Water Activity and Moisture: The Complexity and Interrelationships Explained

Water is a critical factor for the physical and chemical stability of solid oral dosage formulations. It is necessary for the growth of microorganisms and can impact a formulation’s stickiness, caking, and structure. Water’s role in stability is in its potential to contribute to and participate in deteriorative reactions¹. Degradation of a drug product’s active pharmaceutical ingredient (API) through hydrolysis before the end of the product’s shelf life can be detrimental to therapeutic efficacy or patient safety.

Traditionally, discussions about water in products or ingredients focus on moisture content, which is a quantitative measure of moisture in a product². In food sciences, however, the qualitative parameter, water activity, is often considered the critical moisture parameter³. Nowadays, the pharmaceutical industry is also acknowledging the importance of measuring water activity, as evidenced by a US Pharmacopeia chapter <1112> Application of Water Activity Determination to Nonsterile Pharmaceutical Products and a proposed chapter <922> Water Activity⁴. For example, water activity is measured to reduce the degree of microbial testing required to control the product or to define packaging criteria to ensure stability over its entire shelf life. 

Definition of water activity 

Water activity (Aw) is a measure of the energy status of water in a material. It is defined as the ratio of the fugacity of the water (fw) to the fugacity of pure liquid water under reference conditions (fw 0). Fugacity is the escaping tendency of a substance and can be replaced by the vapor pressure (p), provided the vapor behaves as an ideal gas⁵. Water activity is therefore often defined as the vapor pressure of water in a material (p) relative to the vapor pressure of pure water (P₀), at the same temperature (T) and pressure: 

The vapor pressure of pure water (P₀) can be understood as follows. When pure water is placed in a closed container, some water molecules will escape from the liquid water into the container’s headspace, as shown at the bottom of Figure 1. The water molecules in the gas phase are in equilibrium with the liquid water; there are constantly vapor molecules returning to the liquid state and liquid water molecules escaping to the gas phase. The molecules in the gas phase are in constant random motion and exert pressure on the inside of the container’s sides and top. The higher the concentration of water molecules in the headspace and the faster the molecules are moving, the higher the vapor pressure. 

The vapor pressure of water in a material (p) is defined in the same way. When a solid material is placed in a closed container, water molecules can escape from the material into the headspace, resulting in a vapor pressure, as shown at the top of Figure 1. The vapor pressure for a solid material will always be less than it is for pure water, due to the association of water to the material. The less the degree of association of water to the material, the higher the water activity, up to a defined maximum of 1.0. 

When a material only contains water molecules that are so tightly bound to the material that they cannot escape into the gas phase (as with crystal water), the material’s water activity is 0.0. Water activity values between 0.0 and 1.0 reflect the difference in fugacity of “free” water in the material. The water activity depends on the total amount of water in the material, as well as on the material’s ability to bind water in its structure—often referred to as hygroscopicity. A material’s hygroscopicity depends on its chemical composition, physical-chemical state, and physical structure. Capillary forces, solutes, surface, and colligative effects can all impact hygroscopicity. 

Water activity and product quality 

To understand the difference between water activity and moisture content for product stability, one should understand that not all water in a material is available for microbial growth or enzymatic and chemical reactions. Completely bound water, like water of crystallization, is not available for microbial growth or reactions⁷. Available water is only the water that is loosely bound to the material. Water activity is the qualitative measure of the energy state of free water in a sample and, therefore, determines the availability of water for microbial growth or enzymatic and chemical reactions. 

Figure 2 shows the relative reaction rate of different types of reactions as a function of water activity. Each type of reaction (or organism growth) has an optimal water activity range at which it proceeds at the highest rate. Non-enzymatic browning reactions, for example, have an optimal water activity of 0.75, and the reaction rate of degradation reactions is exponentially related to water activity^8,9. Microorganisms have critical water activity values below which they cannot grow. At water activity values below 0.6, all growth of microorganisms is inhibited¹⁰. 

Water activity and moisture migration 

Differences in water activity, or fugacity, between two discrete materials or one material and the environment are the driving force for moisture migration. Water will always migrate from high water activity to low water activity until reaching equilibrium, even if the material with lower water activity has higher moisture content.

Water migration also occurs when a sample has a different water activity than the environmental relative humidity (RH) in which it is stored. If the sample has water activity of 0.5 and is stored in a controlled environment with an RH of 80 percent, the sample will absorb water until reaching equilibrium and a water activity of 0.8. If a sample with water activity of 0.8 is stored in a controlled environment with an RH of 50 percent, it will expel water until reaching equilibrium and a water activity of 0.5. A sample’s actual water activity values can therefore contain useful information about the sample’s most-recent history, such as the protectiveness of packaging, intermediate opening and closing, and storage conditions. 

