Journal of Dairy Science
Volume 89, Issue 7 , Pages 2386-2396, July 2006

Effect of Formulation and Manufacturing Parameters on Process Cheese Food Functionality—I. Trisodium Citrate

Minnesota–South Dakota Dairy Foods Research Center, Department of Food Science and Nutrition, University of Minnesota, St. Paul 55108

Received 2 November 2005; accepted 5 February 2006.

Article Outline

Abstract 

The objective of this research was to use a Rapid Visco Analyzer to study the effect of natural cheese age, trisodium citrate (TSC) concentration, and mixing speed on process cheese food (PCF) functionality. In this study 3 replicates of natural cheese were manufactured, and a portion of each cheese was subjected to 6 different PCF manufacturing treatments at 2, 4, 6, 12, and 18wk of ripening. These treatments were factorial combinations of 3 levels of TSC (i.e., 2.0, 2.5, and 3.0%) and 2 mixing speeds during manufacture (450 and 1,050rpm). Functional properties of the PCF evaluated included manufacturing properties [apparent viscosity after manufacture (VAM)], unmelted textural properties (firmness), melted cheese flow properties [hot apparent viscosity (HAV)], and cheese thickening during cooling [time at 5000cP (T5)]. All 4 parameters (VAM, firmness, HAV, and T5) were significantly affected by natural cheese age and mixing speed, whereas VAM, HAV, and T5 were also significantly influenced by the amount of TSC. The VAM and firmness decreased as cheese age increased, whereas T5 values increased as cheese age increased. Similarly, VAM, HAV, and firmness values increased because of the increased mixing speed, whereas T5 values decreased. The age×mixing speed interaction was significant for VAM and firmness. The age×concentration of the TSC interaction term was significant for VAM, whereas the age×age×TSC concentration term was significant for HAV. The results demonstrate that natural cheese age, mixing speed during manufacture, and concentration of TSC have a significant impact on process cheese functionality.

Key words: process cheese, functionality, unmelted texture, emulsifying salts

 

Back to Article Outline

Introduction 

Process cheese is used as an ingredient in many applications, such as a slice on a burger or a dip for snacks. Consequently, it is important to control process cheese functional properties (i.e., texture, melt, and flow) that are inherent to its proper functioning as an ingredient. The functional properties of process cheese are determined by the ingredients used in the formulation (i.e., type of natural cheese, age of natural cheese, amount of natural cheese, type and amount of emulsifying salt) as well as processing conditions (i.e., cooking temperature, cooking time, and mixing speed during manufacture).

The effect of natural cheese characteristics on process cheese functionality is largely attributed to the level of intact casein present in the natural cheese (Berger et al., 1998). Intact casein is the amount of casein that has not been hydrolyzed by enzymes during ripening. It is generally expressed in terms of relative casein content, a ratio of the amount of unhydrolyzed casein nitrogen to the total nitrogen in the cheese (Berger et al., 1998). During ripening, the protein present in natural cheese is progressively hydrolyzed, thereby decreasing the intact casein content of the natural cheese. Intact casein forms the structural network of process cheese, and a high level of intact casein indicates the potential for extensive protein–protein and protein–fat interactions (Berger et al., 1998). Consequently, process cheese made from excessively ripened natural cheese (i.e., with lower intact casein levels) is soft, whereas process cheese made from unripened cheese (i.e., with a higher level of intact casein) is very firm (Templeton and Sommer, 1930).

In addition to natural cheese age, other important formulation parameters that influence process cheese functionality include the type and amount of emulsifying salt (Thomas, 1973). The purpose of the emulsifying salt is to chelate calcium from the natural cheese paracasein network, thereby solubilizing the protein. Subsequently, solubilized caseins are free to interact with the lipid and water phase, which results in the formation of an emulsion (Caric et al., 1985). Emulsifying salts have different calcium-chelating properties that influence the final emulsion properties (Gupta et al., 1984). The Code of Federal Regulations (U.S. Department of Health and Human Services, 2003) identifies 13 emulsifying salts that can be used in making process cheese products. However, most process cheese products are manufactured with trisodium citrate (TSC), disodium phosphate (DSP), or a combination of both. Disodium phosphate is primarily used in loaf applications that require storage at room temperature, whereas TSC is primarily used in slice applications in which the process cheese is cooled with a chill belt. Numerous studies have been conducted to compare and contrast the effects of TSC and DSP on process cheese texture and melt properties (Caric et al., 1985; Cavalier-Salou and Cheftel, 1991; Sutheerawattananonda and Bastian, 1998). In general, process cheese made with DSP is characterized by larger fat pockets, a grainy texture, and a weaker body, as compared with process cheese made with TSC (Templeton and Sommer, 1936; Rayan et al., 1980). Additionally, the amount of emulsifying salt has been reported to influence process cheese flow properties (Templeton and Sommer, 1936). However, more recent experiments have reported that no significant changes in the body and textural properties result from an increase in TSC concentration from 2 to 4% (Thomas et al., 1980).

In addition to formulation parameters, the temperature and mixing conditions used during manufacture also influence the functional properties of process cheese. Previous researchers have demonstrated that fat particle size distribution is affected by the mixing speed and temperature during manufacture (Rayan et al., 1980; Glenn et al., 2003; Lee et al., 2003). Under a constant cooking time and temperature, a higher mixing speed results in a larger number of small, evenly distributed fat globules, as compared with a low mixing speed (Lee et al., 2003). It is theorized that a large number of small, evenly distributed fat globules enhance fat–protein and protein–water interactions and thus facilitate the formation of a stronger network (Drake, 1973; Lee et al., 2003). However, increasing the mixing speed beyond certain limits will enhance fat–protein and protein–protein interactions to an extent that the casein molecules coagulate into a “pudding-like” structure. This phenomenon is generally referred to as “overcreaming” and is indicated by an increase in viscosity and reduction in meltability (Drake, 1973; Berger et al., 1998; Glenn et al., 2003).

