A Heat Transfer Textbook

Sixth Edition

Solutions Manual for Chapter 11

by

John H. Lienhard V
and
John H. Lienhard IV

Phlogiston
Cambridge
Press Massachusetts
Professor John H. Lienhard V
Department of Mechanical Engineering
Massachusetts Institute of Technology
77 Massachusetts Avenue
Cambridge MA 02139-4307 U.S.A.

Professor John H. Lienhard IV

Department of Mechanical Engineering
University of Houston
4800 Calhoun Road
Houston TX 77204-4792 U.S.A.

Copyright ©2024 by John H. Lienhard V and John H. Lienhard IV

All rights reserved
Please note that this material is copyrighted under U.S. Copyright Law. The
authors grant you the right to download and print it for your personal use or
for non-profit instructional use. Any other use, including copying, distributing
or modifying the work for commercial purposes, is subject to the restrictions
of U.S. Copyright Law. International copyright is subject to the Berne
International Copyright Convention.
The authors have used their best efforts to ensure the accuracy of the methods,
equations, and data described in this book, but they do not guarantee them for
any particular purpose. The authors and publisher offer no warranties or
representations, nor do they accept any liabilities with respect to the use of
this information. Please report any errata to the authors.

Names: Lienhard, John H., V, 1961– | Lienhard, John H., IV, 1930–.
Title: A Heat Transfer Textbook: Solutions Manual for Chapter 11
/ by John H. Lienhard, V, and John H. Lienhard, IV.
Description: Sixth edition | Cambridge, Massachusetts :
Phlogiston Press, 2024 | Includes bibliographical references
and index.
Subjects: Heat—Transmission | Mass Transfer.

Published by Phlogiston Press

Cambridge, Massachusetts, U.S.A.

For updates and information, visit:

http://ahtt.mit.edu

This copy is:

Version 1.07 dated 3 April 2024
 Problem 11.1 Derive eqns. (11.9).

Solution Using eqns. (11.2,4,6,8):

𝜌𝑖 𝑐 𝑀 𝑥 𝑖 𝑀𝑖
𝑚𝑖 = = 𝑖 𝑖 =
𝜌 𝑐 𝑀 ∑𝑘 𝑥𝑘 𝑀𝑘
𝑐𝑖 𝜌 𝑀 𝑚𝑖 /𝑀𝑖
𝑥𝑖 = = 𝑖 =
𝑐 𝜌 𝑀𝑖 ∑𝑘 𝑚𝑘 /𝑀𝑘

Problem 11.2 A 1000 liter cylinder at 300 K contains a gaseous mixture composed of
0.10 kmol of NH3 , 0.04 kmol of CO2 , and 0.06 kmol of He. (a) Find the mass fraction for
each species and the pressure in the cylinder. (b) After the cylinder is heated to 600 K, what are the
new mole fractions, mass fractions, and molar concentrations? (c) The cylinder is now compressed
isothermally to a volume of 600 liters. What are the molar concentrations, mass fractions, and
partial densities? (d) If 0.40 kg of N2 is injected into the cylinder while the temperature remains at
600 K, find the mole fractions, mass fractions, and molar concentrations.

Solution
a) By eqn. (11.6), noting that the total number of moles is 0.1 + 0.04 + 0.06 = 0.2 kmol,
𝑥NH3 = 0.1/0.2 = 0.5
𝑥CO2 = 0.04/0.2 = 0.2
𝑥He = 0.06/0.2 = 0.3
and with eqn. (11.8)
𝑀 = (0.5)(17.03) + (0.2)(44.01) + (0.3)(4.003) = 18.52 kg/kmol
Thus, from eqn. (11.9)
𝑚NH3 = (0.5)(17.03)/(18.52) = 0.460
𝑚CO2 = (0.2)(44.01)/(18.52) = 0.475
𝑚He = (0.3)(4.003)/(18.52) = 0.0648
The pressure, from eqn. (11.13), is
𝑝 = 𝑐𝑅∘ 𝑇 = (0.2 kmol/m3 )(8314.5 J/kmol⋅K)(300 K)
= 4.99 × 105 Pa

b) The mass fractions are unchanged. The total pressure, like the temperature, is doubled,
𝑝 = 2(4.99 × 105 ) = 9.98 × 105 Pa. Then, with eqn. (11.15),
𝑝NH3 = (0.5)(9.98 × 105 ) = 4.99 × 105 Pa
𝑝CO2 = (0.2)(9.98 × 105 ) = 2.00 × 105 Pa
𝑝He = (0.3)(9.98 × 105 ) = 2.99 × 105 Pa

