Incoprea Research · Original hypothesis
Reproduced supplied model · 30 October 2026 · Not peer reviewed
In the supplied ideal salt-solution calculation, recovering a larger share of the feed as fresh water increases reversible work per unit of product. The reproduced values are approximately 0.825 kWh/m³ in the vanishing-recovery limit and 1.143 kWh/m³ at 50% recovery. These are conditional model values, not a universal energy floor for every seawater source or desalination process.
The posted hypothesis predicts that specific separation energy rises with recovery. DeepSeek tested that prediction using an idealized NaCl solution, rather than plant measurements. Our review reproduces the supplied calculation and preserves its useful thermodynamic comparison while correcting the reach of the headline.
Minimum separation work is an established thermodynamic benchmark. Mistry and colleagues' original analysis, also available through MIT's research repository, explains why least work and irreversibilities need to be distinguished when evaluating desalination. This note does not introduce a new thermodynamic law.
The feed contains 35 grams of NaCl per litre at 298.15 K. The calculation uses molar mass 58.44 g/mol, gas constant 8.314 J/(mol K), and a van't Hoff factor of two, representing complete dissociation into two solute particles. These are fixed inputs, not salinity or temperature observations from a treatment plant.
The resulting ideal osmotic pressure is π0 = cRT, approximately 29.69 bar. Dividing pressure by 3.6 million converts joules per cubic metre into kWh/m³, giving 0.8248 kWh/m³ before recovery effects. The model assumes pure product water, retained salt, constant temperature and additive solution volumes. It ignores activities, precipitation and concentration-dependent properties.
Let r be the product-water volume divided by the initial feed volume. After an intermediate fraction x has been removed, the remaining solution volume is proportional to 1 − x. Under the stated ideal assumptions its osmotic pressure is π0/(1 − x).
Integrating the reversible work from zero removal to r gives π0[−ln(1 − r)] per unit initial feed. Dividing by the product volume gives W(r) = π0[−ln(1 − r)]/r. The distinction between feed volume and product volume matters: it explains the factor of r in the denominator. The expression is not simply an unspecified ratio of pressure to recovery.
The dimensionless factor −ln(1 − r)/r approaches one as recovery tends to zero and increases thereafter. At 50% recovery it equals 2 ln 2, about 1.3863. Thus the supplied model requires about 38.6% more reversible work per product volume at 50% recovery than in its vanishing-recovery limit.
The supplied script evaluates 201 recovery fractions from 0.02 to 0.95. Running the unchanged calculation in an isolated copy reproduces both its results JSON and its CSV exactly. Its selected result at 50% is 1.143 kWh/m³ when rounded to three decimals; the package's prose summary instead says 1.144. This note uses the reproducible code value.
At 90% recovery the formal ideal expression gives approximately 2.110 kWh/m³. That number must be read as an extrapolation of this equation. It does not validate ideal dilute-solution behavior at the resulting concentrated brine, much less predict a real plant's electricity consumption. The accompanying public code labels it accordingly.
The equation diverges as r approaches one. That mathematical limit assumes the ideal model can be extended indefinitely. It is not evidence that work diverges at a physical saturation concentration: finite solubility and phase changes are absent from the model. The original saturation wording is therefore excluded.
The defensible result is that recovery increases specific reversible work within this defined ideal-solution model. Real feed chemistry, temperature, pressure recovery, membrane behavior and energy accounting require additional treatment. DeepSeek supplied the calculation; Codex verified it and narrowed the interpretation. The standalone standard-library code uses no external data, and no novelty, plant validation or external peer review is claimed.