of Camembert cheese by the conventional 
method and the UF method, using brine-salting and retentate-salting respectively. It is 
predictable that the manufacturer would have determined the salting parameters in 
both methods, to produce cheese with similar salt levels based on experiments in 
order to achieve consistency. This study applied both types of salting methods in the 
manufacture of UF Camembert, and surprisingly, no significant differences in salt 
uptake were observed. The major differences between the two salting methods are that 
retentate-salting affects the growth of the LAB starters and lactic acid production (Fox 
et al., 2004) and prolongs the acidification period up to 48 h, compared to a maximum 
of 24 h in brine salting. In contrast to the extensive acidification time in UF cheese 
due to increased buffering capacity (Walstra et al., 2006), acidification and curd 
formation in conventionally-made Camembert only takes about 30 min. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  93 
The behaviour of the salt gradient in retentate-salted cheese is also different to those 
that were brine-salted, where a salt gradient is created from the outer portion of the 
cheese block and diffuses inwards (Fox et al., 2000). Instead, the concentration 
gradient in retentate-salted cheese exists between dispersed salt crystals and the 
surrounding retentate. This creates patches of highly salted areas in the retentate, 
where clotting occurs simultaneously in this stage. The displacement rate of moisture 
within the retentate may be decreased as paracasein linkages form, but small drainage 
can occur and the salt eventually diffuses evenly within the fresh cheese (Geurts et al., 
1980). 
 
The growth of LAB starters is generally inhibited by the introduction of NaCl (Scott, 
1998; Fox et al., 2000). However, it has been reported that Lactococcus spp. have 
higher NaCl tolerance than thermophilic Streptococcus thermophilus and 
Lactobacillus spp. such as Lb. helveticus and Lb. delbrueckii subsp. lactis. 
(McSweeney, 2007). Therefore, retentate-salting may have a profound effect on the 
LAB starter profile during acidification and curd formation in UF Camembert, which 
subsequently affects proteolysis during cheese ripening. In this study, however,  
 
Effect of the final acidification pH 
It has been indicated that the increase in pH during curd formation allows better 
hydration of the casein micelles due to the higher negative charges. Hence, there is a 
better solubilisation of proteins which facilitates the diffusion of salt (Floury et al., 
2009). This probably explains the significantly higher salt uptake in cheese samples 
that were acidified to pH 5.2, as the rate of ion exchange was considerably higher 
(Walstra et al., 2006). 
 
Effect of the mould type 
The salt content in both tube and small moulds was not significantly different. The 
similar dimensions of the cheese samples produced using both moulds would have 
allowed comparable salt uptake during brining; and the amount of salt added per 
volume of retentate were the same in retentate-salted treatments. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  94 
5.3. Proteolysis 
During the ripening of cheese, casein is broken down into smaller peptides and N 
compounds by a variety of enzymes (Weimer, 2007). The degradation of protein plays 
a significant role in determining the biochemistry, texture, and organoleptic properties 
of the cheese (Walstra et al., 2006). The assessment of proteolysis by chemical 
analysis involves estimating different N fractions in the cheese samples. These 
include SN which consists of non-casein proteins and peptides, extracted at the 
isoelectric pH of paracasein (pH 4.6); and NPN which consists mainly of medium to 
short peptides and amino acids, obtained by treating the cheese extracts with TCA 
(e.g. 12%) solutions or with 70% aqueous ethanol (Ard?, 1999). The determination of 
N in these fractions can be carried out using the Kjeldahl method (Ard?, 1999), 
several forms of gel electrophoresis (Yeti?meyen et al., 2006; Hayaloglu et al., 2008), 
high performance liquid chromatography (Piraino et al., 2008), and gas 
chromatography (Hayaloglu et al., 2008). 
 
In this study, the NPN/TN and SN/TN ratios were used to estimate the level of 
proteolysis in cheese. All results showed increasing trends in NPN/TN and SN/TN 
ratios as the cheeses matured. It is apparent that the levels of proteolysis increases 
with the age of cheese. It can be observed that the differences in NPN/TN ratios 
between treatments of each processing variable were consistently larger than those 
observed in SN/TN, where significant differences in SN/TN between treatments were 
only present at a certain period during maturation. It may be reasonable to mention 
that SN produced mostly by the surface mould flora, Penicillium camemberti, were 
more readily degraded by bacterial enzymes, which produce smaller NPN molecules 
(Walstra et al., 2006). 
 
Effect of the level of fat 
It is generally acknowledged that having higher fat content increases the softness of 
cheese due to disruption of the casein matrix (Gunasekaran and Ak, 2003). The direct 
effects of fat content on the level of proteolysis in UF Camembert have not been 
reported much. However, studies on Cheddar cheese indicated that the fat content of 
cheese did not influence proteolysis (Kilcawley et al., 2007; Fenelon et al., 1999). In 
contrast, results in this study showed that the high-fat cheese samples had 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  95 
significantly higher NPN/TN and SN/TN (week 7) ratios, suggesting higher activities 
of proteolysis. This may be a result of reduced paracasein linkages due to increased 
fat content, allowing proteolysis to occur more readily (Schlesser et al., 1992). 
 
Effect of the salting method 
The two salting methods were presumed to have profound effects on the level of 
proteolysis, as discussed earlier. Retentate-salted cheese samples experienced a 
significantly longer acidification period during curd formation in order to reach 
desired acidification (21 ? 2 h for pH 5.2 and 45 ? 2 h for pH 4.9 compared to 7 ? 2 h 
and 20 ? 2 h respectively in brine-salted treatments). This can be attributed to the 
presence of salt in the clotting retentate, which effectively slows the growth of LAB 
starters. As a result, lactic acid production was significantly hindered by the presence 
of salt in addition to the already high buffering capacity of UF retentate (Walstra et 
al., 2006). The NPN/TN and SN/TN (week 5) ratios were significantly higher in the 
brine-salted cheese samples. This may be explained by the relatively better growth 
conditions for the LAB starters in the clotting retentate. In contrast, starter growth in 
the salted retentate could have been suppressed, thus reducing the production of 
proteolytic enzymes. 
 
Effect of the final acidification pH 
The interaction between clotting enzymes and caseins is important in cheese 
manufacture. Clotting enzymes specifically cleave the Phe105-Met106 bond of ?-casein 
during the clotting of the milk (Walstra et al., 2006; Larsson and Andren, 1997). At 
the same time, some of the clotting enzymes become adsorbed onto paracasein, 
becoming part of the cheese (Walstra et al., 2006). During cheese ripening, the 
clotting enzymes retained promote proteolysis in addition to the indigenous milk 
enzymes and enzymes from the starter cultures and secondary flora (Larsson and 
Andren, 1997; Trujillo et al., 2000). 
 
Cheese samples acidified to pH 4.9 had higher NPN/TN and SN/TN (week 7) ratios 
than those acidified to pH 5.2. The casein was readily coagulated at about pH 6.3 with 
the presence of Fromase, an aspartyl protease from Rhizomucor miehei commonly 
used as a chymosin substitute (Preetha and Boopathy, 1997). Further acidification 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  96 
occurs when LAB starters deplete the available lactose and produce higher levels of 
lactic acid (Fox et al., 2000). The time required for the cheese pH to achieve 4.9 is 
also considerably longer compared to pH 5.2. The increased proteolysis in cheese 
samples acidified to 4.9 can be attributed to a combination of increased retention of 
the clotting enzyme and prolonged acidification time. When the clotting enzyme is 
mixed with ?-casein, the highly negatively charged casein macropeptide (Met106-
Val169) is cleaved (Walstra et al., 2006). This forms para- ?-casein with an isoelectric 
pH >7, hence it is positively charged at pH below 7 (Larsson and Andren, 1997). The 
acid aspartyl protease from R. miehei, Fromase, and other common clotting enzymes 
such as chymosin have an acidic isoelectric pH (~4.6) (Larsson and Andren, 1997; 
Preetha and Boopathy, 1997), and are negatively charged at pH above 4.6. Therefore, 
there is a higher affinity of the protease on the paracasein at lower pH, which 
increases its retention in cheese and subsequently encourages proteolysis during 
cheese ripening (Preetha and Boopathy, 1997).  
 
The prolonged acidification time required to achieve pH 4.9 in the cheese would have 
allowed extensive growth of LAB starters and the production of cell envelope 
proteinases (Walstra et al., 2006). Proliferation of the starter cultures may get to a 
point where growth ceases and autolysis occurs, enabling the liberation of 
intracellular peptidases (Walstra et al., 2006). Additionally, the initial salt content in 
cheese samples acidified to pH 4.9 was lower. Therefore, the inhibitory effect of salt 
on the many of the cheese non-starter microflora decreases and their proteolytic action 
may be subsequently higher (McSweeney, 2007). 
 
Effect of the mould type. 
The effects of ?mould type? on proteolysis during cheese maturation would have been 
thought to be either none or minimal, as the types of moulds had little to do with the 
biochemistry of the cheese. No apparent differences were shown in the NPN/TN and 
SN/TN ratios between the two mould types, as expected. 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  97 
5.4. Textural and organoleptic properties 
The ripening of cheese contributes to the final texture due to the chemical and 
structural changes over time, which is primarily related to the hydrolysis of proteins. 
The paracasein network is degraded over time, resulting in a softer and more 
deformable cheese (Gunasekaran and Ak, 2003). It has been reported that proteolysis 
generally occurs in two stages, where about 20% of the casein network is hydrolysed 
in the initial stage, and more gradual breakdown takes place in the second stage 
(Rogers et al., 2009). Degradation of the paracasein in Camembert is facilitated 
mainly by the surface mould, such as Geotrichum and P. camemberti (Rash and 
Kosikowski, 1982). Extracellular proteases produced by these organisms hydrolyse 
proteins and lipids to peptides, amino acids, and fatty acids, which cause inward 
migration of softening in the cheese body (Walstra et al., 2006). In this study, 
softening in the UF Camembert (<7 weeks of shelf-life) occurred too rapidly 
compared to Camembert produced conventionally by the manufacturer (9 weeks of 
shelf-life), which effectively reduced its shelf-life and ability to meet consumer 
expectations. Other defective attributes such as rind discolouration, rind deformation, 
and thick rind also became apparent as the samples matured. 
 
5.4.1. Instrumental analysis 
Uniaxial compression tests were conducted to measure the firmness of the UF 
Camembert samples by measuring the force required to achieve 50% deformation 
during the maturation period. Results found no significant differences between 
treatments in each processing variable. Similar testing parameters were used in other 
studies on the texture of Camembert (Antoniou et al., 2000; Schlesser et al., 1992), 
with variations in the extent of height deformation. The major difference which 
differentiated the UF Camembert from the other cheeses such as Emmentaler or 
Cheddar was the rapid development of softness during maturation. Results showed 
that the force required to compress test samples to 50% height deformation decreased 
over time and then started to stabilise at week four. The cheese samples softened to an 
extent where the inner body of the cheese had obtained a continuous, flowing core 
texture. This texture was generally observed at weeks five and six for all cheese 
treatments. This posed an issue during the measurement of cheese firmness, as the 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  98 
sample pieces could no longer support their own weight after the sample had attained 
ambient temperature. 
 