312
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 c) The mass fractions are still unchanged. The molar concentrations are

𝑐NH3 = 0.1 kmol/0.6 m3 = 0.167 kmol/m3

𝑐CO2 = 0.04 kmol/0.6 m3 = 0.0667 kmol/m3
𝑐He = 0.06 kmol/0.6 m3 = 0.100 kmol/m3
From eqn. (11.4), the partial densities are
𝜌NH3 = (0.167)(17.03) = 2.84 kg/m3
𝜌CO2 = (0.0667)(44.01) = 2.94 kg/m3
𝜌He = (0.100)(4.003) = 0.400 kg/m3

d) We have added (0.4 kg)/(28.01 kg/kmol) = 0.0143 kmol N2 .

𝑥NH3 = 0.1/0.2143 = 0.467
𝑥CO2 = 0.04/0.2143 = 0.187
𝑥He = 0.06/0.2143 = 0.280
𝑥He = 0.0143/0.2143 = 0.0667
The molar mass of the mixture is
𝑀 = (0.467)(17.03) + (0.187)(44.01) + (0.280)(4.003) + (0.0667)(28.01)
= 19.17 kg/kmol
Thus,
𝑚NH3 = (0.467)(17.03)/(19.17) = 0.415
𝑚CO2 = (0.187)(44.01)/(19.17) = 0.429
𝑚He = (0.280)(4.003)/(19.17) = 0.0585
𝑚N2 = (0.0667)(28.01)/(19.17) = 0.0975
The molar concentrations calculated in part (c) are unchanged. For N2 :

𝑐He = 0.0143 kmol/0.6 m3 = 0.0238 kmol/m3

Problem 11.3 The pressure of Jupiter’s atmosphere increases with depth. The famous
clouds of Jupiter are in a layer called the troposphere, in which the pressure rises from
0.1 bar to 10 bar. The top of the troposphere, called the tropopause, is about 50 km above
the clouds and is at 0.1 bar and 110 K. The atmospheric mole fractions of hydrogen, helium,
and methane are 𝑥H2 = 0.86, 𝑥He = 0.136, and 𝑥CH4 = 0.0018. Other species have small but
localized concentrations, e.g., in the troposphere’s clouds of ammonia ice. (a) Calculate the
molar concentrations and the partial densities of H2 , He, and CH4 at the tropopause. (b) Find
the number of hydrogen atoms per unit volume (number density), 𝒩H2 at the tropopause.

313
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 (c) Estimate 𝒩H2 at the base of the Jovian troposphere, where the pressure is 10 bar and the
temperature is 340 K.