Although many studies have been conducted on the effects of formulation and processing parameters on process cheese quality, these studies have been limited and have typically evaluated the effect of 1 or 2 parameters owing to limitations imposed by manufacturing techniques, raw material requirements, or time constraints. Moreover, it is difficult to choose a particular cooker design to simulate the effect of all the potential manufacturing parameters (Berger et al., 1998). Consequently, a small-scale manufacturing system, with the ability to vary the manufacturing profile, would be desirable for studying the effect of various parameters on process cheese functionality.

Recently, an instrument called the Rapid Visco Analyzer (RVA; Newport Scientific Pvt. Ltd., Warriewood, Australia) has been used to manufacture process cheese on a small scale (Metzger et al., 2002; Kapoor et al., 2004; Kapoor and Metzger, 2005). The RVA is a computer-integrated manufacturing instrument that can be used to determine the apparent viscosity of food products during manufacture using a sample size of approximately 25g. Its versatility lies in its ability to vary cooking time, cooking temperature, and mixing speed. In a study by Kapoor et al. (2004), significant differences in textural and melt properties were reported between process cheese samples manufactured on a pilot plant scale and the RVA. The differences between the samples were attributed to manufacturing and sampling techniques. Appropriate modifications in the sampling and manufacturing techniques were reported in a subsequent study (Kapoor and Metzger, 2005). Furthermore, Lai et al. (2000) have suggested that the RVA could be used to characterize the rheological properties of non-Newtonian fluids. Subsequently, researchers (Rosenberg and Metzger, 2003; Biswas et al., 2005; Prow and Metzger, 2005) have demonstrated that the RVA can be used to manufacture process cheese and subsequently analyze its melt properties.

The objective of this study was to determine the effect of the age of natural cheese, the amount of emulsifying salt, and the mixing speed during manufacture on process cheese food (PCF) functionality.

Back to Article Outline

Materials and Methods 

Experimental Design and Statistical Analysis 

A split-plot design was used to study the effect of natural cheese age, concentration of TSC, and mixing speed during manufacture on PCF functionality. Three replicates of natural cheese were manufactured, and a portion of each cheese was subjected to 6 different treatments at 2, 4, 6, 12, and 18wk of ripening. These treatments were factorial combinations of 3 levels of TSC (2.0, 2.5, and 3.0%) and 2 mixing speeds during manufacture (450 and 1,050rpm). Five replicates of PCF were manufactured for each treatment. Statistical analysis was performed by treating natural Cheddar cheese replicates as blocks, age as the whole-plot factor, and 6 different treatments as split-plot factors.

Age of natural cheese was considered a continuous variable and was analyzed for linear effects, quadratic effects, and deviations from the quadratic effects. Multicollinearity of the age terms (continuous variables) was reduced by using “mean-centered” data. Mean-centering of continuous variables involves subtracting the mean from each of the individual time points during ripening (Glantz and Slinker, 2001). Furthermore, separate pairwise comparison tests were performed, using least significance difference (LSD), for each functional property at each time point across all treatments, and also for each treatment across all time points. Macanova 4.12 (School of Statistics, University of Minnesota, Minneapolis) was used to perform all statistical analyses. The level of significance was P<0.05 throughout the paper.

PCF Formulations 

A blend was prepared by mixing all the ingredients, except TSC, in a Blentech twin-screw process cheese cooker (Blentech Corporation, Rohnert Park, CA) at 120rpm. The blend formulation developed with each of the 3 natural cheese replicates is shown in Table 1. Each blend formulation was developed using Techwizard, an Excel-based formulation program provided by OWL Software (Lancaster, PA). The formulation program was used to balance the salt, moisture, and fat for each replicate to 1.8, 44.0, and 25.0%, respectively. The 3 replicates of natural cheese used for this study were manufactured at the University of Minnesota pilot plant facility and were analyzed for fat using the Mojonnier method (Atherton and Newlander, 1977), for moisture content using a forced-draft oven method (Bradley and Vanderwarn, 2001), for salt content using a Corning 926 chloride analyzer (Corning Glass Works, Medfield, MA), and for pH using a Corning pH/ion meter model 450 (Corning Glass Works) fitted with a Thermo Orion combination pH probe (Thermo Electron Corporation, Louisville, CO). Additionally, the intact casein of the natural cheese at 2, 4, 6, 12, and 18wk of ripening was determined by measuring the amount of pH 4.6 soluble nitrogen and subtracting it from the total nitrogen (Bynum and Barbano, 1985). Other ingredients in the formulation were nonfat dry milk (low heat; Dairy America, Fresno, CA), sodium chloride salt (Cargill Inc., MN), anhydrous butter oil (MidAmerica Farms, Springfield, MO), trisodium citrate duohydrate (Archer Daniels Midland Company, Decatur, IL), and water.

Table 1. Blend formulations used to manufacture process cheese food
IngredientReplicate 1Replicate 2Replicate 3
Cheddar cheese, %78.9979.0975.51
Water, %14.4014.4015.05
Skim milk powder (nonfat dry milk), %5.395.398.70
Salt, %0.820.770.74
Butter oil, %0.400.350.00