Antoniou et al. (2000) measured Camembert cheese firmness at 10% height 
deformation, and Schlesser et al. (1992) at 20% height deformation, whereas other 
studies have measured firmness of other cheese types at a much higher height 
deformation (80%) (Gaya et al., 1990; Vinas et al., 2007). In this study, 50% height 
deformation was used, as neither the low nor high deformation heights adequately 
measured the cheese firmness. The thick rind of UF Camembert prevented accurate 
measurements of firmness on the cheese body at 10 and 20 % height deformations. At 
80% height deformation, the measurements obtained only represented the rind 
compaction. Hence, measurements of 50% height deformation were used as they 
adequately represented firmness of the cheese body until a flowing core texture was 
developed, where accuracy was compromised. Although the purpose of instrumental 
analysis was to mimic the behaviour of human panels, changes should be made 
depending on the prevailing conditions. It is therefore suggested that instrumental 
measurements of cheese with such extensive softness should be conducted at a cooler 
temperature, for example at 15 ?C. This would allow more accurate comparison 
between samples of different treatments and eliminate variations caused by 
inconsistent cheese texture at high temperatures. 
 
Effect of the level of fat 
It has been reported that lower fat content contributes to a firmer texture due to a 
denser protein network (McSweeney, 2007). In Cheddar cheese, it has been shown 
that low-fat Cheddar exhibited firmer texture when compared to full-fat Cheddar 
throughout maturation (Rogers et al., 2009). This behaviour is generally applicable to 
most cheese varieties, as more casein is replaced by fat globules, as less paracasein 
linkages are present. Even though fat globules have demonstrated the ability to be 
adsorbed in ?-casein and establish hydrophobic interactions (Su and Everett, 2003), 
they would still replace the stronger paracasein linkages. During cheese ripening, the 
reduced amount of paracasein is further degraded, resulting in a softer cheese texture. 
In the case of UF Camembert, the whey proteins incorporated in the cheese act 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  99 
similarly to fat globules, which further reduces the amount of paracasein linkages 
(Hinrichs, 2001; Harper et al., 1989). 
 
Effect of the salting method 
The salting methods did not show significant differences in terms of firmness. There 
is a lack of studies which uses retentate-salting in the manufacture of UF cheeses. 
This is not surprising as salting is known to have inhibitory effects on the starter 
growth during acidification (Fox et al., 2000). The combined mean of cheese 
produced using the two salting methods were not significantly different. However, 
results on the N fractions from this study showed that brine-salted cheese exhibited 
higher levels of proteolysis, and should decrease firmness of the cheese. This, 
however did not occur as the number of replicates may not be sufficient to produce 
significant texture results. 
 
Effect of the final acidification pH 
Although textural results did not indicate any significant differences between the final 
acidification pH, proteolysis in cheese samples which were acidified to pH 4.9 was 
significantly higher at weeks 5 and 7. This in theory indicates that a change in texture 
is probable, and however not exhibited in the results due to high variations.  
 
Effect of the mould type 
In terms of the mould type, no significant difference was observed between the mould 
types. The initial hypothesis was that cheese formed in separate small moulds rather 
than long tube moulds would allow more whey drainage during acidification of the 
retentate, thereby attaining a firmer cheese texture. However, this was not the case 
and TS between the two mould types did not differ significantly. 
 
5.4.2. Cheese grading 
Expert grading systems have been used in the industry by cheese manufacturers to 
ensure the quality of their products are consistent and ?fit for purpose? (Walstra et al., 
2006). Generally, most grading systems assess products based on two elements, 
including a descriptive analysis to determine the sensory attribute profile of the 
product; and whether the attribute profile meets the final product specifications 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  100 
(Dodds et al., 2002). Cheese is often graded periodically until the end of its shelf-life 
specified by the manufacturer, as this reflects the quality of the cheese on the market 
and in the hands of the consumers. This is of particular importance in the types of 
cheese which require extensive ripening as their sensory properties can change 
drastically. 
 
Cheese grading results at week four of the maturation period was used to investigate 
the quality of UF Camembert when it is likely to be consumed. Initially, the grading 
system used by the manufacturer was based on an exhaustive list of defective 
attributes, where scores were deducted when a defect was present in a cheese. For the 
purpose of this study, nine of the significant defects were presented as percentages of 
occurrences of all graded samples. The rate at which UF Camembert cheeses 
deteriorated appeared to have occurred fairly rapidly with the emergence of 
undesirable characteristics such as rind discolouration, rind deformation, core 
unevenness, excessive softness, and other flavour defects; hence effectively reducing 
the shelf-life of the product. This section attempts to discuss various defects detected 
in the cheese and associate them with effects of the processing variables used. 
 
Rind discolouration 
The defect of rind discolouration includes browning and bolding of the surface mould 
(Figure 5.1). A study with respect to mould browning in Camembert has been 
reported (Carreira et al., 2002). It was reported that P. camemberti was able to 
produce brown pigments in the presence of tyrosine and Mn2+, and was more intense 
in the more proteolytic strain. Additionally, the inoculation of a competing microflora, 
Yarrowia lipolytica, showed less browning. This suggests that browning is dependent 
on the proteolytic activity, the growth rate and density of the P. camemberti (Carreira 
et al., 2002). 
 
Rind discolouration occurred more in low-fat cheese samples, retentate-salted 
samples, and tube mould samples than their alternatives, as NPN/TN ratios suggested 
otherwise. The higher occurrence in cheese samples acidified to pH 4.9 was in 
agreement with proteolysis results. Rind discolouration was observed with >50% 
occurrence in all treatments. The high occurrence of this defect indicates that it is 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  101 
generally common among UF cheeses and an attempt to explain difference between 
the treatments may be difficult. The development of browning gives the consumers an 
unclean impression, and discolouration of the cheese affects the uniformity of the 
white mould. 
 
 
Figure 5.1: The rind discolouration defect, shown by browning of the mould, 
with some discolouration on the edges. 
 
Rind deformation 
Rind deformation includes unevenness, misshaping, and sometimes peeling of the 
rind (Figure 5.2). This may be attributed to the presence of calcium phosphate 
deposition at the surface of the cheese (Hannon et al., 2009). Large quantities of 
calcium salts are released into the retentate after UF, and they assist in the linkages of 
paracasein during curd formation. When the proteolysis of casein occurs, the calcium 
ion precipitates onto the surface of the cheese as calcium phosphate due to higher pH 
(Fox et al., 2000). This acts as a barrier and significantly reduces the rate of diffusion 
for various substrates required for surface mould growth (Hannon et al., 2009).  
 
Rind deformation was shown to be more apparent in high-fat cheese samples, which 
is in agreement with the corresponding proteolysis results. Retentate-salted samples 
and small mould samples also had higher occurrence of rind deformation, but they 
were not correlated with proteolysis results. The small mould cheese samples were the 
only treatment with 50% occurrence with this defect, the rest had <50% occurrence. 
Rind deformation is a major concern in the appearance of the cheese, as unevenness 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  102 
and peeling of the rind give consumers the impression that the cheese is not in optimal 
condition for consumption. 
 
 
Figure 5.2: The rind deformation defect, shown by unevenness and concaving of 
the rind. 
 
Thick rind 
Thick rind is a defect associated with the extensive thickness and height of the surface 
mould mycelium, P. camemberti. In UF Camembert, the thickness of the rind was 
higher than the thickness observed in Camembert of the same age manufactured 
conventionally (Figure 5.3). Thick rind was present in all treatments with >80% 
occurrence. This may be attributed to the higher growth rate of the surface mould in 
UF Camembert. This contributes to a hard papery chewing sensation as described by 
the judges during cheese grading which sometimes associated with bitterness. 
 
 
Figure 5.3: The thick rind defect, shown in the UF Camembert cheese (left) 
comparing to the Camembert made conventionally (right). 
 
 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  103 
Core unevenness 
The growth of the surface mould flora produces extracellular proteases which 
hydrolyse proteins and lipids into molecules with lower molecular weight. These 
enzymes soften the cheese inwards from the outer portion (Walstra et al., 2006). As 
the structure of the protein network in UF Camembert is already weak, the degraded 
portion of the cheese becomes apparent. This was shown by the flowing texture of the 
cheese paste in the outer portion, and a core with a texture which remained relatively 
firm or sometimes chalky (Figure 5.4). 
 
Core unevenness is often observed in young UF Camembert during the first few 
weeks of its shelf-life, and gradually disappears over time as the entire cheese paste 
obtains a flowing texture. Results showed that retentate-salted cheese samples and 
small mould samples had higher occurrences of core unevenness than their alternative 
treatments. Presumably at week four of the maturation period, most cheese samples 
would have obtained considerable softness throughout the cheese paste, shown by 
<30% occurrence in all treatments. However, samples from retentate-salting and small 
mould seemed to have resisted such textural changes in the core at this age, 
suggesting that the level of proteolysis had not attained full potential within the centre 
of these cheeses. 
 
Figure 5.4: The core unevenness defect, shown by the overripened soft and 
flowing texture in the outer portion and firm (still chalky) texture in the inner 
portion of the cheese paste. 
 
Softness 
The softness of the cheese results from a weak casein network structure. The 
compactness of the casein network and the amount of casein linkages present play a 
significant role (McSweeney, 2007). The compactness of the casein network is 
affected by the amount of casein compared with levels of fat, moisture, and other 
filler molecules such as whey. This influences the amount of paracasein linkages 
formed and therefore the overall firmness of the cheese (Walstra et al., 2006). 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  104 
Secondary factors include the retention of calcium levels, which is involved in the 
formation of paracasein linkages, by increasing the pH at whey drainage 
(McSweeney, 2007). The levels of clotting enzymes retained, proteases from lactic 
starters and secondary flora, and milk proteinases also influence the degree of 
breakdown in paracasein linkages during cheese ripening (Larsson and Andren, 1997; 
Grappin et al., 1985). 
 
In the UF Camembert cheese produced, the incorporation of whey proteins and the 
higher levels of proteolytic enzymes retained due to lack of the whey drainage step is 
likely to be the main factors contributing to the initial weak casein network and rapid 
protein degradation during maturation. Although the amount of retained whey 
proteins and proteolytic enzymes in the fresh cheese had not been analysed in this 
study, the rapid development of overripening characteristics and softness appeared to 
support this assumption. Softness in the cheese at week four was shown to be 
prevalent in all cheese treatments (>50% occurrence). Extreme cases of softening 
results in a flowing texture of the cheese paste (Figure 5.5). 
 
Occurrence of softness in cheese samples of high-fat and brine-salted treatments were 
slightly higher than their alternatives. This corresponds with their relatively higher 
NPN/TN and SN/TN ratios, indicating a higher level of proteolysis. Occurrence of 
softness in cheese samples acidified to pH 4.9 was much higher than those acidified to 
pH 5.2, which is in agreement with high NPN/TN and SN/TN ratios. The association 
of the final pH during the acidification of UF Camembert with the retention of 
calcium and clotting enzymes discussed earlier may contribute to such softness 
(McSweeney, 2007). Cheese samples made using small moulds also had higher 
occurrence in softness. This was also observed in the uniaxial compression test where 
small mould samples had significantly softer texture. 
 
 
Figure 5.5: Results of softening at a mild degree (left), and over softening (right) 
with the development of a flowing texture of the cheese paste. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  105 
Saltiness 
This is a defect where excessive saltiness is perceived in the cheese. Although 
published data suggest that the concentration of salt is directly related to the degree of 
saltiness (McSweeney, 2007), some results obtained from cheese grading suggested 
otherwise. It was observed that the initial salt content was significantly high in the 
low-fat cheese samples. However, the saltiness defect perceived in the high-fat 
samples had higher occurrence than that in the low-fat samples. 
 
