Frozen food is defined as “tempered” when its temperature is raised from that of a solidly frozen condition to some higher temperature, still below the initial freezing point, at which it is still firm but can readily be further processed.  In other cases, unfrozen, chilled meat can benefit from “tempering” (partial freezing) to aid processing.  At different temperatures in the frozen state, the relative proportions of ice and unfrozen water in food vary.  As the percentage of water that is ice decreases, the mechanical properties of the food change and it becomes less hard and brittle.  This allows further processing, for example flaking, grinding or dicing.  Studies at Langford have shown that a difference of less than 2°C in mincing temperature can produce detectable texture differences in the final product (Sheard et al., 1988).

Meat and butter tempering have been some of the most successful applications of microwave heating in the food industry.  Investigations showed that 25 kg cartons of meat could be raised to an even temperature of between -5 and -2°C in 3 to 7 minutes using microwaves at a frequency of approximately 900 MHz (James, 1986).  Tempering time was a function of fat content and the initial temperature of the material.  In the past 20 years a number of continuous and batch tempering systems have been installed (James, 1995).

Although microwave-tempering systems have been available for many years the majority of the industry still uses rather uncontrolled air based tempering systems.  Our studies have produced design charts relating tempering time and final temperature gradients to the amount of fat in beef blocks and the air temperature and velocity used (Brown, 1997; Brown & James, 2006).  Even using high velocities and temperatures, 5 ms-1 and 3°C, blocks require a minimum of 24 h.  If a very small gradient is required, this will require 120 h at a lower temperature and velocity.

Over the past 20 years an increasing proportion of bacon has been pre-sliced and packed before it is delivered to wholesalers and retailers.  To achieve the required throughput, slicers have to operate at very high speeds, and to maximise the yield of high quality slices, the bacon has to be sliced in a semi-frozen, tempered, state.  Investigations carried out with commercial companies at Langford clearly showed that the optimum slicing temperature was a function of both the bacon and the slicer.  Obtaining the correct temperature throughout the bacon middle is crucial for a high yield of undamaged slices (James & Bailey, 1987).  High-speed photography was used to clearly demonstrate the effect of incorrect slicing temperature.  When the temperature was too low the hard bacon shattered and blade wear was excessive; when too high the soft bacon stuck to the blade and the fat was torn away from the lean.

Studies looked initially at the performance of long, single stage, tempering processes and then moved on to the development of more efficient two-stage processing systems (James & Bailey, 1987; James, 1997).  The correct design and operation of such systems was shown to be critical to the cost effectiveness of the slicing process.

In a two-stage tempering process, an initial blast freezing operation is followed by a separate period of temperature equalisation.  It is critical that the desired amount of heat is extracted in the initial blast freezing operation.  To provide accurate data for such purposes we constructed a multi-temperature facility to measure the rate of heat exchange between the air and with foods during heating and cooling operations (James & Bailey, 1982b).  In the two-stage bacon tempering investigations it allowed us to determine relationships between the environmental conditions, and weight loss and final equalised temperatures.  In each case, the maximum temperature difference across the backs was less than 1°C after 3 h in the equalisation room.

Critically, we were also able to provide comprehensive data on the change in refrigeration load with time during the blast freezing operation and the total amount of energy to be extracted.  We were able to provide a large refrigeration contractor with a total set of process design data for the two-stage process.  In its first year of operation, the capital cost of the plant was recovered due to the cost savings from reduced weight loss, lower slicing losses and lower energy consumption.

Further design data on two-stage bacon tempering was published in 2003 (Brown et al., 2003).  This showed that optimising a two-stage tempering system using a purely empirical technique is a costly, labour intensive process.  Reliable predictive techniques developed at Langford have been shown to substantially improve the design of such processes (James, Palpacelli & James, 2003).

This worked showed the importance of knowing the exact freezing point of the bacon.  The survey work by James & Bailey (1987) had found that mean percentage salt and water contents of bacon varied considerably, and initial freezing points, and consequently the optimum slicing temperatures, could therefore vary by 5°C or more.  We knew that there was a relationship between salt content and freezing point.  However the relationship we used was published in the 1920s (Calow, 1929) and the graph not particularly accurate.  Little other data appeared to have been published since.  Therefore experiments were carried out using the “Karlsruhe test substance” (“Tylose”) with varying salt concentrations as a cured meat substitute (James et al., 2005a).  This determined the relationship between freezing point (Tf) and salt content (xs = mass fraction of salt) to be: Tf=-106.55xs-0.9241.  This correlated to within ±0.5°C of published values for cured pork and within ±0.9°C of theoretical predictions, showing that Tylose modified in this way could be used to simulate the freezing of cured meat.