Solution
(a) From eqn. (11.13)
𝑝 0.1 × 105 Pa
𝑐= = = 0.01093 kmol/m3
𝑅∘ 𝑇 (8314.5 J/kmol⋅K)(110 K)
and with 𝑥𝑖 = 𝑐𝑖 /𝑐,
𝑐H2 = (0.86)(0.01093) = 0.00940 kmol/m3 = 9.40 mol/m3
𝑐He = (0.136)(0.01093) = 0.00149 kmol/m3 = 1.49 mol/m3
𝑐CH4 = (0.0018)(0.01093) = 0.0000197 kmol/m3 = 0.0197 mol/m3
With eqn. (11.4),
𝜌𝑖 = 𝑀𝑖 𝑐𝑖
and molar concentrations in kmol/m3
𝜌H2 = (2.016)(0.00940) = 0.0190 kg/m3
𝜌He = (4.003)(0.00149) = 0.00596 kg/m3
𝜌CH4 = (16.04)(0.0000197) = 0.000316 kg/m3
where the molar masses, in kg/kmol, can be taken from Table 11.3 or elsewhere.
(b) The number of molecules of species 𝑖 per unit volume is 𝑁𝐴 𝑐𝑖 where 𝑁𝐴 is Avogadro’s
number. So, working in mol,
𝒩H2 = (6.0221 × 1023 molecules/mol)(9.40 mol/m3 )
= 5.66 × 1024 molecules/m3
(c) From eqn. (11.13),
𝑝 10 × 105
𝑐= = = 0.354 kmol/m3
𝑅∘ 𝑇 (8314.5)(340)
If we estimate the mole fraction to be the same as for the upper atmosphere, and now
working in kmol
𝒩H2 = 𝑁𝐴 𝑥H2 𝑐
= (6.0221 × 1026 molecules/kmol)(0.86)(0.354 kmol/m3 )
= 1.83 × 1026 molecules/m3
Misc. facts about Jupiter’s atmosphere: The atmosphere of Jupiter continues smoothly
from the gas phase to a supercritical fluid phase as the pressure rises. Ammonia is present
in the troposphere, and condenses to form clouds. Water is present deeper in the tropo-
sphere. Neon is dissolved into liquid helium and transported downward by helium rain.
Photochemistry in the upper atmosphere can produce ethane and ethyne. Elements present
in the upper atmosphere are fully reduced by combination with hydrogen. Convection in
the troposphere transports some molecules vertically, giving rise to geographic patterns of
molecular abundance.

314
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 References:
F. W. Taylor, S. K. Atreya, Th. Encrenaz, D. M. Hunten, P. G. J. Irwin, T. C. Owen,
The Composition of the Atmosphere of Jupiter. In Jupiter: The Planet, Satellites and
Magnetosphere, F. Bagenal, et al., eds., Cambridge Univ. Press, pp. 59-78, 2004. See Table
4.2.
S. K. Atreya, M. H. Wong, T. C. Owen, P. R. Mahaffy, H. B. Niemann, I. de Pater,
P. Drossart, Th. Encrenaz, A comparison of the atmospheres of Jupiter and Saturn: deep
atmospheric composition, cloud structure, vertical mixing, and origin. Planetary and Space
Science, 47(10–11): 1243–1262, 1999. https://doi.org/10.1016/S0032-0633(99)00047-1.
D. Grassi, A. Adriani, A. Mura, S. K. Atreya, L. N. Fletcher, J. I. Lunine, et al. On the
spatial distribution of minor species in Jupiter’s troposphere as inferred from Juno JIRAM
data. J. Geophysical Research: Planets, 125:e2019JE006206, 2020.
https://doi.org/10.1029/2019JE006206

Problem 11.4 In Example (11.2), suppose that the only gases at the 𝑠-surface are CO,
O2 , and N2 . As before, assume that 𝑚O2,𝑠 is very small. Find 𝑗N2,𝑠 , 𝑛N2,𝑠 , and 𝑚N2,𝑠 .

Solution From eqn. (11.21)

∑ 𝑗𝑖⃗ = 0 = 𝑗CO,𝑠 + 𝑗O2,𝑠 + 𝑗N2,𝑠
𝑖
Answer
𝑗N2,𝑠 = −𝑗CO,𝑠 − 𝑗O2,𝑠 = −45.7 + 26.7 = −19.0 g/m2 s ⟵−−−−−−−
Answer
No N2 flows through the 𝑠-surface or reacts there, so 𝑛N2,𝑠 = 0. ⟵−−−−−−−
The diffusion flux of N2 toward the wall simply opposes the mass-average velocity away
from the wall. Therefore, with eqn. (11.20)
𝑛N2,𝑠 = 0 = 𝑚N2,𝑠 𝑛𝑠 + 𝑗N2,𝑠
𝑗N ,𝑠 19.0
𝑚N2,𝑠 = − 2 = = 0.95
𝑛𝑠 20.0
Another way to get the same answer is with eqn. (11.3):
Answer
𝑚N2,𝑠 = 1 − 𝑚CO,𝑠 − 𝑚O2,𝑠 = 1 − 0.05 − 0 = 0.95 ⟵−−−−−−−

