Introduction & Context

The Brunauer–Emmett–Teller (BET) isotherm is the standard model for monolayer-multilayer vapour sorption on solid surfaces. In process engineering it is used to estimate the monolayer moisture content \(X_\mathrm{m}\) (g water per g dry matter) and the sorption energy constant \(C\) from experimental water activity (\(a_\mathrm{w}\)) and equilibrium moisture (\(X\)) data. Knowledge of \(X_\mathrm{m}\) is critical for dryer design, packaging selection, shelf-life prediction and for establishing critical control points in food and pharmaceutical processes.

Methodology & Formulas

Transform experimental data into the BET linear coordinate
\[ \Phi = \frac{a_\mathrm{w}}{X\,(1-a_\mathrm{w})} \] where \(a_\mathrm{w}\) is the water activity (dimensionless) and \(X\) is the corresponding moisture content (g water g⁻¹ dry matter).
Perform ordinary least-squares regression
Fit the straight line \[ \Phi = m\,a_\mathrm{w} + b \] with slope \(m\) and intercept \(b\) obtained from the usual linear regression formulas: \[ m = \frac{n\sum a_\mathrm{w}\Phi - \sum a_\mathrm{w}\sum\Phi} {n\sum a_\mathrm{w}^{2} - (\sum a_\mathrm{w})^{2}}, \qquad b = \frac{\sum\Phi - m\sum a_\mathrm{w}}{n} \] where \(n\) is the number of data points.
Extract BET parameters
Monolayer moisture content: \[ X_\mathrm{m} = \frac{1}{m + b} \] Sorption energy constant: \[ C = \frac{m}{b} + 1 \]

Regime	Water-activity range	Model validity
BET region	0.05 ≤ a_w ≤ 0.45	Linear transform accurate within ±5 %
Capillary condensation	a_w > 0.45	Deviation from linearity; GAB or other models recommended

The Brunauer–Emmett–Teller (BET) isotherm describes multilayer adsorption of water on solid surfaces. In process engineering it is used to predict water activity (a_w) by relating the equilibrium moisture content (X) to the relative pressure (p/p₀) through the equation:

a_w = p/p₀ = (X·(1−(n+1)·(X/X_m)ⁿ+n·(X/X_m)ⁿ⁺¹)) / (X_m·(1−(X/X_m)))
X_m is the monolayer moisture content and n is the number of adsorbed layers; both are obtained by nonlinear regression of experimental sorption data.

Once X_m and n are known, the model predicts a_w at any moisture content, allowing engineers to design drying or hydration steps without direct a_w measurement.

Collect at least 8–10 equilibrium points covering the water activity range 0.05–0.45 (below capillary condensation). Required data:

Moisture content X (kg water / kg dry solids) determined gravimetrically.
Corresponding water activity a_w measured at constant temperature (20–40 °C) using a chilled-mirror dew-point or capacitive sensor.
Temperature must be constant; if seasonal variation is expected, generate separate isotherms at 5 °C intervals.

Dry the sample to <0.02 a_w before starting the adsorption branch to avoid hysteresis errors.

Apply the BET transformation plot:

Plot [a_w / (X·(1−a_w))] versus a_w in the range 0.05–0.35.
A linear regression R² ≥ 0.995 indicates good fit; slopes outside this range suggest capillary effects or dissolution invalidating the model.
Compare predicted X_m with the moisture content at which the product first shows maximum stability (often the minimum in a chemical-rate versus a_w plot).

If the linear region is narrow or R² < 0.98, switch to GAB or Langmuir-type models.

Yes; the monolayer moisture content X_m corresponds to the tightest physically adsorbed water layer, giving the lowest molecular mobility. Industry practice:

Target process a_w ≤ the a_w at X_m plus 0.05 to account for measurement uncertainty.
For enzyme or vitamin degradation, verify that reaction rate minima coincide with X_m; if not, tighten the specification to 0.9·X_m.
Package the product in a moisture barrier whose transmission rate keeps the a_w below the specification for the desired shelf-life under worst-case temperature/humidity.

Worked Example – Estimating Water Activity with the BET Isotherm

A snack-food plant needs to predict the safe equilibrium moisture content for a new baked chip that will be packed at 20 °C. Three experimental data points (water activity vs. moisture content) are available from the lab. The engineering team will use the BET model to obtain the monolayer moisture \(X_m\) and the constant \(C\), then predict sorption behaviour at other humidities.

Knowns

Temperature: 20 °C (293.15 K)
Experimental point 1: \(a_{w1}=0.10\), \(X_1=0.045\) kg water kg⁻¹ dry solids
Experimental point 2: \(a_{w2}=0.25\), \(X_2=0.072\) kg water kg⁻¹ dry solids
Experimental point 3: \(a_{w3}=0.40\), \(X_3=0.095\) kg water kg⁻¹ dry solids

Step-by-step calculation

Transform each data pair into the BET linear co-ordinates: \[ \Phi_i=\frac{a_{w,i}}{X_i(1-a_{w,i})} \] \[ \Phi_1=\frac{0.10}{0.045(1-0.10)}=2.469 \] \[ \Phi_2=\frac{0.25}{0.072(1-0.25)}=4.630 \] \[ \Phi_3=\frac{0.40}{0.095(1-0.40)}=7.018 \]
Prepare data for linear regression (slope–intercept form \(\Phi = \text{slope}\cdot a_w + \text{intercept}\)): \[ n=3,\quad \sum a_w=0.75,\quad \sum\Phi=14.116 \] \[ \sum a_w^2=0.233,\quad \sum a_w\Phi=4.211 \]
Compute slope \(m\) and intercept \(b\): \[ m=\frac{n\sum a_w\Phi-\sum a_w\sum\Phi}{n\sum a_w^2-(\sum a_w)^2}=15.161 \] \[ b=\frac{\sum\Phi-m\sum a_w}{n}=0.915 \]
Extract BET parameters: \[ X_m=\frac{1}{m+b}=\frac{1}{15.161+0.915}=0.062\ \text{kg water kg⁻¹ dry solids} \] \[ C=\frac{m}{b}+1=\frac{15.161}{0.915}+1=17.568 \]

Final Answer

The BET model gives a monolayer moisture content \(X_m=0.062\) kg water kg⁻¹ dry solids and a constant \(C=17.568\) at 20 °C. These values can now be used to predict equilibrium moisture for any water activity below 0.45 using the standard BET equation.

"Un projet n'est jamais trop grand s'il est bien conçu." — André Citroën

"La difficulté attire l'homme de caractère, car c'est en l'étreignant qu'il se réalise." — Charles de Gaulle