The speed at which water activity equilibrates depends on the difference in water activity between the materials, the ability of water molecules to diffuse within the materials, the available surface area for moisture migration, the amount of moisture that needs to migrate, and the moisture barrier properties of the packaging¹³. Table 1 shows the water activity inside different packaging types after storage for 6 months at 40°C and 75 percent RH. As indicated by the difference between 0.75 Aw in polyvinyl chloride compared to 0.2 Aw in polychlorotrifluoroethylene, water activity inside packaging is highly dependent on the packaging material’s moisture vapor transmission rate (MVTR)¹⁴. The MVTR of packaging can be used to predict a product’s speed of water uptake and set shelf-life and storage conditions. 

Moisture content 

The loss on drying (LOD) and Karl-Fisher (KF) methods are two commonly used methods of measuring moisture content. LOD is a gravimetric method involving the heating of a known quantity of material until all free water is evaporated. For lactose, for example, this method includes 2 hours of heating at 80°C¹⁵. By weighing the sample before and after drying, you can determine the amount of free moisture in the sample. 

The KF method uses titration to measure the total amount of water in a sample. The method is based on the fundamental reaction involving the reduction of iodine by sulfur dioxide in the presence of water. The sample is titrated with an iodine-containing KF solution. When water is present in the system, all added iodine is consumed. When no water remains, excess free iodine is detected by two platinum electrodes immersed in the solution¹⁶. You calculate the amount of water in the sample from the amount of KF solution added before the excess iodine appeared. 

For excipients such as lactose, you can use the LOD and KF values to calculate the amount of crystal water in a material. The difference between the total amount of water measured by KF and the total amount of free water measured by LOD provides the amount of crystal water in the sample, as shown in Figure 3. 

Relating moisture content to water activity

Water activity and free moisture content are related by the sorption isotherm, which is measured by dynamic vapor sorption (DVS). In a DVS experiment, a material is forced to different water activity values by exposure to an atmosphere with controlled water vapor pressure and temperature. The sample’s weight is constantly monitored, and changes are the result of moisture being absorbed into or expelled from the material. 

A DVS experiment generally starts by drying the sample at 0 percent RH to a water activity of 0.0. After that, the sample is exposed to a series of RH values, while the mass change is monitored. The RH only changes to the next value after the sample mass has reached equilibrium. 

In this way, the equilibrium mass of free moisture in the system can be plotted as a function of the RH. In the isotherm graphs, the RH equals the water activity times 100 percent, as these points are defined when the system is in equilibrium. Sorption isotherms are unique for each material and are an intrinsic material property. If sorption isotherms are available, the relationship between free moisture content and water activity is defined, and both parameters can be derived from the other by interpolation. Looking at a sorption isotherm, you can evaluate a sample’s actual moisture status by either measuring the free moisture content (y-axis, by LOD) or the water activity (x-axis), as shown in Figure 4. You must be careful with sorption isotherms, however, as some materials might undergo structural changes due to drying or exposure to high humidity, resulting in a change in the sorption isotherm afterwards (see next section).

Reversibility of moisture sorption isotherms 

A material’s sorption behavior depends on its chemical composition, physical-chemical state, and physical structure and is influenced by many factors, including capillary forces, surface, and colligative effects. 

Some materials have different adsorption and desorption profiles, which is called hysteresis. A hysteresis often exists with desorption giving higher equilibrium moisture content than sorption. A hysteresis can be the result of different interactions and forces during sorption or desorption, as with capillary condensation¹⁷. Cellulosic material, for example, often shows a hysteresis that may be related to the mechanics of shrinkage and swelling. Cellulose swells to form water-cellulose bonds upon adsorption, but these bonds do not break upon desorption at the same chemical potential¹⁸. 

Another change in sorption behavior occurs when structural modification happens because of exposure to a relative humidity. In this case, the material’s structure changes during a dynamic vapor sorption measurement. Amorphous lactose, for example, is crystallized to lactose monohydrate upon exposure to RH greater than 50 percent (at 25°C). This structural change reduces the material’s hygroscopicity, modifying the sorption isotherm while it is being measured. The initial material (partial amorphous lactose) and the crystallized material will have different sorption behavior as well as material properties. Measuring DVS sorption isotherms can therefore be very useful to reveal if a product is changing upon exposure to high RH. If the original structure is important for functionality, you can identify a critical humidity value.