PCF Manufacture 

The PCF were manufactured with an RVA as described by Kapoor and Metzger (2005). Samples were prepared in RVA canisters according to the blend formulations in Table 1, and TSC was added at rates of 0.5, 0.625, and 0.75g into 24.5, 24.325, and 24.25g of process cheese blend, respectively. This corresponded to 2.0, 2.5, and 3.0% TSC, respectively. An additional 0.5g of water was added to each RVA canister to compensate for moisture loss during manufacture. The RVA canister was held in a water bath at 40°C for 7min prior to manufacture. The manufacturing profile included holding the RVA canister at 85°C during the entire period of manufacture and subjecting the contents to 1 of 2 mixing speeds (450rpm for low and 1,050rpm for high) for 6min, and then a mixing speed of 160rpm for 1min. The 450 and 1,050rpm mixing speeds used during the first 6min of manufacture were chosen based on previous research, which concluded that a mixing speed of 450rpm in the RVA produced process cheese that was similar to process cheese produced in a twin-screw pilot-scale process cheese cooker operating at 140rpm (Kapoor and Metzger, 2005). Consequently, the 450rpm mixing speed represents a typical mixing speed used to manufacture process cheese, whereas 1,050rpm represents a higher than normal mixing speed. Each mixing speed and TSC combination was manufactured 5 times. Immediately after manufacture, the PCF samples were transferred to electroplated copper cylinders (20mm in diameter and 30mm in height). The cylinders were wrapped in Reynolds food service film (Nogg Chemicals and Paper, Omaha, NE) and were stored in a refrigerator (4°C) for 18h prior to further analysis. A preliminary study (data not shown) was conducted to ensure that there was no difference in the moisture content of the PCF manufactured using the various mixing speeds and TSC concentrations utilized in this study. The pH of each process cheese was determined 24h after manufacture and was between 5.65 and 5.75 for all formulations at all ripening times (data not shown).

PCF Functionality Analysis 

The functional properties evaluated included manufacturing properties [apparent viscosity after manufacture (VAM)], unmelted textural properties (firmness), melted cheese flow properties [hot apparent viscosity (HAV)], and cheese thickening during cooling [time at 5,000cP (T5)].

Apparent viscosity after manufacture values for all PCF samples were determined during manufacture in the RVA as described by Kapoor and Metzger (2005). These values are a measure of process cheese apparent viscosity immediately after manufacture and were determined at 160rpm, which corresponds to a shear rate of 53.6s−1. For unmelted texture analysis, the PCF samples were removed from the electroplated copper cylinders (20mm diameter) and cut to a height of 20mm, and texture profile analysis was performed at 4°C. The texture profile analysis was performed using a TA.XT2 texture analyzer (Texture Technologies Corp., Scarsdale, NY/Stable Microsystems, Godalming, UK). A uniaxial single-bite compression was performed with a 50-mm cylindrical flat probe (TA-25) at 80% compression with a cross-head speed of 0.8 mm/s. For the HAV and T5 analyses, the PCF samples were ground to a particle size of 2 to 3mm and analyzed with the RVA melt test, as described by Kapoor and Metzger (2005). In the RVA melt test, a 14-g sample of the ground process cheese was weighed into an RVA canister, along with 1g of propylene glycol. Subsequently, the RVA melt test was performed by raising the temperature of the canister from 25°C to a peak temperature of 85°C in 5min, holding for 3min at the peak temperature, and finally cooling to 25°C in 6min. The stirring speed was held at 0rpm for 30s, 20rpm for 30s, 100rpm for 1min, and 300rpm for the remainder of the test. The minimum apparent viscosity during the holding period and the time required to reach an apparent viscosity of 5,000cP during the cooling period were collected from the apparent viscosity vs. time curve. These data points collected during the RVA melt analysis are referred to as HAV and T5, respectively, and were determined at a mixing speed of 300rpm, which corresponds to a shear rate of 100.5s−1.

Back to Article Outline

Results and Discussion 

Natural Cheese Intact Casein 

Changes in the level of intact casein during ripening, for each replicate of natural cheese, are shown in Figure 1. The intact casein content in the natural cheese decreased with an increase in age for all replicates of process cheese that were manufactured. A large decrease in intact casein was observed between 2 and 12wk of ripening, but remained stable between 12 and 18wk of ripening. Intact casein is predominantly (80%) made up of αs1 and β fractions of casein (1:1 ratio). During ripening, the casein fractions are hydrolyzed, with a half-lives of 2 and 37wk, respectively (De Jong, 1976; Basch et al., 1989). Because αs1 has a half-life of 2wk, most of this fraction is hydrolyzed within the first 6 to 12wk of ripening, and its hydrolysis is associated with the large decrease in the intact casein content of the natural cheese during early ripening. Although the intact casein content continues to decrease in later ripening, the decrease is small compared with the first several weeks.

VAM 

The change in VAM as a function of ripening time is shown in Figure 2. Apparent viscosity after manufacture was significantly (P<0.05) influenced by age, mixing speed, TSC concentration, and the interaction terms age×age, age×mixing speed, and age×TSC concentration (Table 2). The mean values and LSD for each treatment, at each ripening point, are shown in Table 3.

  • View full-size image.
  • Figure 2. 

    Effect of mixing speed (low: ——; high: – – – –) and trisodium citrate concentration (2.0%: ●; 2.5%: ■; 3.0%: ▴) on process cheese food viscosity after manufacture (VAM) during ripening of natural cheese.

Table 2. Mean squares and P values (in parentheses) of melted and unmelted functional properties for process cheese food1
ComponentdfVAMFirmnessHAVT5
Whole plot
Replicate (Rep)29.8×105 (0.131)910.27 (0.24)29,113* (0.02)0.010 (0.80)
Age (A)13.42×107* (<0.01)31,122* (<0.01)57,341* (<0.01)2.40* (<0.01)
Age×age (A2)13.34×106* (0.02)983.71 (0.21)105,490* (<0.01)0.31* (0.02)
Deviations from quadratic (Ads)22.46×105 (0.54)81.54 (0.85)3,663.9 (0.45)0.018 (0.67)
Whole-plot error83.74×105530.244,117.10.042
Split plot
Mixing speed (MS)18.65×106* (<0.01)15,830* (<0.01)91,278* (<0.01)0.894* (<0.01)
A×MS12.91×105* (0.01)1,426.5* (<0.01)256.07 (0.61)0.005 (0.30)
A2×MS119,110 (0.52)59.40 (0.11)292.19 (0.59)0.002 (0.54)
Ads×MS21.65×105* (0.03)4.167 (0.83)733.32 (0.48)0.001 (0.82)
TSC concentration (C)26.51×105* (<0.01)12.47 (0.58)60,706* (<0.01)0.065* (<0.01)
A×C24.08×105* (<0.01)13.455 (0.56)1,671.9 (0.19)0.001 (0.71)
A2×C268,359 (0.23)20.126 (0.42)3,359.8* (0.04)0.007 (0.19)
Ads×C417,233 (0.82)10.753 (0.76)146.58 (0.96)0.001 (0.90)
MS×C28,136.3 (0.83)75.57* (0.04)584.9 (0.56)0.002 (0.61)
Ads×MS×C832,526 (0.671)13.77 (0.77)548.59 (0.80)0.002 (0.92)
Error5045,08422.72985.90.004

