Seed Moisture Isotherms, Sorption Models, and Longevity

Abstract

Seed moisture sorption isotherms show the equilibrium relationship between water content and equilibrium relative humidity (eRH) when seeds are either losing water from a hydrated state (desorption isotherm) or gaining water from a dry state (adsorption isotherm). They have been used in food science to predict the stability of different products and to optimize drying and/or processing. Isotherms have also been applied to understand the physiological processes occurring in viable seeds and how sorption properties differ in relation to, for example, developmental maturity, degree of desiccation tolerance, or dormancy status. In this review, we describe how sorption isotherms can help us understand how the longevity of viable seeds depends upon how they are dried and the conditions under which they are stored. We describe different ways in which isotherms can be determined, how the data are modeled using various theoretical and non-theoretical equations, and how they can be interpreted in relation to storage stability.

Keywords: equilibrium relative humidity (eRH); isotherm; moisture content; seed longevity; storage stability.

Figure 1

A seed isotherm (blue line), the relationship between moisture content and water activity, is typically sigmoidal, with three regions (region I, II and III in blue text) indicative of the how the water is bound within the seed tissues. The imbibition curve (green line), which follows the change in seed moisture content over time, is also tri-phasic (phase I, II and III in green text). Seeds are usually “air-dry”, typically within region II of the seed isotherm, when they are sown. Water uptake which takes the seeds into region III of the isotherm, is initially fast, but is then more or less constant until radicle emergence at the end of phase II of the imbibition curve. Water content increases in phase III as the seedling develops. If the seeds have dried to a lower moisture level, approaching or within region I of the isotherm, rapid influx of liquid water may result in irreparable damage to the cell membranes. To avoid such imbibition injury, rehydration in a humid atmosphere, for example, over water, is recommended (Bewley et al., 2013). This schematic is broadly based on data for rice seeds from Hay and Timple (2016) and Zhao et al. (2020).

Figure 2

Schematic showing relative reaction rate (food stability diagram), seed longevity and moisture content plotted against equilibrium relative humidity. The sigmoidal teal and green solid sigmoidal lines illustrate desorption isotherms for barley and lettuce seeds, respectively (with moisture content as right y-axis). The dashed teal curve illustrates the adsorption isotherm (for barley seeds only), showing the effect of hysteresis. The solid straight lines indicate the relative longevity (left y-axis) of barley (teal) and lettuce (green) seeds in hermetic storage based on Roberts and Ellis (1989). At very high eRH, when oxygen is freely available (+O₂) longevity increases due to macromolecular repair (Figure 1). The black zig-zag curve shows the food stability isotherm, with relative reaction rate for aging reactions on the leftmost y-axis and an indication of the types of reactions occuring in the presence of oxygen (Rahman and Labuza, 2007). The orange shading also reflects this stabilty diagram, with lighter shading indicating greater stability.

Figure 3

Combining the concepts of glass transition theory and isotherm modeling to understand storage stability. (A,C) Show the state diagrams (glass transition temperature [T_g ] vs. moisture content) for cotyledon slices of pea cotyledons (from Buitink et al., 1999) and sunflower seeds (from Lehner et al., 2006), respectively, with the shaded area indicating the temperature-moisture content region where the tissue would be in a glassy state. Also shown are isotherms based on (A) fitted equations from Vertucci and Leopold (1987) at 15 and 35°C or (C) Cromarty's equation (Cromarty et al., 1982) at 45 and 60°C. The horizontal gray lines indicate these respective temperatures and the isotherms are positioned such that they meet T_g at the corresponding moisture content. The same relations are shown in (B,D), respectively, plotted against eRH rather than moisture content. It should be noted, however, that these glass transition temperatures were determined using differential scanning calorimetry in which samples are not in equilibrium and hence actual relationships between T_g and eRH (and be extension, moisture content) are not known.

Figure 4

Examples of different equations fitted to isotherm data [moisture content vs. equilibrium relative humidity (eRH)] from Hay et al. (2003) for seeds of Arabidopsis thaliana at 45°C. In the various equations MC is the moisture content and in isotherm equations, is usually expressed as g g⁻¹ dry weight (see Figure 1 for conversions between fresh and dry weight basis); a_w is the water activity (where a_w ≅ eRH/100, but see text); T is temperature (°C); OC is oil content; and letters in italics are parameters for estimation (variously, M_m, C, a, b, c, K′, K, k′ and k, according to convention in the literature). Equations shown in (A,C) represent theoretical models, in (A), the Brunauer-Emmett-Teller (BET) and Gugenheim-Anderson-de Boer (GAB) equations and in (C), the D'Arcy-Watt equation. The D'Arcy-Watt equation has three terms with parameters (K, K′, c, k, k′) that can be interpreted as representing the number and strength of strong (light blue), weak (mid-blue) and multi-molecular (dark blue) water binding sites, respectively. One of the parameters estimated when fitting the BET and GAB equations, is the monolayer value, M_m, and differences in the estimates from the two models are apparent in (A). This monolayer value is also different than the amount of water that is strongly bound as determined by fitting the D'Arcy-Watt equation (C). (B) shows the results of fitting various (semi-) empirical equations: Chung-Pfost (CP), modified-Oswin (MO) and Peleg (P). As well as taking different forms, it can be seen how the fit is improved when there are more parameters in the equation (the Peleg equation has four, whereas the other two equations have two). In (D), the results of fitting Cromarty's equation to the same data (limited a_w range) is shown, with oil content (OC) as a parameter that is estimated. This is not how Cromarty's equation is normally used. Rather, it is used to predict the moisture content after drying at a particular relative humidity and temperature, depending on seed oil content. This equation illustrates how the isotherm shifts to lower moisture contents as oil content increases [inset graph, for seeds with oil contents of 0.05, 0.20, and 0.35 (dry weight basis); axes as for main graph].