Designing ice cream and gelato mixes and how to undertake mix calculations
Details
Written by Michael Mullan
This section is aimed at product development technologists who need to know how to design ice cream or gelato mixes. It should also be of interest to small scale ice cream or gelato manufacturers who want to replace commercial ice cream mixes (bought in mixes) or additives by sourcing their own ingredients and additives. The article provides a basic introduction to the scientific principles involved in producing ice cream or gelato recipes using key product components e.g. fat, solids-not-fat (SNF).
Manufacturers produce ice cream or gelato to meet the requirements of consumers.
In many situations ice cream manufacturers will have a final product specification to meet. An ice cream recipe must be produced from this specification. One element of this specification will be the chemical composition of the final product including any legal compositional requirements e.g. for fat, milk protein and total solids. Details of some standards or advisory guidelines for the composition of ice cream are given here.
Since the composition of ice cream is subject to regulation, and regulations are subject to frequent review, manufacturers are advised to consult appropriate regulatory authorities in their jurisdiction to ensure legislative compliance. A compositional specification will typically specify the fat, milk solids not fat (MSNF), sugar, emulsifier and stabiliser concentrations in the final product. In this situation the manufacturer will select ingredients that can supply the above components, blend these and then process to produce a finished ice cream. Calculations will be required to ensure that the ingredients are correctly formulated to meet the final product specification.
In other situations the manufacturer may want to produce a new or improved product and will have to devise their own product compositional specification. This specification must be developed so that the fat, sugar and MSNF components are balanced.
Fat /Sugar balance
Fat adds certain taste and other qualities to food. In ice cream fat is balanced by sugar and there is a relationship between fat concentration and sugar concentration.
The relationship between fat and sugar concentrations is also influenced by the type of freezer used (Table 1).
TABLE 1: Recommended fat and sugar concentrations for ice cream mixes using various types of freezer
Freezer Type
Fat concentration (% w/w)
Sugar concentration (% w/w)
Vertical
6
12
Vertical
7
12.5
Vertical
8
13
Horizontal
9
13.5
Horizontal
10
14
Continuous
10
14
Continuous
12
15
From Rothwell (1985)
Note the fat/sugar ratios above should be regarded as approximate and will need to be modified to meet differing consumer requirements. The market has also changed since Rothwell's work in 1985 and some of the more popular types of dairy ice cream may contain high concentrations of fat. Also, sweeteners other than sucrose have become more extensively used. However, from a product development perspective the information in table 1 provides a useful start.
Guinard et al. (1996) studying the influence of fat and sugar in a vanilla ice cream used university students to taste and rate on a nine-point hedonic scale the texture and mouth feel, flavour (taste and odour), and overall degree of liking (DOL) for nine samples of vanilla ice cream varying in sugar (8 to 18 %, w/w) and fat (10 to 18 %, w/w) concentration.
The hedonic ratings differed significantly among samples, and the best-liked sample for texture and mouth feel, flavour, and overall degree of liking contained 13.54 % sugar and 14.99 % fat. Response surface methodology, in simple terms a three dimensional graphical analysis of overall degree of liking versus sugar and fat concentrations, was used to relate hedonic ratings to sugar and fat percentages in the ice cream. Dome-shaped response surfaces,e.g. figure 1, were obtained for all three degree of liking parameters, and optimal sugar and fat, respectively, were 13.16 % and 14.02 % for degree of liking of texture and mouth feel, 14.07 % and 15.35 % for degree of liking of flavour, and 14.30 % and 14.77 % for overall degree of liking.
The results from figure 1 show clearly that particular sugar concentrations are required to obtain maximum overall DOL ratings for particular fat concentrations and show the value of this simple research tool.
The Hedonic scale method developed by Peryam and Pilgrim (1957) uses the descriptors defined below. Surprisingly, Guinard et al. (1996) did not report values towards the higher end of the scale probably limiting the utility of their work. However, the authors did emphasise that the range and values of the mean hedonic ratings in this trail were surprisingly small and low, respectively.
