Not all milk products are the same – and neither are their sampling methods. A sample pulled haphazardly from a bulk tank of liquid milk will give you very different results than one taken from the bottom of a ghee container or the center of a frozen ice cream tub. Standard Methods for the Examination of Dairy Products makes it clear: the best analytical technique in the world cannot rescue a sample that was poorly collected. Each milk product – liquid milk, milk powder, butter, ghee, and frozen dairy – has its own physical properties and storage conditions that demand a tailored sampling approach. Getting it right is the first step toward meaningful quality control.
Table of Contents
Why one sampling method doesn’t fit all dairy products
Milk products differ widely in physical state, composition, and packaging. Liquid milk stratifies due to fat rising to the surface. Milk powder segregates by particle size. Butter has uneven moisture and salt distribution. Ghee separates into distinct layers within containers. Ice cream contains multiple phases – ice crystals, fat globules, and air cells – that are extremely sensitive to temperature changes. The International Dairy Federation (IDF) emphasizes that standardized sampling methods are essential for achieving comparable, reliable results across the dairy chain, from primary production through to final product release.
The core requirement across all products is representativeness – the sample must reflect the true composition of the entire batch. Because it is expected that a sample represents the entire mass sampled, the initial concern is always adequate mixing. Samples from insufficiently prepared products, even if collected and transported perfectly, will not meet this requirement.
Sampling liquid milk from large storage tanks
Liquid milk presents an immediate challenge: fat is less dense than the other constituents of milk, so it naturally rises to form a cream layer at the surface. If a sample is drawn without addressing this, the result will over-represent fat content and give a distorted picture of milk composition. The solution is thorough agitation before sampling.
For small quantities, pouring milk from one clean container to another three or four times is sufficient to achieve uniform mixing. For larger batches – particularly bulk storage tanks – mechanical or hand stirrers are used to thoroughly agitate the milk before any sample is drawn. Bulk tanks must be thoroughly mixed to disperse the milk fat before a sample is taken using a plunger or dipper. One important precaution: milk churns readily at temperatures between 26.5°C and 29.5°C, so agitation near this temperature range should be avoided to prevent physical damage to the fat globules.
Once the milk is uniformly mixed, samples are collected using sterile sampling equipment from strategically positioned valves. In some operations, multiple samples from different points are combined into a composite sample, which further strengthens representativeness by averaging out any minor localized variations within the tank.
Sampling milk powder
Milk powder cannot be sampled the way liquid milk is. Powder particles of different sizes and densities tend to segregate during storage and transport – larger particles migrate toward the top or sides of a container, while finer particles settle differently. This stratification means that a sample taken from one spot in a bag or drum will not represent the whole.
The sampling approach for milk powder involves selecting a specified number of containers from the batch rather than sampling every container. According to IS 11546:1999 (which aligns with ISO 707:2008), the number of containers to be sampled depends on the size of the lot – a proportional representation approach that ensures the sample is drawn from a cross-section of the batch. A sampling spear or trier is the key tool here. It is inserted diagonally or vertically into the container with its ports closed, then opened to allow powder to fill channels at multiple depths, and finally withdrawn and sealed. This technique captures material from different layers within the container – top, middle, and bottom – in a single extraction.
The material collected from each selected container is then pooled to form a composite sample. This composite approach averages compositional differences between containers while providing a single, workable sample for laboratory analysis. Note that this standard procedure is not suitable for powder stored in large bulk silos; for those, small samples should be drawn continuously during loading or unloading operations to capture any variation across the entire silo volume.
Sampling butter
Butter is a semi-solid product with a tendency toward uneven distribution of moisture and salt throughout the mass. These variations are not always visible, which makes random sampling from multiple locations within each container essential.
Like milk powder, butter sampling begins with selecting the right number of containers from the lot. A random sampling plan is followed based on lot size: for lots of 1 to 9 bulk units, 2 containers are selected; for 10 to 49 units, 3 are selected; for 50 to 99 units, 4 are selected; and so on, scaling upward with lot size. From each selected container, a special butter boring tool (also called a butter trier or core sampler) is used to extract a cylindrical core by boring through the product at different locations and depths. This ensures that both surface and interior material are represented in the sample, capturing any gradient in composition that may exist.
Temperature control during butter sampling is critical. If butter is too cold, it becomes hard and difficult to bore through cleanly, risking incomplete extraction. If it is too warm, the fat structure softens and moisture can redistribute, affecting both the sampling process and the analytical result. Samples collected are then warmed gently – to a temperature not exceeding 39°C – and shaken vigorously to form a homogeneous fluid emulsion before chemical analysis. For butterfat analysis specifically, a portion of the emulsified butter is heated to 50-60°C until the fat separates cleanly.
Sampling ghee
Ghee is clarified butterfat that is liquid at room temperature but can solidify when cooled, and it has a significant tendency to stratify within containers during storage. Different components – moisture traces, fine solids, and fat fractions of varying composition – may concentrate at different depths. This layered nature makes it essential that any sample collected contains material from all strata of the container, not just the top layer.