1Model equation: Y = Rep + A + A2 + Ads + E (Rep. Ads) + R + A×R + A2×R + Ads×R + C + A×C + A2×C + Ads×C + R×C + Ads×R×C. VAM = viscosity after manufacture; HAV = hot apparent viscosity; T5 = time at 5,000cP; TSC = trisodium citrate.

*Statistically significant (P<0.05).

Table 3. Mean (n = 3) melted and unmelted functional properties of process cheese food manufactured at 3 levels (2.0, 2.5, and 3.0%) of trisodium citrate (TSC) and 2 mixing speeds1
Ripening pointLow mixing speedHigh mixing speed
2.0%2.5%3.0%2.0%2.5%3.0%
——— (VAM, cP) ———
Week 25,127A,ab4,758A,b4,549A,b5,805A,a5,119A,ab4,986A,ab
Week 44,144B,ab3,690Bb3,681B,b4,544B,a4,490B,a4,239B,ab
Week 63,774B,ab3,472BC,b3,553B,b4,151BC,a3,929C,ab3,927B,ab
Week 122,933C,b2,878C,b2,877C,b3,718C,a3,747C,a3,810B,a
Week 182,997Cb2,928Cb2,774Cb3,667C,a3,624C,a3,679B,a
——— (Firmness, N) ———
Week 295A,b102A,b104A,b141A,a135A,a142A,a
Week 486A,b88AB,b89AB,b123A,a119AB,a117B,a
Week 670B,b74BC,b73BC,b98B,a97BC,a95C,a
Week 1263B,b68C,ab67Cb85BC,a85C,a85C,a
Week 1860B,a61C,a58C,a76C,a77C,a78C,a
——— (HAV, cP) ———
Week 2608A,b657A,bc715A,ab685A,bc703A,ab781A,a
Week 4542B,c572AB,bc592B,bc589B,bc624AB,ab661B,a
Week 6508BC,a559B,ab579B,ab577B,ab634AB,a650BC,a
Week 12474Cb553B,ab594B,a563B,a594B,a619C,a
Week 18513BC,d595AB,c630AB,bc590B,b653AB,b721AB,a
——— (T5, min) ———
Week 211.26B,a11.25C,a11.12C,ab11.05D,bc11.02D,bc10.92D,c
Week 411.51AB,a11.49B,ab11.44B,abc11.31C,bc11.25C,c11.26C,c
Week 611.64A,a11.55AB,b11.54AB,b11.40BC,c11.38B,c11.35BC,c
Week 1211.74A,a11.66A,ab11.59AB,abc11.48AB,bc11.44AB,c11.43AB,c
Week 1811.71A,a11.68A,a11.63A,ab11.56A,ab11.52AB,ab11.45A,b

a–dMeans within the same row not sharing a common superscript are different (P<0.05).

A–DMeans within the same column (separate for each functional property) not sharing a common superscript are different (P<0.05).

1VAM = viscosity after manufacture; HAV = hot apparent viscosity; T5 = time at 5,000cP.

The significance (P<0.05) of the age term indicates that there was a change in VAM as a function of the ripening period, whereas the significance (P<0.05) of the interaction term age×age indicates that these changes were not linear. As shown in Figure 2, all treatments had a substantial decrease in VAM during the first 12wk of ripening; however, there were no significant (P>0.05) differences between the mean values for each treatment at 12 and 18wk of ripening (Table 3). The trends in VAM (Figure 2) are similar to trends in intact casein values (Figure 1), which indicate that changes in VAM may be related to intact casein.

Previous researchers have demonstrated that the viscosity of process cheese, at the end of manufacture, is largely determined by protein–protein interactions and protein–fat interactions (Lee et al., 2003). However, the extent of protein–protein and protein–fat interactions in process cheese is dependent on the characteristics of the protein (mainly casein) that forms the structure and framework of process cheese (Guinee et al., 2004). The presence of intact (unhydrolyzed) casein results in extensive protein–protein and protein–fat interactions and a fibrous casein network, whereas the presence of hydrolyzed casein results in weaker protein–protein and protein–fat interactions and a nonfibrous casein network (Taneya et al., 1980). Because young natural cheese has a larger amount of intact casein as compared with a mature natural cheese, process cheese manufactured with young natural cheese should have more extensive protein–protein and protein–fat interactions, and consequently a higher viscosity at the end of manufacture as compared with process cheese manufactured with a mature natural cheese.

The significance (P<0.05) of the mixing speed term indicates that there were differences in VAM between high and low mixing speed treatments; however, the significance (P<0.05) of the age×mixing speed interaction term indicates that the differences in VAM between high and low mixing speeds were not constant across all ripening periods (Figure 2). As shown in Table 3, the high mixing speed treatment had significantly (P<0.05) higher VAM values at each TSC level, as compared with the low mixing speed treatment at 12 and 18wk of ripening. However, the differences were not significant (P>0.05) at 2, 4, and 6wk of ripening. These results indicate that a high mixing speed influences VAM when matured natural cheese (>6wk) is used to manufacture PCF, but is less critical when young natural cheese is used.

