Fig. 2: Performance assessment of single-pass thermodiffusive desalination.

a The concentration difference ΔC between the top and bottom collection bottles as a function of the average measured temperature difference along the channel ΔT_meas for different mean temperatures T_mean. The experimental results (solid markers) are plotted with horizontal and vertical error bars representing the standard deviation of ΔT measurements (with six thermocouple pairs) and the uncertainty in ΔC measurements (with phase-shifting interferometry, PSI). The insets are PSI images following the experimental procedure reported in^35,44. CFD results (dashed lines) with experimental temperature values as boundary conditions are included. Monotonic dependencies are revealed. b The effect of flow rate Q on ΔC is reported for T_mean ≈ 41 °C and ΔT_meas ≈ 34 K both experimentally and numerically. The errors in the experimental ΔC are standard deviations from PSI measurements. CFD Case 1 uses the thermocouple temperatures measured in each experiment as boundary conditions, whereas CFD Case 2 uses the common wall temperature profile of the experiment with lowest volumetric flow rate of 1 mL min⁻¹. ΔC remains roughly the same when Q < 5 mL min⁻¹. The effect of slight variation in the temperature field on the separation is prominent for small Q. c CFD Case 1 concentration contours for different ΔT_meas and Q. Larger ΔC are observed for greater ΔT_meas and smaller Q. d CFD computation of the concentration difference between collection bottles. Temperature-dependent S_T yields a quasi-linear concentration profile within the cell (d1), together with the parabolic velocity profile (d2), yields the ion mass flux as a function of the height (d3). Dividing the integrated ion mass flux in the top and bottom halves with the corresponding flow rate yields the concentration at the collection bottles.