Predicting the water activity of mixtures 

Water activity will always try to reach equilibrium within a system. Water will migrate from regions with high Aw to regions of low Aw for all system components, including the ambient atmosphere. However, you cannot predict the amount of water migration by looking at water activities alone. It also depends on each component’s ability to bind water in its structure. Understanding moisture migration in mixtures requires an understanding of the components’ sorption isotherms at the temperature of interest. Sorption isotherms are unique for each material and product¹⁹, and they provide all the information needed to predict the final amount of moisture migration in a system. Figure 5 provides examples of sorption isotherms for some commonly used excipients. 

You can determine a sample’s final moisture status using sorption isotherms based on the following two facts: 1. Moisture will migrate from high water activity to low water activity, until achieving equilibrium. 2. The mass balance of the total system is zero. This means that the total amount of free moisture in a closed system (which can include air) is not changing upon moisture migration. 

Knowing the amount of free water at different water activities for all components is critical for predicting moisture migration. For simplification, the following description and examples use reversible sorption isotherms, but similar calculations can be made for irreversible sorption isotherms. 

The total amount of free water defines the sample’s final moisture status. Upon preparation of a blend of N components, the final moisture content (M_^f) of the blend can be calculated with the fraction of component i (x_i) and the initial moisture content of component i (Mi) via:

The expected final water activity can then be interpolated from the mixture’s sorption isotherm.

Examples of predicting water activity of mixtures 

The following example demonstrates how to calculate moisture migration for a blend of two components. The components include a coarse-grade lactose monohydrate (Pharmatose 80M) with low hygroscopicity and an API, which are blended together in an 80:20 ratio (% w/w). The initial water activity of the API is 0.1, and the initial water activity of the lactose is 0.3. The sorption isotherm of the lactose is measured using a ProUmid SPS-1μ advanced system operated at 25°C with 20 minutes between weighing cycles, equilibrium condition 0.003 percent w/w per 80 minutes, and boundary conditions of 8-12 hours. The isotherm of the API is imaginary. The component isotherms are combined using Equation (2) into the sorption isotherm of the proposed 80:20 blend, as shown in Figure 6. 

From the sorption isotherms, one can see that the API contains 0.060 percent moisture at 0.1 water activity, and the lactose contains 0.005 percent moisture at a water activity of 0.3. The total moisture content of the blend will therefore be: 

(80% ∙ 0.005%) + (20% ∙ 0.060%) = 0.016%

This moisture content value of 0.016 percent can be used to look up the corresponding water activity value in the sorption isotherm of the blend. Assuming the change of moisture content will be linear over the range of 0.1 to 0.2 Aw, the final water status at which the sample equilibrates will be approximately 11 percent RH or 0.11 Aw. Note that the free moisture content is generally not linearly correlated with the water activity, however. You can obtain more accurate results by increasing the number of measurement points of the isotherms. 

To evaluate the impact of excipient hygroscopicity on a blend’s final moisture content, a second example two-component blend contains the same API but a more hygroscopic microcrystalline cellulose (MCC) excipient (Pharmacel 101). The MCC is blended with the API in an 80:20 ratio (% w/w). The initial water activity of the API is 0.1, and the initial water activity of the MCC is 0.3. The sorption isotherm of the MCC is measured using a ProUmid SPS-1μ advanced system with the same parameters used for the lactose, and, again, the API isotherm is imaginary. The component isotherms are combined using Equation (2) into the sorption isotherm of the proposed 80:20 blend, as shown in Figure 7. 

From the sorption isotherms, one can see that the API contains 0.060 percent moisture at 0.1 water activity, and the MCC contains 2.855 percent moisture at 0.3 water activity. The total moisture content of the blend will therefore be: (80%  2.855%) + (20%  0.060%) = 2.296% 

This moisture content value of 2.296 percent can be used to look up the corresponding water activity value in the sorption isotherm of the blend. Assuming the change of moisture content will be linear over the range of 0.20 to 0.30 Aw, the final water status at which the sample equilibrates will be 0.297 Aw. Again, free moisture content is generally not linearly correlated with water activity, and you can obtain more accurate results by increasing the number of measurement points of the isotherms. 

As these examples show, the water activity of ingredients alone is not enough to predict the resulting water activity and, therefore, stability of a blend. For predicting stability, full knowledge of moisture behavior combined with control of environmental conditions and packaging is key.