No previous research describes the effect of mixing speed on process cheese viscosity immediately after manufacture. However, previous researchers have used empirical and rheological melt tests to describe the effect of mixing speed on process cheese flow characteristics during melting (Taneya et al., 1980; Glenn et al., 2003; Lee et al., 2003). It is plausible that the results for flow characteristics during melting may be related to VAM; therefore, we believe the underlying intermolecular interactions governing these functional properties to be similar. Moreover, a study by Kapoor and Metzger (2005) demonstrated a correlation (R2 = 0.89) between the apparent viscosity of process cheese immediately after manufacture and the apparent viscosity of process cheese during melting; both were measured using the RVA. Therefore, the results found by previous researchers for the flow characteristics during melting may be relevant to viscosity immediately after manufacture and will be included in the discussion.

A study by Glenn et al. (2003) demonstrated that during manufacture, an increase in mixing speed from 50 to 150rpm, at a constant temperature, reduces the melt characteristics of the process cheese. The process cheese formulation in their study used 75% young Cheddar cheese and 14% medium Cheddar cheese; the manufacturing conditions included 3 different temperatures (74, 80, and 86°C) and 3 mixing speeds (50, 100, and 150rpm). However, the exact age of the young and medium Cheddar, and their intact casein contents, were not reported. These results are similar to the results obtained in our study for process cheese manufactured using mature natural cheese (>6wk), in which a significant (P<0.05) difference was observed between the high and low mixing speeds. Previous researchers have demonstrated that in some process cheese formulations the protein structure is deformable when a high mixing speed is used during manufacture (Taneya et al., 1980). The deformation at high mixing speeds leads to a decrease in fat globule size and a more homogeneous fat globule distribution within the protein structural network (Lee et al., 2003). The reduction in fat globule size and increase in homogeneity of fat globule distribution is thought to increase the viscosity of process cheese at the end of manufacture for 2 reasons: 1) It allows more proteins to be absorbed at the fat–water interface, and 2) it strengthens the protein matrix by including fat globules in the protein matrix (Glenn et al., 2003; Lee et al., 2003).

The significance (P<0.05) of the TSC concentration term indicates that there were differences in VAM between various levels of TSC, and the significance (P<0.05) of the age×TSC concentration term indicates that these differences were not constant across all periods of ripening (Figure 2). Additionally, the differences between each TSC level, at each ripening point and mixing speed, were not significant (P>0.05; Table 3). However, at 2, 4, and 6wk of ripening, the lowest level of TSC (within each ripening point and mixing speed) had the highest VAM values. These results indicate that the level of TSC may have some effect on VAM when young natural cheese (<12wk) is used in a formulation.

The function of TSC in a process cheese formulation is to chelate calcium from the natural cheese paracaseinate complex by disrupting the calcium bridges that facilitate solubilization and dispersion of the casein (Meyer, 1973; Guinee et al., 2004). A high level of TSC in a formulation may lead to an extensive disintegration of the paracaseinate complex (Meyer, 1973). Conversely, a lower level of TSC in a formulation may leave a portion of the paracasein complex aggregated, which, upon heating and mixing, may result in extensive protein–protein interactions and a fibrous protein network, leading to a more viscous product. Previous researchers have reported that the amount of calcium bound to the proteins in natural cheese decreases significantly after the first 4wk of ripening (Hassan et al., 2004). It is plausible that after 4wk of ripening, even the lowest level of TSC may cause complete disintegration of the natural cheese paracaseinate complex, whereas at less than 4wk of ripening, the lowest level of TSC may not cause complete disintegration of the paracaseinate complex. Consequently, TSC concentration may play an important role in controlling the VAM when young natural cheese (<4wk) is used in a process cheese formulation.

Firmness 

The change in firmness as a function of ripening time is shown in Figure 3. The mean values and LSD for each treatment, at each ripening point, are shown in Table 3. Firmness was significantly (P<0.05) influenced by the age of natural cheese, mixing speed during manufacture, and the interaction terms age×mixing speed and mixing speed×TSC concentration.

  • View full-size image.
  • Figure 3. 

    Effect of mixing speed (low: ——; high: – – – –) and trisodium citrate concentration (2.0%: ●; 2.5%: ■; 3.0%: ▴) on process cheese food firmness during ripening of natural cheese.

As shown in Figure 3, all treatments exhibited a substantial decrease in firmness during the first 6wk of ripening. Overall, there were no significant (P>0.05) differences in firmness among the mean values for each treatment at 6, 12, and 18wk of ripening; however, the PCF samples manufactured with a combination of high mixing speed and 2% TSC showed significant differences in firmness between 6 and 18wk of ripening (Table 3).

A decrease in process cheese firmness, owing to an increase in the age of the natural cheese used to manufacture process cheese, has been associated with the decreasing level of intact casein in natural cheese during ripening (Templeton and Sommer, 1930; Vakaleris et al., 1962; Piska and Stetina, 2003). Young natural cheese with a high level of intact casein is typically used to produce process cheese when a firm body is desired (Templeton and Sommer, 1930; Palmer and Sly, 1943; Berger et al., 1998). In addition to our study, other researchers (Vakaleris et al., 1962) have shown a substantial decrease in process cheese firmness (measured with a penetrometer) as the age of the natural cheese increases. This decrease in firmness may not be related solely to intact casein, but could be linked to the rapid hydrolysis of αs1-casein during early ripening. Previous researchers have observed a decrease in the firmness of process cheese with the hydrolysis of αs1-casein in natural cheese (Tamime et al., 1990; Acharya and Mistry, 2005).

The significance (P<0.05) of the mixing speed term indicates differences in firmness between the high and low mixing speed treatments; however, the significance (P<0.05) of the age×mixing speed interaction term indicates that these differences in firmness (between the high and low mixing speeds) were not constant across all ripening periods (Figure 2). Moreover, the significance (P<0.05) of the mixing speed×TSC concentration interaction term suggests that the differences in firmness, as a result of varying concentrations of TSC, were influenced by mixing speed. As shown in Table 3, the high mixing speed treatments had significantly (P<0.05) higher firmness values at each TSC level, as compared with the low mixing speed treatments at 2, 4, 6, and 12wk of ripening. However, at 18wk of ripening the differences were not significant (P>0.05), even though the PCF manufactured with high mixing speeds were substantially firmer than the PCF manufactured at the low mixing speeds. Previous researchers have reported similar effects of mixing speed on imitation process cheese, where an increase in firmness was observed with an increase in the mixing speed (Norohna et al., 2005).