The first step in ghee sampling is selecting the right number of containers from the lot. A standard random sampling procedure for ghee specifies: 1 container for a lot of 1; 2 containers for lots of 2-40; 3 for 41-110; 5 for 111-300; 7 for 301-600; and 10 for lots of 601 and above. If there is any indication of wide variation within the lot – for instance, ghee from multiple individual producers combined in a consignment – every container must be sampled individually.
The key sampling tool for ghee is a sampling tube (also called a zone sampler or ghee sampling tube). It is inserted vertically through the full depth of the ghee mass in the container, with the tube open so that material from all layers enters as it is pushed down. Once at the bottom, the tube is sealed and withdrawn, preserving a column of ghee that proportionally represents every layer. Equal quantities of ghee are then taken from each selected container and pooled to form the composite sample. For moisture testing and organoleptic evaluation, the sample is mixed without heating – directly in the container – until it is homogeneous. For sensory assessment, a small quantity may be rubbed on the back of the hand and the aroma evaluated as the warmth from friction releases volatile compounds.
Sampling frozen dairy products: ice cream
Ice cream is the most structurally complex dairy product to sample. It contains ice crystals, fat globules, air cells, and various mix-ins – all held together in a frozen matrix. Any disruption to the frozen state, however brief, causes ice crystal growth, air cell collapse, and fat redistribution. The result is a sample that no longer reflects the true composition or quality of the original product.
The cardinal rule for ice cream sampling is simple: maintain the frozen state at all times. Samples must be collected in insulated containers and transported to the laboratory using dry ice or refrigerated vehicles. Even minor temperature fluctuations of 2°C outside the recommended storage range are enough to trigger ice crystal growth in ice cream, compromising both the physical structure and the accuracy of subsequent analysis. The IS 11546:1999 standard specifies that the storage and transport temperature for frozen dairy products should be −18°C or lower, and in some cases even lower depending on the product and intended analysis.
Samples are drawn from multiple locations within a container and from multiple containers within the batch, following the same principle of proportional representation used for other products. The sampling equipment – spoons, scoops, or corers – must be pre-chilled before use so they do not introduce heat into the product at the point of contact. Once collected, samples are immediately sealed and stored frozen. Research confirms that ice cream maintained at −18°C shows considerably more ice crystal growth over time than products stored at −50°C or lower, which underscores why temperature discipline during sampling and transport is non-negotiable.
When preparing the sample for laboratory analysis, it should be allowed to soften only slightly – just enough to make it workable – before being thoroughly mixed to homogenize the fat and water phases. This controlled softening and mixing step ensures that test results reflect the product’s true average composition rather than a fat-rich or ice-rich pocket within the container.
Universal principles across all milk product sampling
Despite the differences between products, a few principles apply universally. After samples are collected, they must be protected from contamination and maintained under specified conditions so that when delivered to the laboratory, they are essentially the same as when collected. Analysis should begin within 24 hours of collection and, in most cases, no later than 48 hours. All sample containers must be food-grade, inert materials – typically stainless steel or approved food-grade plastics – that will not chemically interact with the product or introduce foreign odours and flavours.
Proper documentation is equally important. Every sample must be clearly labelled with the product name, lot or batch number, date and time of collection, collection temperature, and the identity of the person who collected it. This chain of information is what allows laboratory results to be traced back to a specific batch and used as the basis for quality decisions and regulatory compliance. International guidelines such as those from ICAR align with ISO 707:2008 in emphasizing that standardized sampling protocols, combined with rigorous quality control in the laboratory, are fundamental to producing results that are comparable across facilities and countries.
Sampling is also a skilled task. It must be carried out by trained personnel who understand the physical behaviour of each product type, can recognize signs of contamination or product abnormality, and know when conditions require deviation from standard procedure – and when they do not. Aseptic and representative sampling plays a vital role not only in finished product testing but also in in-process monitoring for contamination within dairy plants. Even well-designed analytical systems will produce misleading results if the sample feeding them is not representative of the actual batch.
What do you think? Given how much the physical state of a dairy product influences its sampling method, do you think the same level of attention is given to sampling at small-scale dairy operations as it is at large industrial plants? And with milk powder being prone to particle segregation in bulk silos, what alternative sampling strategies do you think could improve representativeness in those situations?
References
- https://ajph.aphapublications.org/doi/10.2105/9780875530024ch03
- https://fil-idf.org/our-work/methods-of-analysis-and-sampling/
- https://mel.cgiar.org/reporting/download/hash/7Rrm89n8
- https://law.resource.org/pub/in/bis/S06/is.11546.1999.pdf
- https://www.slideshare.net/slideshow/samplingproceduresofmilkandmilkproductspptx/258516998
- https://www.dairyfoods.com/articles/96278-best-practices-in-ice-cream-handling-key-when-distributing-ice-cream
- https://pmc.ncbi.nlm.nih.gov/articles/PMC6335139/
- https://www.icar.org/Guidelines/12-Milk-Analysis.pdf
- https://qualitru.com/dairy-plants/raw-milk-quality-2/
Leave a Reply