Overall DOL ranged from 3.8 to 6.5 with differences between men and women; the mean DOL was 5.7 for men and 5.5 for women on the nine-point scale. The lowest rating, 3.8 indicates "Dislike Moderately" whereas the highest rating 6.5 indicates “like slightly”. The authors provided the equations for the RSM curves. Each curves was described by a multiple regression model that included linear, quadratic, and cross product effects. The equation for overall DOL was :
DOL= 1.76 X_{1} + 0.49 X^{2}- 0.06 X_{1}^{2} - 0.01 X_{2}^{2} - 10.05 . Where X_{1} is the sugar concentration (%) and X_{2} is the fat concentration (%). This author has found this regression equation useful (although limited) and readers may find the calculator programmed using it to be of interest.
The author has also provided a simple tool based mainly on James Rothwell's workthat enables the calculation of the approximate sucrose (sugar) concentration required to balance a particular fat value. This works fairly well at lower fat concentrations but tends to produce slightly sweeter mixes at higher levels particularly at above 12% fat (w/w). Nevertheless it is a useful tool for product development.
Some research has been published on the optimal fat and sugar concentrations required for maximum acceptability of ice cream, table 2. Surprisingly, perhaps, there appears to be agreement in that fat concentrations around 14% and a sugar concentration of around 15% give the most acceptable ice cream.
TABLE 2. Optimal fat and sugar levels required for maximum acceptability of ice-cream
Country
Fat (% w/w)
Sugar (% w/w)
Reference
Egypt
14
15
Salam et al . (1981)
US
14.3
14.77
Guinard et al. (1996)
More information on sweetness is given in the article on the relative sweetness of ice cream. Note. Much of the original work done on sweetness used sucrose as the sweetener. For product development purposes consider balancing the 'sugar' concentration discussed previously with the equivalent concentration of sweetener or sweeteners determined using their relative sweetness values.
MSNF / Water balance
Ice cream is a complex colloidal system and includes ice crystals in a concentrated unfrozen aqueous phase. This aqueous phase contains a concentrated lactose solution along with other sweeteners and salts.
When the MSNF concentration is optimal the ice cream has a smooth body, a characteristic ice cream taste and will have a satisfactory shelf life.
At high MSNF levels the lactose concentration in the unfrozen aqueous phase may be so high that the lactose comes out of solution and that crystals of lactose grow. This may result in ‘sandiness' in the final product. This is not usually a problem if skim milk powder (SMP) is used but could be an issue of whey powder has been used to replace part of the SMP for economic reasons. Tharp and Young (2013) suggest that sandiness can be avoided if the total lactose in a mix is <7% or <11% based on the weight of the lactose concentration in the mix. Assuming a 100 g of mix contained 6 g of lactose and 60 g of water the % lactose based on the lactose concentration in the mix would be = 6/(60+6) x 100 = 9%.
At low MSNF there is a tendency for the ice cream to taste ‘watery' and to lack characteristic ice-cream flavour. Growth of ice crystals may also be promoted.
Rothwell (1985) has illustrated how to calculate the maximum, note not necessarily the optimal, concentration of MSNF; all the solids apart from MSNF are summed and subtracted from 100. The product is then divided by 7 to give the maximum acceptable value. MSNF absorb about 6 times their own weight of water. As an example a mix containing 8% fat, 13% sugar, 1% stabiliser/emulsifier should have a maximum MSNF of
= 100-(fat +sugar+ emulsifier + stabiliser) 7
= 100 - (8+13+1) 7
= 11.14 % MSNF.
You can check your calculation using the calculator provided. Note that the maximum safe MSNF value calculated using this relationship may be lower than those used in current commercial ice cream manufacture since use of polysaccharide based stabilisers and added milk proteins have not been taken into consideration; these may enable higher concentrations of MSNF to be used without lactose crystallisation problems. With the use of appropriate stabilisers, good manufacturing practice, a short shelf life and excellent temperature control it is possible to produce ice cream with surprisingly high levels of lactose with few problems!
Organic ice-cream makers using cream and or butter to raise the fat content of on-farm produced ice cream but who do not use SMP (also called non fat milk powder) or concentrated milk as a source of additional MSNF will find it difficult to meet ideal MSNF recommendations. To some extent the use of organic stabilisers can help overcome some of the problems that can arise with low MSNF ice cream during storage.