The effect of mixing speed on firmness could be due to the impact of mixing speed on fat globule size and distribution in the protein matrix, as well as to protein–protein interactions (Aguilera and Kessler, 1989; Lee et al., 2003). As discussed in the VAM section, using a high mixing speed during manufacture results in small homogeneous fat globules within the protein matrix. Upon cooling of process cheese after manufacture, these small and homogeneous fat globules further strengthen the protein matrix by acting as fillers (Aguilera and Kessler, 1989; Fox et al., 2000). These fat globules will also crystallize within the protein matrix during cooling and may contribute to process cheese firmness (Fox et al., 2000). Moreover, a higher mixing speed could result in a more ordered protein network with an increased level of protein–protein interactions, and upon cooling, may result in a firmer product (Aguilera and Kinsella, 1991; Zhong et al., 2004).

As shown in Table 2, the mixing speed×TSC concentration interaction term was significant (P<0.05), indicating that changes in firmness with respect to TSC concentration for both mixing speeds were dissimilar, although the differences among the various TSC concentrations were not significantly different (P>0.05).

HAV 

Hot apparent viscosity is a measure of the apparent viscosity of melted process cheese. The change in HAV as a function of ripening time is shown in Figure 4. The mean values and LSD for each treatment, at each ripening point, are shown in Table 3. Hot apparent viscosity was significantly (P<0.05) influenced by age, mixing speed, TSC concentration, and the interaction terms age×age, age×mixing speed, and age×age×TSC concentration (Table 2).

  • View full-size image.
  • Figure 4. 

    Effect of mixing speed (low: ——; high: – – – –) and trisodium citrate concentration (2.0%: ●; 2.5%: ■; 3.0%: ▴) on process cheese food hot apparent viscosity (HAV) during ripening of natural cheese.

The significance (P<0.05) of the age term indicates that there was a change in HAV as a function of the ripening period, whereas the significance (P<0.05) of the interaction term age×age indicates that these changes were not linear. As shown in Figure 2, all treatments had a substantial decrease in HAV during the first 12wk of ripening. The trends in HAV between 2 and 12wk of ripening (Figure 4) are similar to trends in VAM (Figure 2). As discussed in the VAM section, the decrease in HAV during ripening could be due to the hydrolysis of the intact casein in the natural cheese during ripening, which in turn results in process cheese with weaker protein–protein and protein–fat interactions, and lower melted viscosity.

It is interesting to note the substantial increase in HAV from 12 to 18wk of ripening (Table 3), although the differences between each treatment at 12 and 18wk of ripening were not significantly different (P>0.05), except for PCF manufactured with a combination of 3.0% TSC and a high mixing speed. The increase in viscosity from 12 to 18wk of ripening may be due to the phenomenon of overcreaming. Overcreaming occurs when process cheese develops into a thick custard-like texture, and is a consequence of the increased solubility and water-binding capacity of protein (Berger et al., 1998; Glenn et al., 2003). However, our results demonstrate that the phenomenon of overcreaming could also be influenced by mixing speed and TSC concentration.

The significance (P<0.05) of the mixing speed term indicates that there were differences in HAV between the high and low mixing speed treatments; however, the significance (P<0.05) of the age×mixing speed interaction term indicates that the differences in HAV between the high and low mixing speeds were not constant across all ripening periods (Figure 2). As shown in Table 3, the high mixing speed treatments had substantially higher HAV values than the low mixing speed treatments at each TSC level and ripening period. Between 2 and 12wk of ripening, the difference in HAV between the low and high mixing speed was not significant (P>0.05). However, at 18wk the HAV with the high mixing speed was significantly (P<0.05) higher than that at the low mixing speed for all TSC concentrations. Again, these results indicate that the effect of mixing speed is magnified in process cheese made from aged (>18wk) natural cheese. Previous researchers have reported a similar effect of mixing speed on imitation process cheese, in which a decrease in meltability was observed with an increase in the mixing speed during manufacture (Norohna et al., 2005). As mentioned previously, a high HAV is an indication of decreased meltability. As discussed in the VAM and firmness sections, the effect of mixing speed on HAV could be due to the effect of mixing speed on fat globule size and distribution within the protein matrix, as well as to protein–protein and protein–fat interactions.

The significance (P<0.05) of the TSC concentration term indicates that there were differences in the HAV between various levels of TSC, and the significance (P<0.05) of the age×age×TSC concentration term indicates that these differences were not constant across all periods of ripening (Figure 2). However, the mean HAV at each TSC level, at each ripening point and mixing speed, were not significantly different (P>0.05; Table 3). It is interesting to note that the lowest level of TSC had the lowest HAV values within each mixing speed at all ripening times. Previous researchers have demonstrated similar results, in which a decrease in the meltability was observed with an increase in TSC concentration (Templeton and Sommer, 1932; Cavalier-Salou and Cheftel, 1991).

The trends with TSC concentration for HAV are the opposite of the trends reported for VAM. The VAM is a measure of the apparent viscosity of process cheese immediately after manufacture, whereas HAV measures the apparent viscosity of the cheese after it has cooled and solidified and then been melted in the RVA melt test. The observed differences between VAM and HAV may be due to the effect of cooling on the microstructure of process cheese. During cooling, the structure of process cheese is further modified and strengthened by protein–protein and protein–fat interactions (Fox et al., 2000; Zhong et al., 2004).