Conclusion 

Moisture is a critical factor for the physical, chemical, and biological stability of many materials, and different moisture parameters exist. For pharmaceutical excipients, total moisture content by KF (for lactose) and free moisture content by LOD are leading moisture parameters from a regulatory perspective. 

Water activity describes the energy status of water in a sample and can be used to provide information about the sample’s recent history and predict relative reaction rates. It can also be used to predict the direction of moisture migration in a blend, although the amount of moisture migration also depends on the hygroscopicity. Knowing the actual moisture status and sorption isotherms allows you to describe a blend component’s full water status and make more reliable predictions. 

References

1. T. P. Labuza and B. Altunakar, “Water activity prediction and moisture sorption isotherms,” Water activity in foods: fundamentals and applications, 2020, pages 161-205. 

2. Decagon Devices, “Fundamentals of Water Activity,” 2006, Accessed via: www.graintec.com.au/ media/12856/Fundamentals.pdf. 

3. Safefood 360, “Water activity (aw) in Foods,” 2014, Retrieved April 2020 from: safefood360.com/resources/ Water-Activity.pdf. 

4. online.usppf.com/usppf/document/GUIDB5DDDC52- 0DC8-4603-B81B-03EA4FB025B8_10101_ en-US. 5. D. S. Reid, “Water activity: fundamentals and relationships,” Water activity in foods: Fundamentals and

 applications, Wiley-Blackwell, 2020, pages 13-26. 

6. DRSJSCHMIDT, “Moisture Content and Water Activity” [video], Available at: www.youtube.com/ watch?v=-LiSankbfSk. 

7. L. Skowronsky, “Inhibition of microbial growth in solid dosages at ICH stability storage conditions,” European Pharmaceutical Review, 2011. 

8. M. Lally, “Introduction to USP <922>: General Chapter for Water Activity Measurement,” Available at: vertassets.blob.core.windows.net/download/e97f8ab4/ e97f8ab4-5bad-46a7-bf18-f40bf301f292/article_ introduction_to_usp_922_general_chapter_for_water_ activity_measurement.pdf.

9. K. C. Waterman, and R. C. Adami, “Accelerated aging: prediction of chemical stability of pharmaceuticals,” International Journal of Pharmaceutics, 2005, Vol. 293, No. 1-2, pages 101-125. 

10. G. V. Barbosa-Cánovas, A. J. Fontana Jr., S. J. Schmidt, and T. P. Labuza (Eds.), Water activity in foods: fundamentals and applications, Wiley-Blackwell, 2020. 

11. M. Lutovska, V. Mitrevski, T. Geramitcioski, V. Mijakovski, and I. Andreevski, “Water Activity vs. Equilibrium Moisture Content,” Journal on Processing and Energy in Agriculture, 2016, Vol. 20, No. 2, pages 69-72. 

12. T. P. Labuza, S. R. Tannenbaum, and M. Karel, “Water content and stability of low moisture and intermediate moisture foods,” Food Technology, 1970, Vol. 24, pages 543-550. 

13. R. Ergun, R. Lietha, and R. W. Hartel, “Moisture and shelf life in sugar confections,” Critical reviews in food science and nutrition, 2010, Vol. 50, No. 2, pages 162-192. 

14. M. Lally, “A Guide to USP <922> For Water Activity Determination” [Webinar], Pharmaceutical online. June 18, 2020. 

15. USP-NF 2020. Lactose monohydrate monograph. European Pharmacopeia 10.0 – lactose monohydrate monograph. 

16. K. Fischer, “Neues Verfahren zur maßanalytischen Bestimmung des Wassergehaltes von Flüssigkeiten und festen Körpern,” Angewandte Chemie, 1935, Vol. 48, No. 26, pages 394-396. 17. B. Ahrenholz, J. Tölke, P. Lehmann, A. Peters, A. Kaestner, M. Krafczyk, and W. Durner, “Prediction of capillary hysteresis in a porous material using lattice-Boltzmann methods and comparison to experimental data and a morphological pore network model,” Advances in Water Resources, 2008, Vol. 31, No. 9, pages 1,151-1,173. 

18. M. Chen, B. Coasne, R. Guyer, D. Derome, and J. Carmeliet, “Role of hydrogen bonding in hysteresis observed in sorption-induced swelling of soft nanoporous polymers,” Nature communications, 2018, Vol. 9, No. 1, pages 1-7. 

19. J. L. Ford and R. Willson, “Thermal analysis and calorimetry of pharmaceuticals,” In Handbook of thermal analysis and calorimetry, Vol. 4, Elsevier Science BV, 1999, pages 923-1,016.