The retentate-salted cheese samples had higher occurrence of saltiness than those 
which were brine-salted, although the initial salt content between the two treatments 
were not significant. It is possible that the distribution of salt can influence the overall 
saltiness. The brine-salted cheeses had initially higher concentrations of salt in the 
outer portion of the cheese body, which gradually diffused inwards (Melilli et al., 
2006; Walstra et al., 2006). Meanwhile, retentate-salted cheese had initially relatively 
homogeneous distribution of salt. It can therefore be stated that the inner portion of 
the brine-salted cheeses would be less salty due to a lower salt concentration, 
depending on the rate of salt diffusion (Fox et al., 2004); and the perception of 
saltiness in the relatively salt-concentrated outer portion of the cheese can be 
influenced by other flavours from the surface flora (McSweeney, 2007). In 
comparison to the more homogenous distribution of salt in retentate-salted cheese, the 
brine-salted cheese may overall be perceived less salty. 
 
The occurrence of saltiness in cheese samples acidified to pH 5.2 was higher than 
those acidified to pH 4.9. This is in agreement with the significantly higher initial salt 
content in pH 5.2 cheese samples. Furthermore, the occurrence of saltiness was equal 
in both mould types, which is also in agreement with their similar levels of initial salt 
content, suggesting that the type of mould does not influence saltiness. 
 
Sourness 
The presence of sourness in cheese may be influenced by the presence of lactate 
(Weimer, 2007; Fox et al., 2004). Depending on the rate of lactate metabolism which 
is affected by various factors, the residual lactate and the perceived sourness may 
reveal the age of the cheese. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  106 
 
The occurrence of sourness in the UF Camembert produced were generally low 
(<30%), with the exception of retentate-salted samples, as it is expected that the 
amount of residual lactate in the cheese at four weeks of maturation would be 
relatively low. Compared to their alternatives, the low-fat cheese samples had 
relatively higher occurrences in sourness. This is in agreement with the corresponding 
NPN/TN and SN/TN ratios, suggesting that lower levels of proteolysis in the low-fat 
cheese exhibited slower rate of lactic acid metabolism (Grappin et al., 1985). The 
higher occurrences of sourness in the brine-salted cheese may be explained by the 
better growing environment for LAB starters, compared to the salty environment 
during curd formation in retentate-salted cheese. Although the salt levels were not 
significantly different between the salting methods, the process and environments 
created by the two methods differ. The cheese samples acidified to pH 4.9 had higher 
occurrences of sourness, which could be attributed to more extensive acid 
fermentation to achieve the lower pH (Fox et al., 2004). Furthermore, there were 
higher occurrences of sourness in the small mould cheese samples, which were not 
expected. However, the differences were small and the occurrence of the defect in 
both mould types were below 30%. 
 
Bitterness 
Bitterness is a serious defect in cheese which results from the excessive accumulation 
of hydrophobic peptides derived from caseins (McSweeney, 2007; Engel et al., 2001). 
The bitter peptides usually have a molecular mass of < 6 kDa and a mean 
hydrophobicity of > 1400 cal per residue.  The development of bitterness in cheese 
may be caused by the undesirable patterns of proteolysis which causes the excessive 
production of bitter peptides, often by the clotting enzymes (Lemieux and Simard, 
1991); or insufficient peptidase activity to degrade hydrophobic peptides into free 
amino acids (McSweeney, 2007; Broadbent et al., 1998). 
 
Generally, all samples except brine-salted cheese had ?50% occurrence of bitterness. 
Published data indicate that low salt level leads to low ionic strength that weakens 
hydrophobic interactions between the caseins (Pastorino et al., 2003). This facilitates 
the action of clotting enzymes on hydrophobic regions of the caseins, which results in 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  107 
the excessive production of hydrophobic peptides (Broadbent and Steele, 2007; 
Broadbent et al., 1998; McSweeney, 2007). The occurrence of bitterness was higher 
in the high-fat cheese samples, as well as those acidified to pH 4.9. This is in 
agreement with the report of McSweeney (2007) as both treatments had significantly 
lower initial salt content. The relatively higher occurrence of bitterness in brine-salted 
samples could be associated with the distribution of salt, as lower salt concentration in 
the centre could have impacted on the levels of bitter peptides produced. The small 
mould cheese samples had higher occurrences in bitterness. The cause of this defect 
was unclear as it was unlikely that the mould type affected the proteolytic activities of 
clotting enzymes (Lemieux and Simard, 1991), the action of peptidases (Broadbent et 
al., 1998), and the salt content (McSweeney, 2007), which are the primary factors that 
influence bitter peptide production. 
 
Blandness 
Blandness is due to the lack of typical Camembert flavours and aromas which causes 
the cheese to be uninteresting and flat in taste (Weimer, 2007). Flavour compounds 
are generated from the action of enzymes from clotting agents, milk, the starter and 
non-starter bacteria, as well as non-enzymatic conversions. The three main pathways 
of flavour formation include glycolysis, lipolysis, and proteolysis. The main sources 
of enzymes involved in these pathways are the starter cultures, such as Lactococcus 
lactis, Lactobacillus spp., Streptococcus thermophilus, Leuconostoc spp.; as well as 
secondary cultures used in surface-ripened cheese including Camembert and Brie, 
such as Penicillium camemberti and Geotrichum candidum (Smit et al., 2005). 
 
The fermentation of lactose mainly produces lactate by LAB, but a portion of the 
intermediate pyruvate can be converted to various flavour compounds such as 
diacetyl, acetoin, acetaldehyde, and acetic acid (Walstra et al., 2006). Free fatty acids 
produced from lipolysis act as precursors of flavour compounds such as 
methylketones, secondary alcohols, esters and lactones. These are mainly produced by 
additional cultures such as moulds in surface-ripened cheeses as they have high 
activities in fat conversion (Smit et al., 2005). A secondary alcohol, 1-octen-3-ol for 
example, has a mushroom type odour with a perception threshold of 0.01 mg per kg 
and is a key aromatic compound in Camembert cheese (Weimer, 2007). The 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  108 
degradation of casein yields small peptides and free amino acids, of which are further 
converted to various alcohols, aldehydes, acids, esters and sulphur compounds. 
 
Table 5.1: Examples of important flavour components in Camembert cheese 
(Fox et al., 2004). 
 
Type Metabolism Flavour Components 
Amino acid 3-Methylbutyrate 
 3-Methylbutanal 
 Methional 
 Methanethiol 
 Dimethylsulphide 
 Benzaldehyde 
 Phenylacetaldehyde 
Sugar 2,3-Butanedione 
Fat 1-Octen-3-ol 
 Butyric acid 
 1-Octen-3-one 
 2-Undecalactone 
 ?-Decalactone 
Combined pathways Phenylethyl acetate 
 
Table 5.1 shows examples of important flavour components in Camembert cheese. 
Blandness is associated with insufficient formation of such typical flavour compounds 
in the cheese, which is a concern in the UF Camembert. Characterisation of the 
flavour and volatile compounds in the UF Camembert has not been conducted in this 
study to investigate the associated metabolic pathways. However, it is speculated that 
as the UF Camembert has a short shelf-life due to the rapid breakdown of its weak 
protein structure, proper flavour development has yet to reach its fullest potential. 
 
It was assumed that higher fat levels and longer acidification times (lower pH) would 
allow the production of more flavour compounds, due to the availability of more 
substrate for lipolysis and higher levels of LAB enzymes, respectively. The higher 
proteolysis in brine-salted samples is also expected to have higher levels of flavour 
formation. However, the occurrence of blandness for these treatments indicated 
otherwise. This suggests the complexity in flavour development of cheese is rather 
high, and can not be sufficiently manipulated by altering a few basic manufacturing 
parameters. It has been suggested that the profile of starter cultures and strains of 
mould flora used can significantly influence the flavour development in Camembert 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  109 
cheese (Smit et al., 2005; Boutrou et al., 2006). This is mainly associated with the 
production of peptidases, which facilitate the conversion of peptides into free amino 
acids and various flavour compounds (Smit et al., 2005; Weimer, 2007). 
 
5.5. Consumer acceptance 
Consumer perception of a cheese consists of several factors, including appearance, 
odour, flavour, and texture (Marchesseau et al., 1997). Panellists in this study rated 
the cheese samples according to appearance, odour, flavour, texture, and overall 
acceptance on a nine-point scale ranging from ?extremely dislike? to ?extremely like?. 
It was observed that the scores were not normally distributed, suggesting that there is 
a wide variation in preference. Although the comments from the sensory evaluation 
were not analysed, a large portion of the panellists defined the cheese samples as ?too 
soft? or ?too runny?. Most panellists in this group generally do not consume 
Camembert regularly and are accustomed to cheeses with the Cheddar-type texture. A 
small group of panellists defined the cheese samples to have good texture but ?lacked 
Camembert flavour?. This group of panellists were generally knowledgeable in 
specialty cheeses. They had a preference for soft and strong-flavoured Camembert, 
whereby blandness of the UF cheeses became an issue. Generally, all cheese samples 
were acceptable as their mean overall acceptance score were above ?neither like nor 
dislike?, with one exception, the HB5.2T. 
 
Appearance 
The appearance was significantly different between samples produced with different 
levels of fat, salting method, and mould type; with preference for the low-fat, brine-
salted, and tube mould samples. During sensory evaluation, the cheese samples were 
cut into wedges; appearance of the cheese generally involved examining the colour of 
the rind, shape of the cheese, as well as the colour and texture of the cheese paste, as 
revealed by the cut.  
 
The high-fat cheese samples were less preferred, possibly due to more apparent rind 
deformation and softness described in cheese grading. A characteristic in the 
overripened cheese is the flowing texture of the cheese paste, often described as ?too 
runny?. At serving temperature (18 ? 2 oC), cheese samples with a flowing cheese 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  110 
paste tended to collapse as they could no longer support their shape, which was 
undesirable. The retentate-salted cheese samples were less preferred, possibly owing 
to higher occurrence in rind discolouration and deformation according to cheese 
grading. Although the dimensions of the small mould cheese samples were similar to 
the tube mould samples, they were not exactly identical. Furthermore, the risk of 
contamination in small mould cheeses was higher due to a higher exposed surface 
area during acidification and curd formation, which could have contributed to the low 
average sensory scores. 
 
Odour and flavour 
The UF Camembert cheeses were generally described as lacking odour and flavour, 
especially the typical mushroom aroma in Camembert described by the judges from 
cheese grading. Although the level of proteolysis in UF Camembert seemed to be 
high, due to the combination of an initially weak protein network (Hinrichs, 2001) and 
increased enzyme retention during manufacture, specific metabolic pathways are 
required to produce the right volatile and flavour components (Smit et al., 2005). The 
odour of the cheese samples was found to be significantly different between the 
mould types. This was surprising as the type of mould was not expected to play a 
significant role in the formation of flavour compounds due to the only variable was 
surface area. The salting method had a significant impact on the panellists? perception 
of the flavour of the cheese. The high occurrence of saltiness in retentate-salted 
samples seemed to be relatively less desirable for the consumers, where excessive 
saltiness masks the complexity of flavours leading to an undesirable cheese. 
 