OVERRUN
Overrun is a measure of the volume of air added to ice cream or gelato during manufacture. Apart from contributing to the profitability of an ice cream business, overrun influences hardness or resistance to scooping and certain other important product characteristics. High levels of overrun can be generated using air pumps but there is a maximum value for each ice cream mix and above this value product shrinkage and other problems may develop on storage. The maximum value is dependent on several factors including the MSNF, fat and total solids concentrations of mixes. Iversen (1983) has suggested the following relationship between MSNF, total solids and fat for deriving the maximum overrun:
maximum overrun = (Fat + MSNF + total solids) x 2.
There are several methods that can be used to determine the quantities of ingredients required to meet a target ice cream mix formulation. However, ingredient costs are also important and there is a balance between cost and quality that often must be considered. One of the simpler methods for performing mix calculations is known as the serum point method and can be learned quite quickly. Mix compositions can also be calculated using linear programming and other more advanced mathematical techniques.
Descriptions of the serum point method can be found in several text books including Hyde and Rothwell (1973), Marshall and Arbuckle (2000), Marshall, Goff and Hartel (2003) and Goff and Hartel (2013). The older books use imperial measures e.g. pounds but they are easy to follow. A particularly clearly written explanation of this method, for a limited number of mixes, is given in an inexpensive booklet by Rothwell (1985). Professor Douglas Goff, University of Guelph in Canada has an excellent website that provides an Ebook on Ice Cream. This has a section on worked examples of mix calculations.
The serum point method is based on the principle that the quantities of MSNF and fat contributed by ‘milk' of any composition can be subtracted from the entire quantity of fat and MSNF required in a mix, leaving the remainder to be supplied by concentrated sources of MSNF or fat.
There is a logical approach to solving problems using this method and the following description has been adapted from that given by Marshall and Arbuckle (2000).
1. List the fat, MSNF, sugar, emulsifier, and stabiliser concentrations, usually as percentages, that are required in the ice cream mix.
2. The quantities of single source ingredients required for 100 kg of mix are then calculated next. For example if 15 % sugar is required, then:
the quantity of ingredient, in this case sugar = 15 kg (15/100*100).
3. Next the weight of serum in the mix (serum is water and MSNF or milk serum solids) and is obtained by subtracting the weights of all of the other ingredients from 100 kg of total mix.
4. To calculate the quantity of concentrated milk needed, it is necessary to know the quantity of serum solids and the quantity of serum in 1 kg of concentrated milk as well as quantities of the same components in the mix. The formula for the quantity of concentrated milk is
Quantity of concentrated milk = (MSNF needed) – serum of mix x 0.09.
(MSNF/kg concentrated milk) – serum/kg concentrated milk) x 0.09.
Note the figure 0.09 represents the approximate % MSNF of skim milk and represents 9% MSNF in milk, if the actual value has been determined by analysis then this should be used. Concentrated milk is a general term for milk powder, condensed or evaporated milks and other sources of concentrated milk solids.
5. If the concentrated milk also supplies sugar or fat, these contributions must be calculated:
Fat contribution = (quantity of concentrated milk) x (% fat)
Sugar contribution = (quantity of concentrated milk) x (% sugar)
6. The quantity of fat required in the mix is calculated from the milk and cream, or milk and cream sources, by subtracting the quantity of fat in the concentrated milk from the total quantity of fat needed in the mix:
7. The quantity of sugar that must be added to the mix is calculated by subtracting the quantity of sugar in the concentrated milk from the total quantity needed in the mix:
8. If there is no fat or sugar in the concentrated milk, steps 6 and 7 are not required.
9. The quantity of milk and cream required are calculated by subtracting the total of all other ingredients from the 100 kg of mix:
Milk and cream = (100) – total quantities of other ingredients)
10. The quantity of cream needed is then calculated as follows:
Cream = (fat needed – [{milk and cream needed x (% fat in milk)}
(quantity of fat/kg cream) – (quantity of fat/kg milk)
11. Calculate the quantity of milk needed by subtracting the quantity of cream from the total quantity of milk and cream.