T5 

Typically, it is desirable to have process cheese that takes a longer time to thicken during cooling when it is used as a dip for snacks, and a shorter thickening time during cooling when it is used as a slice on burgers. Consequently, postmelt characteristics (e.g., thickening during cooling), as measured using T5 values, can be an important assessment of the functional properties of process cheese. The T5 values measured during the RVA melt test measure the time required for a melted process cheese to thicken to an apparent viscosity of 5,000cP during cooling. The change in T5 values as a function of ripening time is shown in Figure 5. The mean values and LSD for each treatment, at each ripening point, are shown in Table 3. The T5 values were significantly (P<0.05) influenced by age, mixing speed, TSC concentration, and the interaction term age×age (Table 2).

  • View full-size image.
  • Figure 5. 

    Effect of mixing speed (low: ——; high: – – – –) and trisodium citrate concentration (2.0%: ●; 2.5%: ■; 3.0%: ▴) on process cheese food time at 5,000cP (T5) during ripening of natural cheese.

The significance (P<0.05) of the age term indicates that there was a change in the T5 values as a function of the ripening period, whereas the significance (P<0.05) of the interaction term age×age indicates that these changes were not linear. As shown in Figure 5, all treatments had a substantial increase in T5 values during the first 6wk of ripening. This indicates that the process cheese required a longer time to thicken during cooling, which is related to the age of the natural cheese that was used to manufacture the process cheese. However, there were no significant (P>0.05) differences between the mean values for each treatment at 12 and 18wk of ripening (Table 3).

The increasing trend in T5 values with age could again be due to the hydrolysis of casein in natural cheese during ripening. When process cheese is manufactured with young cheese, it is more likely to have longer protein strands and more protein–protein interactions than process cheese manufactured with ripened cheese (Berger et al., 1998). An increase in protein–protein interactions should result in a cheese that thickens faster during cooling. This may explain why process cheese manufactured with a young cheese thickens faster during cooling as compared with process cheese made with ripened cheese. However, the increase in T5 values was not linear. The T5 values increased faster during the first 6wk of ripening as compared with 6 to 18wk of ripening. The trends in T5 are similar to firmness, in which a large decrease was observed during the first 6wk of ripening. A larger increase in T5 values during initial ripening may again be related to degradation of certain casein fractions, especially αs1-casein, during early ripening.

The significance (P<0.05) of the mixing speed term indicates that there were differences in T5 values between high and low mixing speed treatments, where increasing the mixing speed significantly (P<0.05) lowered T5 values. Moreover, the amount of emulsifying salt also had a significant (P<0.05) impact on T5 values, with higher TSC concentrations resulting in lower T5 values. The reasons for these differences could be the same as explained in the VAM and firmness section, in which process cheese manufactured with high mixing speeds or high emulsifying salt concentrations is characterized by small and homogeneous fat globules and a more ordered protein network (Lee et al., 2003; Zhong et al., 2004). Upon melting, these small and homogeneous fat globules may aid in faster thickening of the protein matrix during cooling as compared with the large and heterogeneous fat globules observed when a low mixing speed or low concentration of emulsifying salt is used.

Back to Article Outline

Conclusions 

This study demonstrated that the flow properties of process cheese immediately after manufacture, unmelted textural properties, flow properties of process cheese during melting, and postmelt characteristics are all influenced by the formulation and processing parameters used during process cheese manufacture. A change in the age of natural cheese used to manufacture process cheese, and a change in mixing speed during manufacture resulted in a change in all the above-mentioned functional properties of process cheese. Additionally, a change in the TSC concentration influenced only the melt and flow characteristics, but not the unmelted texture properties. Consequently, it is important to control these formulation and processing parameters simultaneously to obtain a product with the desired texture and melt properties.

Back to Article Outline

Acknowledgments 

The authors would like to thank the Midwest Dairy Association Inc. (St. Paul, MN) for funding this work.

Back to Article Outline

Supplementary data 

Download file

download text

Interpretive summary.