Texture 
The level of fat significantly influenced the consumers? perception in the texture of 
the cheese. Higher fat content contributes to a weaker protein structure, and hence a 
softer cheese, whereas in extreme cases, it is undesirable to consumers who prefer 
firmer cheeses. This corresponded to the higher occurrence of softness in the high-fat 
cheese samples, and higher NPN/TN ratios. The texture of high-fat samples were 
generally described as ?too soft? or ?too runny?, which also resulted in the collapsing 
of the ?cut sample? during sensory evaluation. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  111 
Overall acceptance 
In terms of overall acceptance of the various UF Camembert treatments, the low-fat 
cheese samples were significantly more preferred than the high-fat samples. Although 
two of the high-fat treatments (HR4.9T and HB5.2S), grouped by cluster analysis, had 
relatively high scores, the quartile ranges of the other high-fat treatments were rather 
broad. In the low-fat treatments, there was generally less variation in the sensory 
scores given, shown by the smaller quartile ranges. This observation suggested that 
the low-fat samples were consistently scored higher, which is desired by 
manufacturers as consistent consumer acceptability is key to product success. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  112 
6. CONCLUSIONS 
 
The application of UF in Camembert cheese-making has created several advantages 
such as improved yield, reduced cost of labour, improved plant efficiency, reduced 
pollution, and the possibility for automation. However, adjustments in the 
manufacturing processes and parameters are required in order to realise the benefits of 
UF. This includes adapting to the high buffering capacity of the retentate, determining 
new quantities of clotting enzymes and starter cultures, and in particular, the removal 
of the whey drainage step. The lack of whey drainage not only allows the 
incorporation of whey proteins into the cheese which improves yield, but it also has 
significant consequences in the quality and ripening characteristics of the product. 
 
In the UF cheese, whey proteins act as a filler component which weakens the cheese 
structure by reducing the compactness of the casein network. The water-binding 
capacity of the proteins in cheese is also altered by the increased whey content. 
Prolonged acidification due to increased buffering capacity influences the profile of 
the starter culture and the composition of their metabolites. Enzymes, particularly 
those from the clotting agent and starter cultures are retained in the cheese due to the 
lack of whey drainage, which contribute to the breakdown of the casein network 
during cheese ripening. The combination of these factors results in a cheese which 
tends to ripen fairly rapidly, and has a short shelf-life with apparent emergence of 
various quality defects. 
 
In this study, manipulation of the process variables has shown that the use of high-fat 
milk or retentate is not desirable. The increased fat content has been indicated to 
reduce the amount of paracasein linkages by acting as filler, which reduces the 
compactness of the casein network. This weakens the structure of the UF cheese, 
causing the cheese to soften relatively quickly with more rapid proteolysis. Salt 
uptake is also hindered as the tortuosity is reduced by fat globules. 
 
The method of salting had significant effects on the quality and ripening 
characteristics of the cheese. Retentate-salted cheese had lower levels of proteolysis, 
possibly due to impaired growth of the starter cultures in the salted retentate. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  113 
However, cheese grading showed higher occurrences of rind discolouration and 
deformation in retentate-salted cheeses, as well as being significantly less preferred in 
terms of flavour by consumers. However, no significant differences were found in the 
overall acceptance of the cheeses when compared to brine-salted cheese. Considering 
the labour-intensiveness in brining and brine maintenance, it is more preferable to use 
retentate-salting over brine-salting. It is however important to be aware that retentate-
salting can influence the growth of starter cultures, such as selectively encouraging 
salt-tolerant strains or specific metabolic pathways, which can consequently influence 
the cheese quality. 
 
The extent of acidification in UF Camembert had a significant influence on cheese 
quality. The higher final acidification pH of 5.2 is more preferable as it is associated 
with lower levels of proteolysis and lower occurrences of sensory defects including 
softness, rind discolouration and deformation. The extent of lactic starter growth is 
not as high due to a shorter acidification time, which leads to a relatively lower 
production in LAB proteases that contributes to proteolysis during cheese ripening. 
The advantage of using pH 5.2 over pH 4.9 as an acidification endpoint significantly 
reduced acidification time, possibly leading to high throughput in production. 
 
The effect of mould type on cheese ripening seems to be insignificant. Although the 
use of small moulds is associated with softer texture and several other defects, it 
offers the possibility for an optimised or automated demoulding process, as well as 
producing Camembert with new shapes. However, the use of tube moulds is still 
recommended as an existing demoulding and cheese cutting system is already in 
place. Cheeses produced from tube moulds also seemed to be more stable with fewer 
defects and had firmer texture. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  114 
7. RECOMMENDATIONS 
 
It is recommended that processing variables which involves improving the protein 
network of the cheese or reducing the level of proteolysis during cheese ripening 
should be studied further. This primarily involves manipulating the enzymes retained 
in the curd by investigating the optimal type and amount of clotting agent used, as 
well as the proteolytic properties of the lactic starter and mould cultures used. This is 
expected to reduce degradation of the protein network in the cheese and hence 
produce a firmer texture. The lactic starter and mould cultures also play an important 
role in flavour development. It has been suggested that strains of L. lactis have 
varying levels of proteolytic and peptidolytic activities. It would be interesting to use 
strains with different proteolytic and peptidolytic profile in the production of UF 
Camembert to find ways of improving flavour formation without encouraging too 
much structural degradation. Additionally, the use of Geotrichum candidum along 
with P. camemberti has been suggested to improve the sensory qualities of 
Camembert cheese and reduce bitterness. It is possible for G. candidum to influence 
the extensive growth of the mould and the thick rind defect of UF Camembert 
cheeses. It may also be able to introduce a new flavour profile to the blandness of the 
cheese. 
 
Shelf-life of the cheese is generally dependent on the degradation of proteins and 
formation of flavour compounds. The metabolic pathways involved are highly 
complex and dynamic, where the result depends on the net effect of numerous 
cascading processes. In UF Camembert, the shelf-life can generally be improved by 
strengthening the protein network and reducing the level of proteolysis during cheese 
ripening. In the context of commercial aspects, considering implementation of 
processes, time, and ease of handling, it is recommended that UF Camembert is 
produced using these factors: low-fat, retentate-salting, final acidification pH of 5.2, 
and tube mould. 
 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  115 
8. REFERENCES 
 