12. Check and confirm that the total weight of all ingredients equals 100 kg.
13. The calculations should be verified by preparing a table listing all the ingredients, their weights and their MSNF, fat and sugar contributions. These should be added and compared with target values to ensure that the calculation has been undertaken correctly.
The finished product! Ice cream on sale in a typical small shop in Rome, Italy.
LIMITATIONS OF SERUM POINT METHOD
The serum point method if used with 2 or 3 sources of MSNF e.g. cream, SMP and whole milk can sometimes appear to slightly over calculate the concentration of MSNF required, particularly for mixes containing 'low fat concentrations' e.g 8% and particular concentrations of MSNF. The over calculation is typically small and is generally insignificant. It is easy to make adjustments to spreadsheets to allow for this. Typically this involves slightly reducing the weight of the most concentrated source of MSNF e.g. condensed milk or milk powder to compensate. Interestingly this, to the best of my knowledge, has not been reported in the literature; probably because spreadsheets were not readily available when the method was developed.
There is also a simple algebraic method for solving the simultaneous equations associated with mixes. This has been well described by the previous authors and a number of worked examples can be viewed on Professor Goff's website. This gives identical approaches to the serum point method including the over calculation of MSNF in some complex mixes. It is easy to write a simple macro to use with spreadsheets to solve 3 variable simultaneous equations. The overestimation of MSNF is not an issue if linear programming solutions are used.
While the serum method will typically only allow the calculation of three ingredients supplying either fat of MSNF it is possible to work with four or more ingredients e.g. milk, cream, skimmed milk powder and whey powder supplying fat or MSNF. Spreadsheets illustrating how to do this can be downloaded from this site.
DAIRY SCIENCE AND FOOD TECHNOLOGY ICE CREAM MIX CALCULATOR
You can use the DSFT ice cream mix calculator (subscription required) to check your calculations. This uses linear programming to solve the mix calculation variables.
The websites and text books below along with the general instructions above provide a good introduction to the principles of ice-cream mix calculation. A number of queries have also been addressed in the Forum.
Primary sources or Journal articles (Students note reviews in Journal articles are not primary sources!)
Iversen, E.K. (1983). Scoopable ice cream. N. Eur. Dairy J. 49, 116-122.
Guinard, J.X., Zoumas-Morse, C., Mori, LL., Panyam, D. and Kilara, A. (1996). Effect of sugar and fat on the acceptability of vanilla ice cream. J. Dairy Sci. 79, 1922–1927.
Peryam, D. R., and Pilgrim, F. J. (1957). Hedonic scale method of measuring food preferences. Food Technol. Supplement. 11:9-14.
Text Books
Goff, H. D. and Hartel, R. W. (2013). Ice Cream. 7th Edn. Springer: New York, U.S.
Hyde, K. A. and Rothwell, J. (1973). Ice Cream. Churchill Livingstone: Edinburgh, Scotland.
Marshall, R. T. and Arbuckle, W. S. (2000). Ice Cream. 5th Edn. Gaithersburg , MD , U.S.
Marshall, R. T., Goff, H. D. and Hartel, R. W. (2003). Ice Cream, 6th Edn. Kluwer Academic: New York, U.S.
Rothwell, J. (1985) Ice cream. Reading University, UK. Published by Dr James Rothwell. Available through the Ice Cream Alliance.
Tharp, B. W. and Young, L. S. (2013). Tharp and Young on Ice cream: An encyclopedic guide to ice cream science and technology. DEStech Publications Inc: Lancaster, U.S.
Mullan, W.M.A. (2007).
[On-line]. Available from: https://www.dairyscience.info/index.php/ice-cream/154-ice-cream-mix.html . Accessed: 28 March, 2023.
Updated: February 2010, August 2010, September 2012, February 2014, March 2015, July 2016, August 2019, June 2020.
We use cookies to improve our website and your experience when using it. Cookies used for the essential operation of the site have already been set. To find out more about the cookies we use and how to delete them, see our Privacy Policy.