Problem 11.5 Small automobile batteries commonly contain an aqueous solution of

sulfuric acid (H2 SO4 ) with lead and lead oxide plates as electrodes. Current is generated by
the reaction of the electrolyte with the electrode material. The reaction at the lead electrode
(the anode) is
Pb(𝑠) + HSO4 − ⟶ PbSO4 (𝑠) + H+ + 2𝑒−
where the (𝑠) denotes a solid phase component. If the current density at such an electrode is
𝑖 = 5 mA/cm2 , what is the mole flux of HSO4 – (bisulfate) to the electrode? Recall that the

315
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 charge of an electron is −1.609 × 10−19 C and that 1 A = 1 C/s. What is the mass flux of
HSO4 – ? At what mass rate is PbSO4 produced? A what rate does H+ flow away from the
electrode?

Solution Two electrons are released for each mole of bisulfate reaching the electrode,
so we may write the current in terms of the mole flux of HSO4 – through an 𝑠 -surface just
over the electrode:
𝑖 = 5 mA/cm2 = 2(1.609 × 10−19 C)(6.0221 × 1026 kmol/kg)(𝑁HSO−4 ,𝑠 kmol/m2 s)
Solving
(5 × 10−3 /10−4 ) A/m2
𝑁HSO−4,𝑠 =
2(1.609 × 10−19 )(6.0221 × 1026 )
Answer
= 2.580 × 10−7 kmol/m2 s ⟵−−−−−−−
The mass flux is just
𝑛HSO−4,𝑠 = 𝑀HSO−4 𝑁HSO−4,𝑠 = (97.06 kg/kmol)(2.580 × 10−7 kmol/m2 s)
Answer
= 2.504 × 10−5 kg/m2 s ⟵−−−−−−−
The electrode reaction creates one mole of PbSO4 for each mole of HSO4 – reaching the
electrode, so
𝑟′′PbSO
̇ 4
= 𝑀PbSO2−
4
𝑁HSO−4,𝑠
= (303.3 kg/kmol)(2.580 × 10−7 kmol/m2 s)
Answer
= 7.825 × 10−5 kg/m2 s ⟵−−−−−−−
H+ ions flow away at the rate of one mole per mole of HSO4 – , so
Answer
𝑁H+,𝑠 = −𝑁HSO−4,𝑠 = −2.580 × 10−7 kmol/m2 s ⟵−−−−−−−

Problem 11.6 In catalysis, one gaseous species reacts with another on the surface of a
catalyst to form a gaseous product. For example, butane (C4 H10 ) reacts with hydrogen on the
surface of a nickel catalyst to form propane (C3 H8 ) and methane (CH4 ). This heterogeneous
reaction, referred to as hydrogenolysis, is
Ni
C4 H10 + H2 ⟶ C3 H8 + CH4
The molar rate of consumption of C4 H10 per unit area in the reaction is
∘
𝑅̇C H = 𝐴(𝑒−Δ𝐸/𝑅 𝑇 )𝑝C H 𝑝−2.4
4 10 H 4 10 2

where 𝐴 = 2.9 × 10 kmol/m s, Δ𝐸 = 1.9 × 10 J/kmol, and 𝑝𝑖 is in atm.

9 2 8

(a) If 𝑝C4H10,𝑠 = 𝑝C3H8,𝑠 = 0.2 atm, 𝑝CH4,𝑠 = 0.17 atm, and 𝑝H2,𝑠 = 0.3 atm at a nickel
surface with conditions of 440°C and 0.87 atm total pressure, what is the rate of con-
sumption of butane?
(b) What are the mole fluxes of butane and hydrogen to the surface? What are the mass
fluxes of propane and methane away from the surface?

316
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 (c) What is 𝑣C4H10,𝑠 What are 𝑣𝑠 and 𝑣𝑠∗ ?
(d) What is the diffusional mole flux of butane? What is the diffusional mass flux of
propane? What is the flux of Ni?