AGRAWAL, P. & HASSAN, A. N. (2008) Characteristics of reduced fat Cheddar 
cheese made from ultrafiltered milk with an exopolysaccharide-producing 
culture. Journal of Dairy Research, 75, 182-188. 
ANTONIOU, K. D., PETRIDIS, D., RAPHAELIDES, S., BEN OMAR, Z. & 
KESTELOOT, R. (2000) Texture Assessment of French Cheeses. Journal of 
Food Science, 65, 168-172. 
AOAC INTERNATIONAL (2005) AOAC Official Methods of Analysis, Gaithersburg, 
AOAC International. 
ARD?, Y. (1999) Evaluating Proteolysis by Analysing the N Content of Cheese 
Fractions. Bulletin of the IDF, 337, 4-9. 
AULDIST, M. J. & HUBBLE, I. B. (1998) Effects of mastitis on raw milk and dairy 
products. Australian Journal of Dairy Technology, 53, 28-36. 
BASTIAN, E. D. & BROWN, R. J. (1996) Plasmin in milk and dairy products: An 
update. International Dairy Journal, 6, 435-457. 
BOUTROU, R., AZIZA, M. & AMRANE, A. (2006) Enhanced proteolytic activities 
of Geotrichum candidum and Penicillium camembertii in mixed culture. 
Enzyme and Microbial Technology, 39, 325-331. 
BRANDSMA, R. L. & RIZVI, S. S. H. (2001) Manufacture of Mozzarella cheese 
from highly-concentrated skim milk microfiltration retentate depleted of whey 
proteins. International Journal of Food Science and Technology, 36, 611-624. 
BROADBENT, J.R. and STEELE, J.L. (2007). Proteolytic enzymes of lactic acid 
bacteria and their influence on bitterness in bacterial-ripened cheeses. ACS 
Symposium Series, 971, 193-203. 
BROADBENT, J. R., STRICKLAND, M., WEIMER, B. C., JOHNSON, M. E. & 
STEELE, J. L. (1998) Peptide Accumulation and Bitterness in Cheddar 
Cheese Made Using Single-Strain Lactococcus lactis Starters with Distinct 
Proteinase Specificities. Journal of Dairy Science, 81, 327-337. 
BUTIKOFER, U., RUEGG, M. & ARD?, Y. (1993) Determination of Nitrogen 
Fractions in Cheese: Evaluation of a Collaborative Study. Food Science and 
Technology, 26, 271-275. 
CARREIRA, A., DILLINGER, K., ELISKASES-LECHNER, F., LOUREIRO, V., 
GINZINGER, W. & ROHM, H. (2002) Influence of selected factors on 
browning of Camembert cheese. Journal of Dairy Research, 69, 281-292. 
COOLBEAR, T., CROW, V., HARNETT, J., HARVEY, S., HOLLAND, R. & 
MARTLEY, F. (2008) Developments in cheese microbiology in New Zealand-
Use of starter and non-starter lactic acid bacteria and their enzymes in 
determining flavour. International Dairy Journal, 18, 705-713. 
DODDS, T. M., GREGORY, S. P. & GREENWOOD, S. L. (2002) Streamlining the 
grading function from market to manufacture. Australian Journal of Dairy 
Technology, 57, 164. 
ELWELL, M. W. & BARBANO, D. M. (2006) Use of microfiltration to improve 
fluid milk quality. Journal of dairy science., 89 Suppl 1. 
ENGEL, E., NICKLAUS, S., SEPTIER, C., SALLES, C. & LE QUERE, J. L. (2001) 
Evolution of the Taste of a Bitter Camembert Cheese during Ripening: 
Characterization of a Matrix Effect. Journal of Agricultural and Food 
Chemistry, 49, 2930-2939. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  116 
EXTERKATE, F. A. (2006) Controlled permeabilization of Lactococcus lactis cells 
as a means to study and influence cheese ripening processes. International 
Dairy Journal, 16, 788-796. 
FAO (2008) Dairy Summit. Food and Agricultural Organisation of the United 
Nations. Rome, Italy. 
FENELON, M. A., GUINEE, T. P., DELAHUNTY, C., MURRAY, J. & CROWE, F. 
(1999) Composition and sensory attributes of retail Cheddar cheese with 
different fat contents. Journal of Food Composition and Analysis, 13, 13-26. 
FERRON-BAUMY, C., MAUBOIS, J. L., GARRIC, G. & QUIBLIER, J. P. (1991) 
Milk and retentate rennet coagulation: Effects of several heat treatments. Le 
Lait, 71, 423-434. 
FIL/IDF (1991) New Applications of Membrane Processes. International Dairy 
Federation, Brussels, Belgium. 
FIL/IDF (2007) World Dairy Situation 2007. International Dairy Federation, Brussels, 
Belgium. 
FLOURY, J., ROUAUD, O., LE POULLENNEC, M. & FAMELART, M.-H. (2009) 
Reducing salt level in food: Part 2. Modelling salt diffusion in model cheese 
systems with regards to their composition. LWT - Food Science and 
Technology, 42, 1621-1628. 
FOUCAUD, C. & JUILLARD, V. (2000) Accumulation of casein-derived peptides 
during growth of proteinase-positive strains of Lactococcus lactis in milk: 
Their contribution to subsequent bacterial growth is impaired by their internal 
transport. Journal of Dairy Research, 67, 233-240. 
FOX, P. F. (Ed.) (1993) Cheese: Chemistry, Physics and Microbiology, London, 
Chapman & Hall. 
FOX, P. F., GUINEE, T. P., COGAN, T. M. & MCSWEENEY, P. L. H. (Eds.) 
(2000) Fundamentals of Cheese Science, Gaithersburg, Aspen Publishers, Inc. 
FOX, P. F., MCSWEENEY, P. L. H., COGAN, T. M. & GUINEE, T. P. (Eds.) 
(2004) Cheese - Chemistry, Physics and Microbiology, Oxford, Elsevier 
Acedemic. 
FSANZ (2008) Australia New Zealand Food Standard Code. Fermented Milk 
Products. Melbourne, Anstat Pty Ltd. 
FSANZ (2009) Microbiological Risk Assessment of Milk Cheese. Risk Assessment 
Microbiology Section. Food Standards Australia New Zealand. 
GAYA, P., MEDINA, M., RODRIGUEZ-MARIN, M. A. & NUNEZ, M. (1990) 
Accelerated ripening of ewe's milk Manchego cheese: The effect of elevated 
ripening temperatures. Journal of Dairy Science, 73, 26-32. 
GEURTS, T. J., WALSTRA, P. & MULDER, H. (1974) Water binding to milk 
protein, with particular reference to cheese. Netherlands Milk Dairy Journal, 
28, 64-72. 
GEURTS, T. J., WALSTRA, P. & MULDER, H. (1980) Transport of salt and water 
during salting of cheese. 2. Quantities of salt taken up and moisture lost. 
Netherlands Milk Dairy Journal, 34, 229-254. 
GRAPPIN, R., RANK, T. C. & OLSON, N. F. (1985) Primary Proteolysis of Cheese 
Proteins During Ripening. A Review. Journal of Dairy Science, 68, 531-540. 
GUNASEKARAN, S. & AK, M. M. (Eds.) (2003) Cheese Rheology and Texture, 
Boca Raton, CRC Press. 
GUO, M. R., GILMORE, J. A. & KINDSTEDT, P. S. (1997) Effect of sodium 
chloride on the serum phase of Mozzarella cheese. Journal of Dairy Science, 
80, 3092-3098. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  117 
HAJIPOUR, M., SOLTANIEH, M. & YAZDANSHENAS, M. (2010) Investigation 
of membrane fouling in cross flow microfiltration of non-alcoholic beer and 
modeling of tubular membrane flow. Desalination, 251, 20-28. 
HANNON, J. A., DEUTSCH, S.-M., MADEC, M.-N., GASSI, J.-Y., 
CHAPOT-CHARTIER, M.-P. & LORTAL, S. (2006) Lysis of starters in 
UF cheeses: Behaviour of mesophilic lactococci and thermophilic lactobacilli. 
International Dairy Journal, 16, 324-334. 
HANNON, J. A., LORTAL, S., TISSIER, J.-P. & FAMELART, M.-H. (2009) 
Limited ripening of low-fat UF-cheese due to CaPO4 barrier? Dairy Science 
and Technology, 89, 555-568. 
HARJU, M., HEIKONEN, MATTI (1989) A process for the demineralization of 
whey or a liquid based on whey and for producing an acid mixture and a base 
mixture useful in food industries., Valio, Meijerien, Finland. 
HARPER, J., IYER, M., KNIGHTON, D. & LELIEVRE, J. (1989) Effects of Whey 
Proteins on the Proteolysis of Cheddar Cheese Slurries (A Model for the 
Maturation of Cheeses Made from Ultrafiltered Milk). Journal of Dairy 
Science, 72, 333-341. 
HAYALOGLU, A. A., BRECHANY, E. Y., DEEGAN, K. C. & MCSWEENEY, P. 
L. H. (2008) Characterization of the chemistry, biochemistry and volatile 
profile of Kuflu cheese, a mould-ripened variety. LWT - Food Science and 
Technology, 41, 1323-1334. 
HERSLETH, M., ILSENG, M. A., MARTENS, M. & N??S, T. (2005) Perception of 
Cheese: A comparison of quality scoring, descriptive analysis and consumer 
responses. Journal of Food Quality, 28, 333-349. 
HERZBERG, M., BERRY, D. & RASKIN, L. (2010) Impact of microfiltration 
treatment of secondary wastewater effluent on biofouling of reverse osmosis 
membranes. Water Research, 44, 167-176. 
HINRICHS, J. (2001) Incorporation of whey proteins in cheese. International Dairy 
Journal, 11, 495-503. 
JANA, A. H. & THAKAR, P. N. (1996) Recombined milk cheese - A review. 
Australian Journal of Dairy Technology, 51, 33-43. 
J?NSSON, A.-S. & TR?G?RDH, G. (1990) Ultrafiltration applications. 
Desalination, 77, 135-179. 
KANG, Y. S. (1999) Membrane industries and R and D in Korea. Proceedings of the 
1998 Membrane Technology/Separations Planning Conference, Newton, 
Massachusetts, USA, 7-9 December 1998, 59-63. 
KARAMI, M., EHSANI, M. R., MOUSAVI, S. M., REZAEI, K. & SAFARI, M. 
(2009) Microstructural properties of fat during the accelerated ripening of 
ultrafiltered-Feta cheese. Food Chemistry, 113, 424-434. 
KELLY, A. L., HUPPERTZ, T. & SHEEHAN, J. J. (2008) Pre-treatment of cheese 
milk: Principles and developments. Dairy Science and Technology, 88, 549-
572. 
KILCAWLEY, K. N., O'CONNEL, P. B., HICKEY, D. K., SHEEHAN, E. M., 
BERESFORD, T. P. & MCSWEENY, P. L. H. (2007) Influence of 
composition on the biochemical and sensory characteristics of commercial 
Cheddar cheese of variable quality and fat content. International Journal of 
Dairy Technology, 60, 81-88. 
KNEIFEL, W., PAQUIN, P., ABERT, T. & RICHARD, J.-P. (1991) Water-Holding 
Capacity of Proteins with Special Regard to Milk Proteins and Methodological 
Aspects - A Review. Journal of Dairy Science, 74, 2027-2041. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  118 
KO, M. K., PELLEGRINO, J. J., NASSIMBENE, R. & MARKO, P. (1993) 
Characterization of the adsorption-fouling layer using globular proteins on 
ultrafiltration membranes. Journal of Membrane Science, 76, 101-120. 
KOSIKOWSKI, F. V. & MISTRY, V. V. (Eds.) (1997) Cheese and fermented milk 
foods, Westport, Edwards Brothers. 
KOUSTA, M., MATARAGAS, M., SKANDAMIS, P. & DROSINOS, E. H. (2010) 
Prevalence and sources of cheese contamination with pathogens at farm and 
processing levels. Food Control, 21, 805-815. 
KRIK, R. S. & SAWYER, R. (Eds.) (1991) Pearson's Composition and Analysis of 
Foods, Edinburgh, Longman Group UK Limited. 
KRISTENSEN, J. M. B. (1999) Cheese Technology - A Northern European 
Approach, Aarhus, International Dairy Books. 
LARSSON, K. I. & ANDREN, A. (1997) Affinity Between Chymosin and Individual 
Caseins at Varying pH-values. International Dairy Journal, 7, 615-618. 
LEE, D. Y., LI, Y. Y., NOIKE, T. & CHA, G. C. (2008) Behavior of extracellular 
polymers and bio-fouling during hydrogen fermentation with a membrane 
bioreactor. Journal of Membrane Science, 322, 13-18. 
LEMIEUX, L. & SIMARD, R. E. (1991) Bitter flavour in dairy products. I. A review 
of the factors likely to influence its development, mainly in cheese 
manufacture. Lait, 71, 599-636. 
LINTON, M., MACKLE, A. B., UPADHYAY, V. K., KELLY, A. L. & 
PATTERSON, M. F. (2008) The fate of Listeria monocytogenes during the 
manufacture of Camembert-type cheese: A comparison between raw milk and 
milk treated with high hydrostatic pressure. Innovative Food Science and 
Emerging Technologies, 9, 423-428. 
LO, C. G. & BASTIAN, E. D. (1998) Incorporation of Native and Denatured Whey 
Proteins into Cheese Curd for Manufacture of Reduced Fat, Havarti-type 
Cheese. Journal of Dairy Science, 81, 16-24. 
LORAIN, O., DUFAYE, P. E., BOSQ, W. & ESPENAN, J. M. (2010) A new 
membrane bioreactor generation for wastewater treatment application: 
Strategy of membrane aeration management by sequencing aeration cycles. 
Desalination, 250, 639-643. 
LUCE, C. & REX DIETERLE, R. D. (2005) Membrane applications for recovery & 
renewable energy. AIChE Annual Meeting, Conference Proceedings. 
MALTHLOUTHI, M., MICHEL, J. F. & MAITENAZ, P. C. (1981) Study of Some 
Factors Affecting Water Vapor Sorption of Gruyere Cheese. Food Science and 
Technology, 14, 163. 
MARCHESSEAU, S., GASTALDI, E., LAGAUDE, A. & CUQ, J.-L. (1997) 
Influence of pH on Protein Interactions and Microstructure of Process Cheese. 
Journal of Dairy Science, 80, 1483-1489. 
MATIS, K. A., LAZARIDIS, N. K., ZOUBOULIS, A. I., GALLIOS, G. P. & 
MAVROV, V. (2005) A hybrid flotation - Microfiltration process for metal 
ions recovery. Journal of Membrane Science, 247, 29-35. 
MAUBOIS, J. L. (2002) Membrane microfiltration: a tool for a new approach in dairy 
technology. Australian Journal of Dairy Technology, 57, 92-96. 
MAUBOIS, J. L. & MOCQUOT, G. (1975) Application of Membrane Ultrafiltration 
to Preparation of Various Types of Cheese. Journal of Dairy Science, 58, 
1001-1007. 
McSWEENEY, P. L. H. (Ed.) (2007) Cheese Problems Solved, Woodhead Publishing 
Limited. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  119 
MEILGAARD, M. C., CARR, B. T. & CIVILLE, G. V. (Eds.) (2007) Sensory 
Evaluation Techniques, Boca Raton, Taylor & Francis. 
MELILLI, C., BARBANO, D. M., CACCAMO, M., TUMINELLO, L., CARPINO, 
S. & LICITRA, G. (2006) Interaction of Brine Concentration, Brine 
Temperature, and Presalting on Salt Penetration in Ragusano Cheese. Journal 
of Dairy Science, 89, 1420-1438. 
MIO?, B., PRPI?, Z., VNU?EC, I., BARA?, Z., SU?I?, V., SAMAR?IJA, D. & 
PAVI?, V. (2008) Factors affecting goat milk yield and composition. 
Mljekarstvo, 58, 305-313. 
MISTRY, V. V. & MAUBOIS, J.-L. (Eds.) (1993) Application of Membrane 
Separation Technology to Cheese Production, London, Chapman & Hall. 
MOREIRA, M., STORANI, E., BEVILACQUA, A. & DE ANTONI, G. (2003) 
Effects of starter culture on the characteristics of tybo cheese. Italian Journal 
of Food Science, 15, 287-294. 
MULDER, M. (Ed.) (1996) Basic principles of membrane technology. Kluwer 
Academic. Boston. 
NAGENDRAN, A. & MOHAN, D. (2008) Protein separation by cellulose 
acetate/sulfonated poly(ether imide) blend ultrafiltration membranes. Journal 
of Applied Polymer Science, 110, 2047-2057. 
ONG, L., HENRIKSSON, A. & SHAH, N. P. (2007) Chemical analysis and sensory 
evaluation of Cheddar cheese produced with Lactobacillus acidophilus, Lb. 
casei, Lb. paracasei or Bifidobacterium sp. International Dairy Journal, 17, 
937-945. 
PASTORINO, A.J., HANSEN, C.L., and MCMAHON, D.J. (2003). Effect of salt on 
structure-function relationships of cheese. Journal of Dairy Science, 86 (1), 
60-69. 
PAYNE, M. R. & MORISON, K. R. (1999) A multi-component approach to salt 
diffusion in cheese. International Dairy Journal, 9, 887-894. 
PIRAINO, P., ZOTTA, T., RICCIARDI, A., MCSWEENEY, P. L. H. & PARENTE, 
E. (2008) Acid production, proteolysis, autolytic and inhibitory properties of 
lactic acid bacteria isolated from pasta filata cheeses: A multivariate screening 
study. International Dairy Journal, 18, 81-92. 
PORTER, M. C. (Ed.) (1990) Handbook of industrial membrane technology, New 
Jersey, Noyes Publications. 
POVEDA, J. M., S?NCHEZ-PALOMO, E., P?REZ-COELLO, M. S. & CABEZAS, 
L. (2008) Volatile composition, olfactometry profile and sensory evaluation of 
semi-hard Spanish goat cheeses. Dairy Science and Technology, 88, 355-367. 
PREETHA, S. & BOOPATHY, R. (1997) Purification and characterization of a milk 
clotting protease from Rhizomucor miehei. World Journal of Microbiology 
and Biotechnology, 13, 573-578. 
RASH, K. E. & KOSIKOWSKI, F. V. (1982) Influence of Lactic Acid Starter 
Bacteria on Enteropathogenic Escherichia coli in Ultrafiltration Prepared 
Camembert Cheese. Journal of Dairy Science, 65, 537-543. 
RIAHI, M. H., TRELEA, I. C., LECLERCQ-PERLAT, M. N., PICQUE, D. & 
CORRIEU, G. (2007) Model for changes in weight and dry matter during the 
ripening of a smear soft cheese under controlled temperature and relative 
humidity. International Dairy Journal, 17, 946-953. 
RICHOUX, R., AUBERT, L., ROSET, G. & KERJEAN, J. R. (2009) Impact of the 
proteolysis due to lactobacilli on the stretchability of Swiss-type cheese. Dairy 
Science and Technology, 89, 31-41. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  120 
ROGERS, N., DRAKE, M., DAUBERT, C., MCMAHON, D., BLETSCH, T. & 
FOEGEDING, E. (2009) The effect of aging on low-fat, reduced-fat, and full-
fat Cheddar cheese texture. Journal of Dairy Science, 92, 4756-4772. 
ROUPAS, P. (2001) On-farm practices and post farmgate processing parameters 
affecting composition of milk for cheesemaking. Australian Journal of Dairy 
Technology, 56, 219-232. 
ROUSH, D. J. & LU, Y. (2008) Advances in primary recovery: Centrifugation and 
membrane technology. Biotechnology Progress, 24, 488-495. 
RUEGG, M. & BLANC, B. (1976) Effects of pH onwater vapor sorption by caseins. 
Journal of Dairy Science, 59, 1019. 
SABOYA, L. V. & MAUBOIS, J. L. (2000) Current developments of microfiltration 
technology in the dairy industry. Le Lait, 80, 541-553. 
SCHLESSER, J. E., SCHMIDT, S. J. & SPECKMAN, R. (1992) Characterization of 
Chemical and Physical Changes in Camembert Cheese During Ripening. 
Journal of Dairy Science, 75, 1753-1760. 
SCHREIER, K., SCHAFROTH, K. & THOMET, A. (2010) Application of cross-flow 
microfiltration to semi-hard cheese production from milk retentates. 
Desalination, 250, 1091-1094. 
SCOTT, R. (1998) Cheesemaking Practice, Gaithersburg, Aspen Publishers, Inc. 
SHARMA, S. K., FERRIER, L. K. & HILL, A. R. (1989) Effect of Modified 
Manufacturing Parameters on the Quality of Cheddar Cheese made from 
Ultrafiltered (UF) Milk. Journal of Food Science, 54, 573-577. 
SHAW, M. B. (1981) The manufacture of soft, surface mould, ripened cheese in 
France with particular reference to Camembert. International Journal of Dairy 
Technology, 34, 131-138. 
SIMOV, Z. I. & IVANOV, G. Y. (2005) Growth and proteolytic activity of 
Lactobacillus delbrueckii ssp. bulgaricus and Streptococcus thermophilus in 
reduced sodium Kashkaval cheese. World Journal of Microbiology and 
Biotechnology, 21, 1285-1289. 
SIRKAR, K. K. (2000) Application of membrane technologies in the pharmaceutical 
industry. Current Opinion in Drug Discovery and Development, 3, 714-722. 
SMIT, G., SMIT, B. A. & ENGELS, W. J. M. (2005) Flavour formation by lactic acid 
bacteria and biochemical flavour profiling of cheese products. FEMS 
Microbiology Reviews, 29, 591-610. 
SU, J. and EVERETT, D.W. (2003). Adsorption of ?-casein onto native milk fat 
globule, latex particle, and emulsion surfaces. Food Hydrocolloids, 17 (4), 
529-537. 
SU?REZ, E., LOBO, A., ?LVAREZ, S., RIERA, F. A. & ?LVAREZ, R. (2006) 
Partial demineralization of whey and milk ultrafiltration permeate by 
nanofiltration at pilot-plant scale. Desalination, 198, 274-281. 
TARAKCI, Z. & TUNCTURK, Y. (2008) The effect of adjunct cultures on some 
chemical and biochemical properties of white-brined cheese. Journal of Food 
Biochemistry, 32, 490-505. 
THOMANN, S., SCHENKEL, P. & HINRICHS, J. (2008) The Impact of 
Homogenization and Microfiltration on Rennet-Induced Gel Formation. 
Journal of Texture Studies, 39, 326-344. 
TRUJILLO, A. J., GUAMIS, B., LAENCINA, J. & L?PEZ, M. B. (2000) Proteolytic 
activities of some milk clotting enzymes on ovine casein. Food Chemistry, 71, 
449-457. 
Ultra Filtration (UF) Process Development for the Production of Camembert Cheese 
Ming Ho Edwin Law  121 
VAHID, A., MOHAMMADI, T., PAK, A., KARBASSIAN, M. & GOLSHAN, M. 
(2006) Dimensional analysis of steady and unsteady state permeation flux for 
microfiltration of oil/water emulsion by a ceramic membrane. AIChE Annual 
Meeting, Conference Proceedings. 
VINAS, M. A. G., BALLESTEROS, C., MARTIN-ALVAREZ, P. J. & CABEZAS, 
L. (2007) Relationship between Sensory and Instrumental Measurements of 
Texture for Artisanal and Industrial Manchego Cheeses. Journal of Sensory 
Studies, 22, 462-476. 
WALSTRA, P., WOUTERS, J. T. M. & GEURTS, T. J. (2006) Dairy Science and 
Technology, Boca Raton, CRC Press, Taylor & Francis Group. 
WEIMER, B. C. (Ed.) (2007) Improving the flavour of cheese, Cambridge, Woodhead 
Publishing Limited. 
WIUM, H., GROSS, M. & QVIST, K. B. (1997) Uniaxial Compression of UF-Feta 
Cheese Related to Sensory Texture Analysis. Journal of Texture Studies, 28, 
445-476. 
YETI?MEYEN, A., GEN?ER, N., G?RSOY, A., DEVECI, O., ?ENEL, E., ?ANLI, 
E. & ?ZTEKIN, F. S. (2006) Protein properties of UF Teleme produced from 
various types of milk. Journal of Food and Drug Analysis, 14, 279-283. 
YORGUN, M. S., BALCIOGLU, I. A. & SAYGIN, O. (2008) Performance 
comparison of ultrafiltration, nanofiltration and reverse osmosis on whey 
treatment. Desalination, 229, 204-216. 
ZWIJNENBERG, H. J., KEMPERMAN, A. J. B., BOERRIGTER, M. E., LOTZ, M., 
DIJKSTERHUIS, J. F., POULSEN, P. E. & KOOPS, G. H. (2002) Native 
protein recovery from potato fruit juice by ultrafiltration. Desalination, 144, 
331-334. 
 
