BROWSE ALL ARTICLES BY TOPIC

RELATED ITEMS

Bookmark and Share

From: Food Quality & Safety magazine, April/May 2010

Moisture Content Analysis

A new method may offer improved precision and reliability

by Julia Mumford

Like many products, nuts have a moisture content (MC) “sweet spot” where they’re dry enough to meet customer quality specs but not so dry that they break during shipment. To check quality reliably as the nuts are processed, shipped, stored, and used as ingredients, manufacturers need a moisture method that is:

  • Highly accurate: Traditional loss-on-drying methods only reach ±0.5% MC. In nuts, that margin of error is often greater than the difference between a nut that breaks because it’s too dry and one that contains too much water.
  • Verifiable: Currently, as one quality control manager put it, “My suppliers tell me what the moisture content was when it shipped, but they can say whatever they want, because we don’t have reliable standards to measure against.”
  • Simple to perform: Operators will vary in their education, training, and skill. The method should be one that any minimally trained person can perform accurately.

A new method of moisture analysis offers promise in meeting those needs. This method measures moisture content using the dewpoint water activity method.

Water Activity and Moisture Content

Water activity quantifies the energy status of water in a product and uses the notation aw. It is a measure of microbial susceptibility accepted by the U.S. Department of Agriculture and a key component in many hazard analysis and critical control point plans. It can be measured with ±0.003 aw accuracy using tests that are both simple and precise (precise meaning whether the instrument gives the same answer each time). But can it be used to find moisture content?

Since work by chemists in the 1930s, food scientists have known that moisture content and water activity are related. The relationship, called a moisture sorption isotherm, is a graph showing how water is adsorbed into, and desorbed from, a product held at constant temperature.

Moisture Contenet Analysis

Isotherms are complex—unique to every product and surprisingly unpredictable. In the 1980s and 1990s, a lot of research went into making isotherms for common products and ingredients, but the process was painstaking—a single point on the isotherm could take a few weeks to produce.

Isotherms were so expensive and time consuming to make that they were beyond the reach of all but a few big labs. And their usefulness was limited because they didn’t have good resolution—a graph might be constructed from just six points, for example.

All that changed with the introduction of isotherm generators that can produce isotherms that include 50 to 100 points in just a day or two. With that breakthrough, it became possible to reliably determine moisture content from water activity.

One Instrument, Two Measurements

Let’s say a nut producer wants to measure the moisture content of pecans. The first step is to collect a data set of water activity and moisture content values to construct an isotherm. Then using a model, several of which are available, the pecan isotherm is converted into an equation. This equation will take the water activity reading of a pecan sample and convert it into a moisture content reading.

Whether the challenge is meeting a spec and certifying that it has been met or identifying where moisture problems happen in the supply chain, solid, reliable moisture content information is crucial.

The process could be simplified by programming the equation into a special instrument. The instrument could determine water activity and convert the reading into a moisture content measurement using a preprogrammed equation.

After processing the pecans, the producer can put a sample into the instrument, run it for five minutes or so, and get both moisture content and water activity readings on his pecans.

Moisture content in pecans is a touchy thing. Their target moisture content is around 4.6%. At moisture contents over 4.8%, they are susceptible to mold. The problem: Most moisture content methods are not precise enough to distinguish between the two. The dewpoint measurement can be much more precise, especially in intermediate moisture foods like pecans. Intermediate moisture foods have an S-shaped isotherm with a flat spot in the middle. That flat spot tends to be the moisture range food producers care about, and, in that range, small changes in moisture content correspond to big changes in water activity.

The difference between 4.6% MC and 4.8% MC translates to a difference between 0.5800 and 0.600 aw—a huge difference to a water activity instrument with 0.003 aw accuracy. Loss-on-drying methods can’t reliably tell the difference between 4.6% MC and 4.8% MC. A dewpoint method instrument can distinguish between 4.60% and 4.62% The magnifying effect of the dewpoint method can mean a 90% or 95% improvement in precision, a boon in some sensitive moisture content applications.

Table 1. The relative strength of the dewpoint moisture content method to accurately predict oven loss-on-drying values as indicated by the standard error of prediction (SEP) and R2 values.
click for large version
Table 1. The relative strength of the dewpoint moisture content method to accurately predict oven loss-on-drying values as indicated by the standard error of prediction (SEP) and R2 values.

Standards Make Comparisons Easier

A customer who gets ingredient supplies from overseas perfectly describes the standards challenge: “What’s the standard for moisture content? My supplier always suggests that the shipment picked up moisture on the boat. I think that’s bull, but what can I say? There’s no independent standard for moisture.”

Whether the challenge is meeting a spec and certifying that it has been met or identifying where moisture problems happen in the supply chain, solid, reliable moisture content information is crucial.

Benchmarking moisture measurements with independent standards is a critical need. There are well-established water activity values for a range of saturated salt solutions such as potassium chloride (0.5 m, 0.984 aw) and sodium chloride (6 m, 0.760 aw).