Solution
(a)
1.9 × 108
𝑅̇C4H10 = (2.9 × 109 ) exp[− ](0.2)(0.3)−2.4
(8314.5)(440 + 273.15)
Answer
= 1.27 × 10−4 kmol/m2 s ⟵−−−−−−−
(b) Let a negative flux be toward the surface and a positive flux be away from the surface.
Then, from stoichiometry:
Answer
𝑅̇C4H10 = −𝑁C4H10 = −𝑁H2,𝑠 = +𝑁C3H8,𝑠 = +𝑁CH4,𝑠 ⟵−−−−−−−
The mass fluxes are 𝑛𝑖 = 𝑀𝑖 𝑁𝑖 . With 𝑀C3H8 = 44.09 kg/kmol and 𝑀CH4 = 16.04 kg/kmol,
we find
𝑛C3H8,𝑠 = (44.09)(1.27 × 10−4 ) = 5.58 × 10−3 kg/m2 s
Answer
= 5.58 g/m2 s ⟵−−−−−−−
𝑛CH4,𝑠 = (16.04)(1.27 × 10−4 ) = 2.03 × 10−3 kg/m2 s
Answer
= 2.03 g/m2 s ⟵−−−−−−−
(c) With eqns. (11.17) and (11.4)
𝑣C4H10,𝑠 = 𝑛C4H10,𝑠 /𝜌C4H10,𝑠 = 𝑁C4H10,𝑠 /𝑐C4H10,𝑠
With eqn. (11.12)
𝑝C H ,𝑠 (0.2)(101325)
𝑐C4H10,𝑠 = 4∘ 10 = = 3.418 × 10−3 kmol/m2
𝑅𝑇 (8314.5)(440 + 273.15)
so the species average speed, which is negative because it is toward the surface, is
−1.27 × 10−4 Answer
𝑣C4H10,𝑠 = = −0.00372 m/s ⟵−−−−−−−
3.418 × 10−3
Answer
No net mass flux occurs at the surface of the catalyst, so 𝑣 = 𝑣∗ = 0. ⟵−−−−−−−
(d) Because 𝑣 = 𝑣∗ = 0, 𝑛 = 𝑁 = 0 and eqns. (11.20) and (11.24) show that the total and
diffusional mass and mole fluxes are equal. Nickel is a catalyst and has no flux.

Comment: In these experiments, the catalyst material contained 5 wt% Ni and had a BET
area of 240 m2 per gram of catalyst. The reported reaction rates were given in mol h−1 g−1
Ni .

Reference:
G. Leclercq, L. Leclercq, L.M. Bouleau, S. Pietrzyk, R. Maurel, Hydrogenolysis of sat-
urated hydrocarbons: IV. Kinetics of the hydrogenolysis of ethane, propane, butane, and
isobutane over nickel, Journal of Catalysis, 88(1):8–17, 1984, https://doi.org/10.1016/0021-
9517(84)90044-7.

317
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.7 Show that 𝒟12 = 𝒟21 in a binary mixture.

Solution From eqn. (11.21), 𝑗1⃗ + 𝑗2⃗ = 0, and so, with eqn. (11.26)
𝑗1⃗ = −𝑗2⃗
−𝜌𝒟12 ∇𝑚1 = +𝜌𝒟21 ∇𝑚2
−𝒟12 ∇𝑚1 = 𝒟21 ∇𝑚2
But with eqn. (11.3)
∇(𝑚1 + 𝑚2 ) = ∇(1) = 0
so −∇𝑚1 = −∇𝑚2 . Substituting into the previous equation
−𝒟12 ∇𝑚1 = −𝒟21 ∇𝑚1
Answer
𝒟12 = 𝒟21 ⟵−−−−−−−

Problem 11.8 Using the definitions of the fluxes, velocities, and concentrations, derive
eqn. (11.30) from eqn. (11.26) for binary diffusion.