 
 APPENDICES 
APPENDIX 1 
Table 1: Descriptive statistics in cheese compositions (mean ? stdev) for each 
type of treatment. Significant differences are highlighted in bold (p ? 0.05). 
 
Cheese Treatments TS Content Salt Content 
Fat Content 
(%) 
Protein 
Content (%) 
Low-Fat 44.56 ? 0.94 1.74 ? 0.10 24.96 ? 0.92 15.09 ? 0.45 
High-Fat 44.64 ? 0.83 1.52 ? 0.25 28.03 ? 0.57 14.10 ? 0.58 
p-value 0.791 0.000 0.000 0.000 
Brine-Salted 44.82 ? 0.88 1.65 ? 0.23 26.62 ? 1.44 14.84 ? 0.72 
Retentate-Salted 44.36 ? 0.87 1.61 ? 0.22 26.37 ? 2.02 14.67 ? 0.65 
p-value 0.081 0.455 0.689 0.381 
pH 5.2 44.91 ? 0.69 1.74 ? 0.19 26.67 ? 1.46 14.86 ? 0.74 
pH 4.9 44.27 ? 0.98 1.52 ? 0.20 26.32 ? 2.00 14.66 ? 0.62 
p-value 0.013 0.000 0.575 0.321 
Tube Mould 44.56 ? 0.83 1.60 ? 0.25 26.40 ? 1.71 14.75 ? 0.70 
Small Mould 44.62 ? 0.97 1.66 ? 0.19 26.59 ? 1.80 14.76 ? 0.68 
p-value 0.820 0.311 0.762 0.941 
 
 
8642 8642
30
15
0
30
15
0
30
15
0
8642
30
15
0
8642
HB4.9S
Overall acceptance
N
u
m
b
e
r
 
o
f
 
p
a
n
e
l
l
i
s
t
s
HB4.9T HB5.2S HB5.2T
HR4.9S HR4.9T HR5.2S HR5.2T
LB4.9S LB4.9T LB5.2S LB5.2T
LR4.9S LR4.9T LR5.2S LR5.2T
 
Figure 1: The frequency of overall acceptance scores by sensory panellists for 16 
cheese treatments. 
 APPENDIX 2 
 
 
 
 
 
 
 
 
 
 
 
 
CHEESE GRADING 
 Collaborated cheese grading results 
  LB5.2T LB5.2T LB5.2S   LB5.2S   
APPEARANCE 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
bolding                         
cabbaging                   
concave                   
tanning                         
CUT                      
chalky                      
crunchy mould                      
core                     
dry rind                   
open texture                   
thick mould                        
TEXTURE                       
firm                        
good texture                   
runny                   
smooth                   
soft                         
FLAVOUR                      
bitter                        
bland                      
creamy                   
mushroom                   
papery rind                   
powdery                   
salty                    
sweet                   
sour                    
milky                   
ammonia                         
   LR5.2T   LR5.2T   LR5.2S   LR5.2S   
APPEARANCE 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
bolding                         
cabbaging                   
concave                   
tanning                         
CUT                    
chalky                        
crunchy mould                     
core                    
dry rind                   
open texture                   
thick mould                        
TEXTURE                        
firm                     
good texture                 
runny                 
smooth                 
soft                        
FLAVOUR                    
bitter                       
bland                  
creamy                    
mushroom                   
papery rind                  
powdery                  
salty                     
sweet                   
sour                  
milky                 
ammonia                         
 
 
   LB4.9T   LB4.9T   LB4.9S   LB4.9S     
APPEARANCE 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 Week 2 
Week 
4 
Week 
6 
bolding                         
cabbaging                     
concave                     
tanning                         
CUT                      
chalky                        
crunchy mould                     
core                     
dry rind                     
open texture                     
thick mould                         
TEXTURE                     
firm                         
good texture                  
runny                     
smooth                     
soft                        
FLAVOUR                      
bitter                       
bland                    
creamy                     
mushroom                   
papery rind                   
powdery                   
salty                     
sweet                   
sour                     
milky                  
ammonia                         
 
