16 January 2015
When presented with the term ice cream, it can be described as a semi-soft, frozen dessert. The physical characteristics of ice cream do make it hard to measure consistently; it would be detrimental to alter the production process to attempt to attain a sample of the frozen finished product and most measurements would be taken in a warm laboratory meaning the sample would be changing consistency as it melts. Therefore, when talking of measuring the colour of ice cream, it is not the frozen matter being measured but rather the liquid substance that gives ice cream its desired colour. This liquid substance can be measured easily and reliably with repeatable data to assure top of the range quality for consumers.
The first variable that must be addressed is
what instrument should be used? The main options available would be to use either a Colorimeter or Colorimetric Spectrophotometer, but which one? A tristimulus colorimeter can measure colour and the technology is built with the older optical filter design. A Colorimetric spectrophotometer, on the other hand carries out colour measurement avoiding the variation of the optical filters and provides a far more detailed analysis of colour by obtaining reflectance or transmittance values over the visible spectrum. Its reliability of measurements for longer periods of time makes it an ideal tool when measuring a product for public consumption.
There are a wide range of spectrophotometers available for a variety of applications but for ease of use and space saving design, a HunterLab ColorFlex EZ Spectrophotometer is perfect. Samples of the product can be sat on top and measured quickly, allowing for many sample measurements to be taken per day to keep up with the masses of ice cream produced. When coupled with EasyMatch QC software, once set-ups are completed it can be as easy as loading the sample, pressing a button and receiving a pass or fail notification. With the numerical data provided, it is possible to see non-conforming results, for example if the sample is too yellow, and make suitable alterations.
When measuring ice cream, the next aspect would be sample preparation. How a sample is prepared for measurement is key and can affect the outcome of the measurements. During the manufacturing process, many operators may be required to take sample measurements, record data and make alterations to batches. A uniform method of sample preparation and measurement procedure is the more reliable way to organise a quality control process.
The key points to consider are: What receptacle is going to be used to contain the sample for measurement?
A circular, glass sample cup would be ideal. This would allow the sample to fit flush against the sample area on the instrument and lets the liquid be distributed evenly for measurements. For a busier environment, plastic containers can be used if there is the likelihood of breakages but these are subject to being scratched more than a glass container which could affect measurement results. Some sample cups also come with marker lines around the external circumference of the cup to ensure the same amount of liquid is added each time.
How is the sample going to be added to the sample cup?
Pouring or pipetting? If the sample is poured, there is the chance of not measuring out the same thickness of a previous sample accurately which could result in varied data, whereas a pipette can allow for gradual additions. However using a pipette, as it is designed, can result in air bubbles being added to the sample. These can also alter the colour of the liquid being measured and provide inaccurate results.
How much of the sample is going to be measured?
Essentially, how thick will your sample be? A thick sample would provide an opaque surface for the instrument to measure the reflected colour which is what a human observer would be looking at. This does use more of the sample however and the thickness would depend on the depth of receptacle available for use. The sample cups should not be full to the brim as this would waste a lot of operator time cleaning any spills and, if some spills are negated, could damage equipment or provide inaccurate results. A thinner sample, although it uses less of the matter, saves waste and is certainly easier to handle, would allow more of the light to travel through the sample, therefore not giving an accurate representation of reflected colour.
What will the sample be covered with?
For increasing the reliability and accuracy of results, samples can either be backed with a white backing tile or covered with an opaque cover, providing the same method is used for all samples. The ideal method would involve using the opaque cover as this would absorb any excess light that may pass through the sample whereas the white backing tile would reflect this light back into the instrument to be measured. The light reflected by the white tile would be included in the colour measurement reading and give erroneous results that do not best represent the colour of the sample.
How many samples of the same batch will be measured?
The accuracy of liquid measurement would depend on how many samples of the same batch are measured. It is beneficial from a time scale point of view to just allow for one measurement to be taken. For the sake of accuracy however, a few sample measurements should be taken to negate any anomalies and even provide an average set of data for analysis. Multiple samples of a batch are also useful when colour matching. If only one sample is taken and proves to be unacceptable, does that mean the entire batch isn’t right or was there something wrong with the sample? Was the sample cup dirty or scratched? Is it worth altering an entire batch from the data given from one sample?
This begs the question, is there enough time for multiple measurements to be taken? Is the quantity of samples measured more important than the quality of results?
The next points to list would be in relation to taking the measurements:
How many measurements will be taken of the standard?
With a perfect batch of the liquid, a standard could be set. This standard would contain the numerical colour data of the perfect substance which every future sample could be compared to. Due to the particles in a liquid, light can be reflected differently so it is advisable to take an average reading of the standard. Four separate sample cups filled in the same way with the same amount of liquid product can be measured to create an average. This average data would then be used as the perfect reference point for all batches of the intended same colour. Standards for different colours can also be taken in the same way, depending on batch colour being measured that day.
How many measurements will be taken of the samples?
In a similar way to the standards, samples of subsequent batches can be measured in the same way by taking an average. This can be done either with four separate sample cups or by turning the sample between each reading. Even this slight variation can result in more accurate results as it gives a better view of the sample as a whole.
What parameters are being evaluated in the results?
Many colour scales can be used, depending on the application. More universally used at present are the CIE L*, a*, b* and Hunter L, a, b colour scales as these give a detailed overview of an object’s appearance. These colour scales give numerical data that can be presented to gauge how light to dark, red to green and yellow to blue a liquid is. The differences between a sample being measured and the standard it is being compared to can also be represented as delta values. These delta values can tell the operator how far off perfect the sample is and what colour needs to be increased/decreased to rectify it. Included with these could be delta E* (dE*) that portrays how different to the standard the sample is in its entirety. The use either of these colour scales in addition to the delta values makes colour matching a simple, cost effective task. Tolerances can also be set depending on the relationship between numerical differences and what is visibly different to a human observer.
How simple would the results need to be presented?
Are the measurements being taken for the purpose of colour matching or simply document the efficiency of the manufacturing process? If it is a case of an operator needing to know if the batch being produced is of the correct colour so they can move forward with the production process, a simple Pass/Fail notification may be all that is required. If the sample is too different from the standard and needs to be altered, the CIE/Hunter LAB values can be used to correct the colour.
The final question needing to be asked would be what is going to be done with the results?
Content Written by Rachael Stothard