Heats of transport of some chlorides in H2O and D2O

On the time-dependent electrolyte Seebeck effect

“…( 9) for α ± is probably not accurate when D + ≈ D − , as the term 1/(D + − D − ) becomes very sensitive to uncertainty (and ρ s -dependence) in the ionic diffusivities. Interestingly, the aforementioned works [7,16,17,45] that determined α + +α − from S early − S late did not suffer from the same diverging factor, probably because these authors included into Eq. ( 1) only the cationic species with which their electrodes reacted electrochemically.…”

“…The above-mentioned literature values for Q * i essentially all derive from Soret effect experiments that measured the sum α + +α − [1,7,16,17,45]. From these data, the single-ion Soret coefficients α + and α − were determined either by (arbitrarily) setting α Cl = 0 as a reference point [1] or by the "reduction rule" of Takeyema and Nakashima [18].…”

“…See the supplementary material for a derivation of Eqs. ( 1) and (7); for a detailed schematic of our exper-imental setup; for a figure of V T (t) for several heatingcooling cycles of TBAOH, KCl, and Li 2 SO 4 ; and for a figure showing S late vs. S early using Q * i data of Ref. [6].…”

“…For electrons in a metal, such movement underlies the Peltier-Seebeck effect, which is used in solid-state devices that refrigerate, measure temperature, or harvest thermal energy; For ions in a nonisothermal fluid, thermodiffusion underlies the analogous electrolyte Seebeck effect [1,2]. Microscopically, ionic thermodiffusion in electrolytes has been ascribed to ionion interactions [3] and to the dynamics and structure of the surrounding fluid [1,[4][5][6][7][8][9]. On mesoscopic length scales, ionic thermodiffusion can perturb the salt density c(x) = ρ + (x) + ρ − (x)-the Soret effect-with ρ ± (x) being the local ionic densities of a binary electrolyte, leading to convection [10,11] and variations in the electrolyte's conductivity [12][13][14] and refractive index [8,9,15].…”

“…This theoretical finding was at odds, however, with experiments that found that V T (t), indeed, develops fast at first [faster than the experimental resolution (seconds)], but then evolves with the slow timescale τ dif [27,37]. Many authors, therefore, explained the dynamics of V T (t) in terms of the time-dependent salt density c(x, t) [7,16,17,27,33,[45][46][47], shown by Bierlein [48] to evolve with τ dif (see also Refs. [49,50]).…”