 
   LR4.9T   LR4.9T     LR4.9S   LR4.9S   
APPEARANCE 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
bolding                         
cabbaging                   
concave                    
tanning                        
CUT                         
chalky                   
crunchy mould                    
core                   
dry rind                   
open texture                    
thick mould                      
TEXTURE                         
firm                        
good texture                   
runny                   
smooth                   
soft                         
FLAVOUR                         
bitter                   
bland                   
creamy                    
mushroom                   
papery rind                   
powdery                   
salty                     
sweet                    
sour                    
milky                   
ammonia                         
 
 
   HB5.2T   HB5.2S   HR5.2T   HR5.2S   
APPEARANCE 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
bolding                         
cabbaging                      
concave                     
tanning                         
CUT                     
chalky                         
crunchy mould                      
core                     
dry rind                     
open texture                     
thick mould                        
TEXTURE                         
firm                     
good texture                     
runny                     
smooth                     
soft                         
FLAVOUR                     
bitter                         
bland                       
creamy                     
mushroom                     
papery rind                     
powdery                     
salty                       
sweet                     
sour                     
milky                      
ammonia                         
 
 
   HB4.9T   HB4.9S   HR4.9T   HR4.9S   
APPEARANCE 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
Week 
2 
Week 
4 
Week 
6 
bolding                         
cabbaging                    
concave                  
tanning                        
CUT                     
chalky                       
crunchy mould                    
core                   
dry rind                  
open texture                  
thick mould                        
TEXTURE                     
firm                  
good texture                      
runny                  
smooth                  
soft                       
FLAVOUR                     
bitter                       
bland                    
creamy                   
mushroom                  
papery rind                  
powdery                  
salty                        
sweet                  
sour                    
milky                     
ammonia                         
 
 Grading Sheet - White Mould Cheeses 
Date Graded                           
Grading Team                           
              
WEEK                         
No of Cheeses   1 2 3 4 5 6 7 8 9 10   
Cheese                         
Batch                         
BB Date                       Total % Defect 
Appearance Crooked or Lopsided                         
Max score Oversized/Undersized                         
8 Bare Mould & Uneven                         
  Too Moist or Wet                         
  Uneven Colour                         
  Brown Edges                         
  SCORE                         
Cut Dull                         
Max Score Holes                         
4 Under Ripe                         
  Over Ripe                         
  Rind                         
  SCORE                         
Texture Soft                         
Max Score Hard/Dry                         
12 Chalky                         
  SCORE                         
Flavour Bitter                         
Max Score Too Acidic                         
16 Ammonia                         
  Lacking Flavour                         
  Chemical Soapy                         
  Too Salty                         
  Metallic                         
  Rancid                         
  SCORE                         
Max Score = 40 OVERALL SCORE                         
 APPENDIX 3 
 
 
 
 
 
 
 
 
 
 
 
 
 
CONSUMER SENSORY EVALUATION 
 INFORMATION SHEET 
 
 
 
Ultrafiltration Process Development for the Production of 
Camemberti Cheese: Consumer Sensory Evaluation 
 
Researcher(s) Introduction 
Researchers 
Name: 
Edwin Law Supervisors 
Name: 
Dr Tony Mutukumira 
Dr Marie Wong 
Contact Details: 021418911 Contact Details: 094140800 ext 41203 
                   ext 41204 
 
You are invited to take part in a consumer sensory evaluation to assist a project based 
on Camembert cheese. 
 
Your participation in this activity will take approximately 10 minutes.  
 
We are selecting people for this exercise who meet the following criteria:  
General consumer 
 
The foods you will be tasting contain the following components that can be harmful 
or cause allergic reactions with certain groups of people. You are requested not to 
partake if you may be adversely affected by the following 
? Alcohol in small quantities 
? Aspartame 
? Quinine 
? Milk and milk derivatives 
? Added Sulphites  
 
People who are pregnant are also requested not to participate in this sensory 
evaluation. 
 
The information collected in this study will be used to complete an assignment in 
partial fulfilment of the Master of Technology in Food Technology. No data linked to 
an individual?s identity will be collected. 
 
If you have any questions about this work, please contact one of the people indicated 
above. 
 
This project has been reviewed and approved by the Massey University Human Ethics Committee, 
PN Protocol HEC: PN Protocol 03/34).  If you have any concerns about the conduct of this 
research, please contact Professor Sylvia V Rumball, Chair, Massey University Campus Human 
Ethics Committee: Palmerston North, telephone 06 350 5249, email S.V.Rumball@massey.ac.nz. 
 CONSENT FORM  
 
Ultrafiltration Process Development for the Production of Camemberti Cheese: 
Consumer Sensory Evaluation 
 
CONSENT FORM 
 
THIS CONSENT FORM WILL BE HELD FOR 12 MONTHS FROM DATE OF SIGNING 
(For minors aged 8-15 consent form to be signed by parent or guardian) 
 
? I have read and understood the Information Sheet and have had the details of 
the study explained to me.  My questions have been answered to my 
satisfaction, and I understand that I may ask further questions at any time. 
 
? I agree to voluntarily participate in this study under the conditions set out in 
the Information Sheet. 
 
? I understand I have the right to withdraw from the study at any time and to 
decline to answer any particular questions. 
 
? I have advised and discussed with the Researcher any potentially relevant 
cultural, religious or ethical beliefs that may prevent me from consuming the 
Foods under consideration. 
 
Participants Signature:  Date:  
 
Full Name - printed  
 
 Sensory Evaluation of Camembert Cheese      
     
 
Instructions 
 
? FIVE samples are presented. 
 
? Please rate one sample at a time from left to right WITHOUT 
COMPARING them. 
 
1. First, observe the APPEARANCE of each sample. Indicate your 
degree of likeness/dislikeness by placing an X in the square. 
 
2. Then smell the sample. Indicate your degree of likeness/dislikeness 
for ODOUR. 
 
3. Then take a bite of the sample and chew three times. 
 
4. Indicate your degree of likeness/dislikeness for FLAVOUR in terms of 
aroma and taste (sweet, sour, bitter, salty, umami). 
 
5. Indicate your degree of likeness/dislikeness for TEXTURE in terms of 
mouth-feel/structure (firmness, consistency etc.) 
 
6. Finally, rate your OVERALL ACCEPTANCE of the sample. 
 
 
SAMPLE 1 Code: 187 
Attributes 
Dislike      Dislike       Dislike       Dislike  Neither like     Like          Like           Like           Like 
extremely very much moderately slightly   nor dislike    slightly  moderately very much extremely 
Appearance 
                                                         
Odour 
                                                         
Flavour 
                                                         
Texture 
                                                         
Overall 
Acceptance                                                          
Comments 
(if applicable) 
 
 
 
 
 
 
Please turn over >> 
  
SAMPLE 2 Code: 142 
Attributes 
Dislike      Dislike       Dislike       Dislike  Neither like     Like          Like           Like           Like 
extremely very much moderately slightly   nor dislike    slightly  moderately very much extremely 
Appearance 
                                                         
Odour 
                                                         
Flavour 
                                                         
Texture 
                                                         
Overall 
Acceptance                                                          
Comments 
(if applicable) 
 
 
 
 
 
 
 
 
SAMPLE 3 Code: 313 
Attributes 
Dislike      Dislike       Dislike       Dislike  Neither like     Like          Like           Like           Like 
extremely very much moderately slightly   nor dislike    slightly  moderately very much extremely 
Appearance 
                                                         
Odour 
                                                         
Flavour 
                                                         
Texture 
                                                         
Overall 
Acceptance                                                          
Comments 
(if applicable) 
 
 
 
 
 
 
 
Please turn over >> 
  
SAMPLE 4 Code: 521 
Attributes 
Dislike      Dislike       Dislike       Dislike  Neither like     Like          Like           Like           Like 
extremely very much moderately slightly   nor dislike    slightly  moderately very much extremely 
Appearance 
                                                         
Odour 
                                                         
Flavour 
                                                         
Texture 
                                                         
Overall 
Acceptance                                                          
Comments 
(if applicable) 
 
 
 
 
 
 
 
 
SAMPLE 5 Code: 851 
Attributes 
Dislike      Dislike       Dislike       Dislike  Neither like     Like          Like           Like           Like 
extremely very much moderately slightly   nor dislike    slightly  moderately very much extremely 
Appearance 
                                                         
Odour 
                                                         
Flavour 
                                                         
Texture 
                                                         
Overall 
Acceptance                                                          
Comments 
(if applicable) 
 
 
 
 
 
 
 
THE END 
Thank you for your participation! 
Hope you?ve enjoyed the cheese! 
 
 
 APPENDIX 4 
 
 
 
 
 
 
 
 
 
 
 
 
ADDITIONAL INFORMATION 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  
 
 
DELVO-TEC? TS-30 
DIRECT-SET? Range 
 
 
Description and Composition 
DIRECT-SET? highly concentrated, free-flowing, deep-frozen pelletised (DSF) lactic acid bacteria starters, 
ready to be added straight to the vat. 
 
The DELVO-TEC? TS-30 DSF DIRECT-SET? range consists of defined strains of Streptococcus 
thermophilus.  
 
The TS-30 cultures have proven to be phage robust. Two bacteriophage unrelated cultures are included in 
the range: 
DELVO-TEC? TS-30A 
DELVO-TEC? TS-30B 
 
These cultures were not genetically modified according to the European Directive 90/220/CEE.  
They are certified as Kosher. 
 
Acidification performance  
 
Test conditions    Substrate: 
Temperature: 
Dosage: 
70?C - RSM 9.5%-solids 
38?C 
2 Units/1000 litres 
Performance Time: 
pH: 
5.0 hours 
5.5 ? 0.5 
 
Acidification curves 
       
      
DELVO-TEC
?
 TS-30 Series
Acidification of 2U / 1000 litres in 9.5% Reconstituted Skim Milk 
5.0
5.5
6.0
6.5
7.0
0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00 4:30 5:00 5:30 6:00
Time (hours)
pH
35?C
37?C
39?C
42?C
s
 
 
NB. The curves above are obtained in experimental laboratory conditions and are only shown for illustration purpose. 
Product Sheet
 
 
Applications  
 
Examples Indicative Dosage Possibly together with 
Various stabilised soft cheeses 1-2 Units /1000 L various DELVO-ADD? cultures 
 
Microbiological properties 
 
Microorganisms Specifications 
Coliforms 
Escherichia coli 
Salmonella 
Listeria monocytogenes 
Staphylococcus aureus (coagulase+) 
Yeasts & Moulds 
Enterococci 
Non lactic acid bacteria 
Absent in 1 g 
Absent in 25 g 
Absent in 25 g 
Absent in 25 g 
Absent in 1 g 
< 10 in 1 g 
< 10 in 1 g 
< 500 in 1 g 
 
Distribution and Storage 
The DELVO-TEC? TS-30 DSF cultures are packed in metallised laminated sachets in standard pack sizes 
and shipped in polystyrene boxes with dry.  
Minimum Order: 300 units 
Storage Conditions: below -45?C. The shelf life is 9 months at -45?C from the date of manufacture.  
The ?Best Before? date is printed on each pack. 
 
Available Pack Sizes Name 
2 units 5 units 10 units 
All DELVO-TEC? TS-30 DSF cultures  < < < 
 
Technical Service 
Our Technical Service Teams can help you in selecting the most suitable starter solutions for your 
application by designing, supervising and discussing the results of laboratory and plant trials.  
 