Solution From eqns. (11.26), (11.19), (11.30) and (11.23)

𝑗1⃗ = −𝜌𝒟12 ∇𝑚1 = 𝜌1 (𝑣1⃗ − 𝑣)⃗ (*)
𝐽1⃗ = −𝑐𝒟12 ∇𝑥1 = 𝑐1 (𝑣1⃗ − 𝑣∗⃗ )
With eqns. (11.16), (11.3), (11.22), and (11.7)
𝑣1⃗ − 𝑣 ⃗ = 𝑣1⃗ − (𝑚1 𝑣1⃗ + 𝑚2 𝑣2⃗ ) = 𝑚2 (𝑣1⃗ − 𝑣2⃗ )
𝑣1⃗ − 𝑣∗⃗ = 𝑣1⃗ − (𝑥1 𝑣1⃗ + 𝑥2 𝑣2⃗ ) = 𝑥2 (𝑣1⃗ − 𝑣2⃗ )
Thus
𝑀
𝐽1⃗ = 𝑐1 𝑥2 (𝑣1⃗ − 𝑣2⃗ ) = 𝑐1 (𝑚1 𝑀/𝑀2 )(𝑣1⃗ − 𝑣2⃗ ) = 𝑗⃗ (†)
𝑀1 𝑀2 1
where we also use eqn. (11.9).
Next, we need to express 𝑗1⃗ in terms of ∇𝑥1 . With eqn. (11.8)
𝑥 1 𝑀1 𝑀 𝑥𝑀
∇𝑚1 = ∇( ) = 1 ∇𝑥1 − 1 2 1 ∇𝑀
𝑀 𝑀 𝑀
𝑀1 𝑥 1 𝑀1
= ∇𝑥1 − (𝑀1 ∇𝑥1 + 𝑀2 ∇𝑥2 )
𝑀 𝑀2
and with ∇𝑥2 = −∇𝑥1 , 𝑀 = 𝑥1 𝑀1 + 𝑥2 𝑀2 , and 𝑥1 + 𝑥2 = 1
𝑀1 𝑀𝑀
= 2
(𝑀 − 𝑥1 𝑀1 + 𝑥1 𝑀2 ) ∇𝑥1 = 1 2 2 ∇𝑥1 (**)
𝑀 𝑀

318
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Hence, upon substituting eqns. (*) and (**) into eqn. (†), we find our result:
𝑀
𝐽1⃗ = (−𝜌𝒟12 ∇𝑚1 )
𝑀1 𝑀2
𝑀 2 𝑀1 𝑀2
= −𝑐𝒟12 ( ∇𝑥1 )
𝑀1 𝑀2 𝑀 2
Answer
= −𝑐𝒟12 ∇𝑥1 ⟵−−−−−−−

Problem 11.9 Fill in the details involved in obtaining eqn. (11.33) from eqn. (11.32).

Solution
𝑗𝐴 || = 𝜂𝜌𝐶 (𝑚𝐴 || − 𝑚𝐴 || ) (11.32)
𝑥0 𝑥0 −𝑎ℓ 𝑥0 +𝑎ℓ
We may use a Taylor expansion of 𝑚𝐴 (𝑥) about 𝑥 = 𝑥0 :
𝑑𝑚𝐴 |
𝑚𝐴 || = 𝑚𝐴 || + | (𝑥 + 𝑎ℓ − 𝑥0 ) + ⋯
𝑥0 +𝑎ℓ 𝑥0 𝑑𝑥 |𝑥 0
0

Likewise
𝑑𝑚𝐴 |
𝑚𝐴 || = 𝑚𝐴 || + | (𝑥 − 𝑎ℓ − 𝑥0 ) + ⋯
𝑥0 −𝑎ℓ 𝑥0 𝑑𝑥 |𝑥 0
0
Substituting these into eqn. (11.32)
𝑑𝑚𝐴 |
𝑗𝐴 || = 𝜂𝜌𝐶 (−2 | (𝑎ℓ))
𝑥0 𝑑𝑥 |𝑥
0

which we can rearrange as

𝑑𝑚𝐴 | Answer
𝑗𝐴 || = −𝜌(2𝜂𝑎𝐶ℓ) | ⟵−−−−−−− (11.33)
𝑥0 𝑑𝑥 |𝑥
0

Comment: The second-order terms in the two Taylor series cancel out, so that the error in
the expression for 𝑗𝐴 is 𝒪(ℓ3 ).

319
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 11.10

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 320

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 321
 11.11

(11.37)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 322

 11.12

11.13 (11.44).
(11.44).