Back to Article Outline

References 

  1. Acharya MR, Mistry VV. Effect of vacuum-condensed or ultrafiltered milk on pasteurized process cheese. J. Dairy Sci. 2005;88:3037–3043
  2. Aguilera JM, Kessler HG. Properties of mixed and filled type dairy gels. J. Food Sci. 1989;54:1213–1217
  3. Aguilera JM, Kinsella JE. Compression strength of dairy gels and microstructural interpretation. J. Food Sci. 1991;56:1224–1228
  4. Atherton HV, Newlander JA. Tests for fat: Babcock, Gerber and Mojonnier. Chemistry and Testing of Dairy Products. 4th ed.. Westport, CT: AVI Publ.; 1977;Pages 71–116
  5. Basch JJ, Farrell HM, Walsh RA, Konstance RP, Kumosinski TF. Development of quantitative model for enzyme catalyzed, time-dependent changes in protein composition of Cheddar cheese during storage. J. Dairy Sci. 1989;72:591–603
  6. Biswas AC, Kapoor R, Metzger LE, Muthukumarrapan K. Evaluation of process cheese food functionality using various melt tests. J. Dairy Sci. 2005;88(Suppl. 1):149;(Abstr.)
  7. Berger W, Klostmeyer H, Merkenich K, Uhlmann G. Process Cheese Manufacture—A JOHA Guide. Germany: BK Guilini Chemie Ladenberg; 1998;
  8. Bradley RL, Vanderwarn MA. Determination of moisture in cheese and cheese products. J. AOAC. 2001;84:570–592
  9. Bynum DG, Barbano DM. Whole milk reverse osmosis retentates for cheddar cheese manufacture: Chemical changes during aging. J. Dairy Sci. 1985;68:1–10
  10. Caric M, Gantar M, Kalab M. Effects of emulsifying agents on the microstructure and other characteristics of process cheese—A review. Food Microstruct. 1985;4:297–312
  11. Cavalier-Salou C, Cheftel JC. Emulsifying salts influence on characteristics of cheese analogs from calcium caseinate. J. Food Sci. 1991;56:1542–1551
  12. De Jong L. Protein breakdown in soft cheese and its relation to consistency. 1. Proteolysis and consistency of ‘Noordhollandse Meshanger’ cheese. Neth. Milk Dairy J. 1976;30:242–253
  13. Drake P. Rheological properties of fat containing foods. In:  Solms J editors. Functional Properties of Fats in Foods. Switzerland: Forster Zurich; 1973;p. 171–192
  14. Fox PF, Guinee TP, Cogan TM, McSweeney PLH. Fundamentals of Cheese Science. Gaithersburg, MD: Aspen Publ.; 2000;
  15. Glantz SA, Slinker BK. Primer of Applied Regression and Analysis of Variance. 2nd ed.. New York, NY: McGraw-Hill; 2001;
  16. Glenn TA, Daubert CR, Farkas BE, Stefanski LA. A statistical analysis of creaming variables impacting process cheese melt quality. J. Food Qual. 2003;26:299–321
  17. Guinee TP, Caric M, Kalab M. Pasteurized processed cheese and substitute/imitation cheese products. In:  Fox PF,  McSweeney P,  Cogan T,  Guinee T editor. Cheese: Chemistry, Physics and Microbiology. 2:Boston, MA: Elsevier Academic; 2004;p. 349–399
  18. Gupta SK, Karahdian C, Lindsay RC. Effect of emulsifier salts on textural and flavor properties of processed cheeses. J. Dairy Sci. 1984;67:764–778
  19. Hassan A, Johnson ME, Lucey JA. Changes in the proportions of soluble calcium and insoluble calcium during the ripening of cheddar cheese. J. Dairy Sci. 2004;87:854–862
  20. Kapoor R, Lethola P, Metzger LE. Comparison of pilot-scale and rapid visco analyzer process cheese manufacture. J. Dairy Sci. 2004;87:2813–2821
  21. Kapoor R, Metzger LE. Small-scale manufacture of process cheese using a Rapid Visco Analyzer. J. Dairy Sci. 2005;88:3382–3391
  22. Lai KP, Steffe JF, Ng PKW. Average shear rates in Rapid Visco Analyzer (RVA) mixing system. Cereal Chem. 2000;77:714–716
  23. Lee SK, Buwalda RJ, Euston SR, Foegeding EA, McKenna AB. Changes in the rheology and microstructure of process cheese during cooking. Lebensm. Wiss. Technol. 2003;36:339–345
  24. Metzger LE, Kapoor R, Rosenberg LA, Upreti P. RVA: Process cheese manufacture. Aust. J. Dairy Technol. 2002;57:136
  25. Meyer A. Processed Cheese Manufacture. London, UK: Food Trade Press; 1973;
  26. Norohna N, O’Riordan D, O'Sullivan M. Influence of processing parameters on the physicochemical properties of imitation cheese. New Orleans, LA: Abstract 31-6 in Book of Abstracts, 2005 IFT Annual Meeting and Food Expo; 2005;
  27. Palmer HJ, Sly WH. Oil separation in processed cheese. Dairy Ind. 1943;8:427–431
  28. Piska I, Stetina J. Influence of cheese ripening and rate of cooling of the processed cheese mixure and rheological properties of processed cheese. J. Food Eng. 2003;61:551–555
  29. Prow LA, Metzger LE. Melt analysis of process cheese spread or product using a rapid visco analyzer. J. Dairy Sci. 2005;88:1277–1287
  30. Rayan AA, Kalab M, Ernstrom CA. Microstructure of process cheese. Scan. Electron Microsc. 1980;3:635–643
  31. Rosenberg LA, Metzger LE. Comparison of the melting properties of process cheese using a Rapid Visco Analyzer (RVA) and Schreiber melt test. J. Dairy Sci. 2003;86(Suppl. 1):40;(Abstr.)
  32. Sutheerawattananonda M, Bastian ED. Monitoring process cheese meltability using dynamic stress rheometry. J. Texture Stud. 1998;29:169–183
  33. Tamime AY, Kalab M, Davies G, Younis MF. Microstructure and firmness of processed cheese manufactured from cheddar cheese and skim milk powder cheese base. Food Struct. 1990;9:23–37
  34. Taneya S, Kimura T, Izutsu T, Buchhiem W. The submicroscopic structure of processed cheese with different melting properties. Milchwissenschaft. 1980;35:479–481
  35. Templeton HL, Sommer HH. Some observations on processed cheese. J. Dairy Sci. 1930;13:203–220
  36. Templeton HL, Sommer HH. Factors affecting the body and texture of processed cheese. J. Dairy Sci. 1932;15:29–41
  37. Templeton HL, Sommer HH. Studies on emulsifying salts used in processed cheese. J. Dairy Sci. 1936;19:561–572
  38. Thomas MA. The Manufacture of Processed Cheese—Scientific Principles. 1st ed.. Australia: New South Wales Department of Agriculture Richmond; 1973;
  39. Thomas MA, Newell G, Abad GA, Turner AD. Effect of emulsifying salts on objective and subjective properties of processed cheese. J. Food Sci. 1980;45:458–466
  40. U.S. Department of Health and Human Services. Code of Federal Regulations. Washington, DC: U.S. Department of Health and Human Services; 2003;Section 133.169.
  41. Vakaleris DG, Olson NF, Price WV. Effects of proteolysis of natural cheese on body and melting properties of pasteurized process cheese spread. J. Dairy Sci. 1962;45:492–494
  42. Zhong Q, Daubert CR, Velev OD. Cooling effects on a model rennet casein gel system: Part I. Rheological characterization. Langmuir. 2004;20:7399–7405

PII: S0022-0302(06)72311-6

doi:10.3168/jds.S0022-0302(06)72311-6

Journal of Dairy Science
Volume 89, Issue 7 , Pages 2386-2396, July 2006

Article Tools

* Image Usage & Resolution