 
Quality Assurance 
? DSM Food Specialties has production facilities worldwide, which comply with ISO 9001:2000 
requirements, and employ cGMP and HACCP throughout. 
? DELVO-TEC? products are manufactured to meet IDF-standards and can be provided with a certificate 
of conformity and specification sheets. 
 
For further information please contact:      or our local representative: 
DSM Food Specialties 
Dairy Ingredients 
P.O. Box 1 2600 MA Delft  
The Netherlands 
Tel.: +31-15-2792355 
Fax: +31-15-2793200 
E-mail: info.dairy-ingredients@dsm.com 
www.dsm-dairy.com 
To the best of our knowledge, the information contained herein is accurate and complete. 
 However, nothing herein contained shall be construed to imply any warranty or guarantee". 
 
DELVO-ADD? 100-X DSF  
DIRECT-SET? Range 
 
 
 
Description and Composition 
DIRECT-SET? highly concentrated, free-flowing, deep-frozen pelletised (DSF) lactic acid bacteria starter, 
ready to be added straight to the vat. 
 
DELVO-ADD? 100-X DSF is an undefined mixed culture of: 
   Lactococcus lactis ssp lactis  
   Lactococcus lactis ssp cremoris  
   Lactococcus lactis ssp lactis biovar diacetylactis 
   Leuconostoc sp. 
 
This culture was not genetically modified according to the European Directive 90/220/CEE.  
It is certified as Kosher. 
 
 
Unit value 
1 unit = 150 g 
 
 
Technological properties and benefits  
 
DELVO-ADD? 100-X DSF is suitable flavour adjunct in various heterofermentative fermented milk and 
cheese applications, for flavour enhancement.  
In cheese applications DELVO-ADD? 100-X DSF may also enhance openness depending on the cheese 
making process and the specific cheese characteristics. 
 
Even though this culture is not primarily used for acidification, the graph below illustrates its acidification 
properties.  
 
 
    
Acidification DELVO-ADD? 100-X DSF
2Units / 1000 litres in 9.5%-solids Reconstituted Skim Milk
4.0
4.5
5.0
5.5
6.0
6.5
7.0
0 2 4 6 8 10 12 14 16 18 20
Time (hours)
pH
 20?C
 30?C
 25?C
 
 
 
NB. The curves above are obtained in experimental laboratory conditions and are only shown for illustration purpose. 
Product Sheet 
Applications  
 
Examples Indicative Dosage Possibly together with 
Sour cream, buttermilk 0.5-1.0 unit / 1000 L 
Fresh cheeses, e.g. quark 0.5-1.0 unit / 1000 L 
Edam, Gouda 0.1-0.3 unit / 1000 L 
Cheddar (Mild, Mature, etc) 0.1-0.3 unit / 1000 L* 
Various DELVO-TEC?, other 
DELVO-ADD? cultures, 
and/or LAFTI? probiotics 
* possibly in combination with a cooking temperature of 40?C 
 
Microbiological properties 
 
Microorganisms Specifications DSF 
Coliforms 
Escherichia coli 
Salmonella 
Listeria monocytogenes 
Staphylococcus aureus (coagulase+) 
Yeasts & Moulds 
Enterococci 
Non lactic acid bacteria 
Absent in 1 g 
Absent in 25 g 
Absent in 25 g 
Absent in 25 g 
Absent in 1 g 
< 10 in 1 g 
< 10 in 1 g 
< 500 in 1 g 
 
Distribution and Storage 
DELVO-ADD? 100-X DSF is packed in metallised laminated sachets in standard pack sizes and shipped in 
polystyrene boxes with dry ice. 
Minimum Order: 300 units 
Storage Conditions: below -45?C. The shelf life is 9 months at -45?C from the date of manufacture.  
The ?Best Before? date is printed on each pack. 
 
Available Pack Sizes Name 
1 unit 5 units 10 units 
DELVO-ADD? 100-X DSF    
 
Technical Service 
Our Technical Service Teams can help you in selecting the most suitable starter solutions for your application 
by designing, supervising and discussing the results of laboratory and plant trials.  
 
Quality Assurance 
? DSM Food Specialties has production facilities worldwide, which comply with ISO 9001:2000 
requirements, and employ cGMP and HACCP throughout. 
? DELVO-ADD? products are manufactured to meet IDF-standards and can be provided with a certificate 
of conformity and specification sheets. 
 
For further information please contact:      or our local representative: 
DSM Food Specialties 
Dairy Ingredients 
P.O. Box 1 2600 MA Delft  
The Netherlands 
Tel.: +31-15-2792355 
Fax: +31-15-2793200 
E-mail: info.dairy-ingredients@dsm.com 
www.dsm-dairy.com 
 
"To the best of our knowledge, the information contained herein is accurate and complete. 
 However, nothing herein contained shall be construed to imply any warranty or guarantee". 
 
Cps dadd 100-X dsf 0505e1 


  
 
Fromase? XL 
Extra thermolabile coagulant for cheese 
 
 
Introduction 
Fromase? XL is an extra thermolabile variant of the 
microbial coagulant derived from the soil fungus 
Rhizomucor miehei. 
In most cases the standard Fromase? TL meets all 
the requirements of the cheesemaker. In a number 
of cases however and for various reasons a 
product with an increased thermolability is needed.  
Such reasons include for instance the need to use 
lower whey pasteurisation temperatures when the 
whey is used for whey protein manufacturing or the 
need to be able to destroy residual clotting activity 
in slightly acidic whey. 
In those cases the normal characteristics of 
Fromase? TL may not be the right ones to satisfy 
the technological requirements. 
Research into increased thermolability led to the 
development of Fromase? XL. 
 
The increased thermolability of Fromase? XL has 
also been shown to be an advantage in the 
production of some cheese varieties such as 
Emmentaler, where the enzyme is already partly 
inactivated during the cooking of the curds. This has 
been shown to lead to a better controlled eye 
formation. Other varieties of cheeses have also 
been produced with good results using  
Fromase?  XL. 
 
 
Technical characteristics 
Fromase? XL is an acid fungal protease with a 
molecular weight of approx. 40,000. 
The overall amino acid composition of  
Fromase? XL is very similar to that of animal rennet. 
Its stability in solution is excellent between pH 3.0 
and 6.5. The specificity of Fromase? XL on the beta 
chain of insulin is very similar to that of animal 
rennet, with a preference for those linkages 
involving aromatic amino acids. 
Like animal rennet, Fromase? XL induces milk 
clotting through hydrolytic cleavage of the 
phenylananine-methionine linkage of kappa casein. 
 
 
1. Effect of temperature  
The relative activity of animal rennet, Fromase? TL 
and Fromase? XL is given in graph 1. 
GRAPH 1 : TEMPERATURE STABILITY 
% Relative activity
! Calf rennet # Fromase? TL (Fromase? XL
0
50
100
150
35 40 45 50 55
TEMP [?C]
#
!
60
(
 
Fromase? XL has the same effect as calf rennet and 
Fromase? TL in the usual clotting temperature range. 
At higher temperatures the effect of the enzymes 
show marked differences. 
The thermolability of Fromase? XL is clearly superior 
to that of calf rennet. Fromase? XL is, like calf 
rennet, partly inactivated during scalding. 
 
 
2. Effect of pH 
Graph 2 shows that Fromase? XL is somewhat more 
sensitive with reference to the pH than Fromase? TL 
or calf rennet in the important pH range for cheese 
manufacturing. 
 
 
GRAPH 2: pH DEPENDANCE
% relative activity
! Calf rennet # Fromase? TL (Fromase? XL
0
100
200
6.20 6.35 6.50 pH 6.65
#
150
50
(
!
 
 
3.  Effect of calcium ion concentration 
Fromase? XL and calf rennet have the same 
sensitivity to the calcium ion concentration, whereas 
Fromase? TL is slightly more sensitive. These 
studies were carried out using Berridge substrate, 
the concentration of 10 mM of CaCl2 corresponds to 
1.11 g of anhydrous CaCl2 per liter. 
Technical Bulletin 
 
 GRAPH 3: EFFECT OF CALCIUM  ION CONCENTRATION
% relative activity
! Calf rennet # Fromase? TL (Fromase? XL
0
300
500
2 4 8 mM CaCl2 10
400
100
(
!
6
200
#
 
 
4.  Flocculation times and enzyme concentration 
Animal rennet is known to follow the law of Storch 
and Segelke: 
 
 
In which: 
t     = flocculation time 
[E]  = enzyme concentration 
C1  = constant 
C2  = constant depending on the type of milk used  
        and the type of rennet. 
 
This is also the case for Fromase? XL and TL.  
However, when raw milk is used, Fromase? XL and 
TL do not follow this law and the deviation observed 
varies with different types of milk. It has been shown 
that all the Rhizomucor miehei enzymes deviate from 
the law of Storch and Segelke, corresponding in fact 
to competitive inhibition by inhibitors present in raw 
milk that are usually destroyed at temperatures 
above 68?C. This characteristic is especially evident 
in the manufacture of Emmental cheese from raw 
milk. Due to this inhibition the addition rates of 
Fromase? XL and TL in raw milk have to be 
increased about 20-25% compared to calf rennet to 
obtain a coagulation time of 25-30 minutes. 
 
5. Inactivation of Fromase? XL in whey 
It is essential to ensure that Fromase? is inactivated 
in whey. Graph 4 shows the activity of Fromase? XL 
and TL with reference to the whey pH. 
 
As long as storage acidification has been avoided 
before pasteurisation, Fromase? TL will be 
inactivated by normal whey pasteurisation conditions 
(15 sec./73?C). Fromase? XL is even superior to calf 
rennet in this respect, when Fromase? XL is added 
directly to reconstituted whey at pH 6.0, less than 
1% residual activity is found after 15 sec. 
pasteurization at 68?C.  
To achieve the same degree of inactivation of calf 
rennet or bovine rennet the temperature must be 2?C 
higher i.e 15 sec. at 70?C. 
GRAPH 4:             INACTIVATION IN WHEY, PASTEURISATION:  15 sec. / 73?C
! Calf rennet # Fromase? TL (Fromase? XL
% relative activity
0
50
100
75
25
(
!
4.80 5.20 6.00pH5.60
#
 
 
Graph 5 shows the influence of pH and temperature 
on the residual activity of Fromase? XL in whey after 
pasteurization. 
 
 
GRAPH 5: pH AND TEMPERATURE EFFECT ON RESIDUAL ACTIVITY 
IN Fromase? XL
TEMP [?C]
0
65
75
5.20 5.40 5.60 pH 6.00
70
60
5.80
< 1 % residual activity
< 10% residual activity
 
 
 
Method of Analysis. 
The activity of Fromase? is expressed in 
International Milk Clotting Units (IMCU) in accordance 
with the Relative Milk Clotting Activity Test 
(REMCAT) of the International Dairy Federation 
(IDF). In the case of Fromase? the IDF method 
176:1996 applies. A copy of the method can be 
obtained from the IDF: 41, Square Vergote, 1030 
Brussels, Belgium or request at your local DSM Dairy 
Ingredients Office 
 
For further information please contact:  
DSM Food Specialties 
Dairy Ingredients 
P.O. Box 1 2600 MA Delft  
The Netherlands 
Tel.: +31-15-2792355   
Fax: +31-15-2793200 
E-mail: info.dairy-ingredients@dsm.com 
www.dsm.com/dfs/dairy 
 
or our local representative: 
?To the best of our knowledge, the information contained herein is accurate and complete.  
However, nothing herein contained shall be contrued to imply any warranty or guarantee.? 
 
EtbfmaseXL0001e1 
 
 
t  =  C  1 * 
1  
[ E  ]  +  C  2