Unfortunately, there are no independent standards for moisture content. But because it measures both water activity and moisture content with a single instrument, the dewpoint method can leverage water activity standards to establish reliability for moisture content as well. Users can confidently compare readings across different instruments, different locations, and different time periods.

Often the quality of moisture measurements depends on the expertise of the person making the measurement. With the dewpoint method, expertise is not as much of a factor. Water activity instruments are simple to operate. In spite of the test’s scientific precision, the actual testing process can be performed accurately with a minimal amount of training. This decreases the risk of operator error and increases testing opportunities along the supply chain.

Many manufacturers measure both water activity and moisture content. Being able to make both measurements simultaneously on a single sample would save significant time and expense.

Table 2. Average precision values for oven loss on drying and dewpoint moisture content for all products analyzed. The values represent an average of standard deviations of triplicate moisture analyses across 10 samples for each product.
click for large version
Table 2. Average precision values for oven loss on drying and dewpoint moisture content for all products analyzed. The values represent an average of standard deviations of triplicate moisture analyses across 10 samples for each product.

Dewpoint Moisture Content

With such obvious promise, food scientists set out to test the dewpoint moisture content method with a high-accuracy water activity meter and an isotherm generator.

Testing a moisture content method is not simple. Since there is no standard for measuring moisture content, a true accuracy can’t be calculated.1 The best way to compare moisture content methods, then, is by comparing their repeatability. The precision of the different methods was calculated as the average standard deviation of triplicate analyses across all samples for a given product.

Nine products, representing a wide variety of types, from homogeneous ingredients to complex final products, were selected for testing. The products included milk powder, flour, dry dog food, chocolate syrup, granola bars, potato flakes, solid dosage tablets, whole wheat, and beef jerky.

The moisture content in triplicate was determined for all of the products using convection oven loss on drying. Time and temperature settings for loss on drying were based on Association of Official Analytical Chemists (AOAC) recommendations when available.2 All moisture measurements are expressed as percent dry basis.

To create samples varying in moisture content, 10 sub-samples were taken for each product, five of which were wetted by exposure to 100% relative humidity in a sealed desiccator, while the other five were dried by exposure to desiccated air inside another sealed desiccator. Sub-samples were removed from the desiccators at different times to create samples varying in moisture content. As the sub-samples were removed, they were sealed in jars and set aside until all sub-samples had been removed from the desiccators.

The time in the wet and dry desiccators for the sub-samples of each product was adjusted based on the diffusion properties of the product. All sub-samples for a product were then analyzed in triplicate for moisture content and water activity.

The isotherm testing results were characterized using the Guggenheim-Anderson-Deboer, or double log polynomial, model. Dewpoint moisture content values were compared to average moisture content values from oven loss on drying. Standard error of prediction (SEP), interpreted as the 95% confidence interval for the predicted value around the actual value (a smaller value is better), and R2 value, interpreted as the goodness of fit of the correlation (closer to one is better) were used for comparisons between different isotherm curve/model combinations.

The relative strength of a secondary method is measured by how well it matches the reference method. For this study, the SEP value can be considered a measure of the ability of the dewpoint moisture content method to correctly match reference data.

Table 3. Commonly reported precision (also reported as accuracy) values for most frequently used moisture content determination methods.
click for large version
Table 3. Commonly reported precision (also reported as accuracy) values for most frequently used moisture content determination methods.

A Good Method

Dewpoint moisture content values agreed well with oven loss-on-drying values for all products, as evidenced by the low SEP and high R2 values (see Table 1, above). Most secondary methods for moisture content consider an SEP of 0.60% or lower to be acceptable. All SEP values except for chocolate syrup and jerky were close to that range, indicating that dewpoint moisture content can be considered a viable secondary method for lower- and intermediate-moisture products. It may be possible to develop better predictions with larger data sets for high-moisture products like chocolate syrup and jerky, but that was not attempted.

Table 2 (see p. 43) shows a comparison between the precision of the oven loss-on-drying method and dewpoint moisture content method. For every product investigated, dewpoint moisture content gave better precision, even though loss on drying is considered the reference method. Table 3 (see above) shows that, in comparison to reported precision values for other methods, dewpoint moisture content has the highest level of precision.

Dewpoint moisture content is an excellent moisture content measuring option and is especially attractive when both water content and water activity measurements are needed on the same sample. A product-specific isotherm, which can be obtained manually or using an isotherm generator, is needed. The precision of this method is the best of any of the secondary methods and exceeds that for loss on drying. It does an acceptable job of matching reference values, but the accuracy cannot be assessed because there is, to date, no absolute method for measuring moisture content. â– 

Mumford is a science editor at Decagon Devices Inc. Reach her at julia@decagon.com or (509) 332-2756.

References

  1. Isengard HD. Water content, one of the most important properties of food. Food Control. 2001;12(7):395-400.
  2. Mulvaney TR. In: Cunniff P, ed. Official Methods of Analysis of AOAC International. 16th ed. Arlington, Va.; 1995:42-1-42-2.

Advertisement

 

Current Issue

Current Issue

April/May 2014

Site Search

Site Navigation

 

Advertisements

 

 

Advertisements