(11.40) (11.42)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 323

 Eqn. (11.44)

11.14

(11.44)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 324

 11.15

(11.46)

11.16

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 325

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 326
 Problem 11.17 Consider one-dimensional, binary diffusion in which isothermal ideal gases 1
and 2 travel in opposite directions along the 𝑧-axis of a tube. Assume that the flow has a zero mass-
average velocity, so that mass transfer purely diffusive. (a) Show that the conservation equations
for mass, species, and momentum are satisfied by 𝑣 = 0, 𝑗1 = −𝑗2 = constant, and 𝑝 = constant.
(b) Use Fick’s law to solve for 𝑚1 (𝑧), if 𝑚1 (0) = 0 and 𝑚1 (𝐿) = 1. (c) Find 𝑗1 in terms of the
molar masses of the gases. (Hint: 𝜌 is not constant; use 𝜌 = 𝑀𝑐 and write 𝑀 in terms of 𝑚1 before
integrating). (d) Write the mole fluxes in terms of the mass fluxes and solve for the mole-average
velocity, 𝑣∗ . Under what condition is 𝑣∗ = 0? (e) Suppose that gas 1 is helium and gas 2 is air
at 𝑇 = 276 K and 𝑝 = 1 atm, with 𝐿 = 1 m. Calculate 𝑣∗ , and 𝑗𝑖 and 𝑁𝑖 for both gases. Are the
molar diffusion rates equal and opposite?

Solution
a) In steady, one-dimensional flow, the mass and species equations, (11.40) and (11.41), may
be simplified to
𝑑
(𝜌𝑣) = 0 eqn. (11.41)
𝑑𝑧
𝑑 𝑑
(𝜌𝑖 𝑣 + 𝑗𝑖 ) = (𝑛 ) = 0 eqn. (11.40)
𝑑𝑧 𝑑𝑧 𝑖
When 𝑣 = 0, the first equation is satisfied and 𝑛 = 0. The second equation shows that
𝜌𝑖 𝑣 = 𝑛𝑖 = constant, and with eqns. (11.18) and (11.20)
0
𝑣 + 𝑗𝑖 = 𝑗𝑖 = constant
𝑛𝑖 = 𝜌𝑖
Finally, eqn. (11.21) shows that 𝑗1 = −𝑗2 , as is always the case in binary diffusion.
The momentum equation will be familiar from past studies of fluid mechanics. If the fluid
has zero velocity, and if gravitational pressure gradients are negligible over the height of
the system, then 𝑝 = constant. (If the fluid flowed through the tube, i.e. had a non-zero
mass-average velocity, then viscous drag would produce an axial pressure gradient.)
b) With Fick’s law,
𝑑𝑚1
𝑗𝑖 = −𝜌𝒟12= constant = 𝑐1
𝑑𝑧
At this point, we need to think a little: the mass density will vary with the mass fractions,
even though 𝑝 and 𝑇 are constant. One the other hand, the molar density, from eqn. (11.13)
will be constant: 𝑐 = 𝑝/𝑅∘ 𝑇. So, we need the molar mass of the mixture from eqn. (11.8):
1 𝑚 𝑚 1 1 1 𝑚 (𝑀 − 𝑀1 ) + 𝑀1
= 1 + 2 = 𝑚1 ( − )+ = 1 2 (*)
𝑀 𝑀1 𝑀2 𝑀1 𝑀2 𝑀2 𝑀 1 𝑀2
Now returning to Fick’s law and rearranging
𝑐𝒟12 𝑀1 𝑀2 𝑑𝑚1
− = 𝑐1
𝑚1 (𝑀2 − 𝑀1 ) + 𝑀1 𝑑𝑧
The integral we need, with 𝑚1 (0) = 0, is
𝑚1 (𝑧)
𝑑𝑚1 1 𝑎𝑚 (𝑧)
∫ = ln ( 1 + 1)
0
𝑎𝑚1 + 𝑏 𝑎 𝑏

327
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 where 𝑎 = (𝑀2 − 𝑀1 ) and 𝑏 = 𝑀1 . Collecting all this
1 𝑎𝑚 (𝑧) 𝑐1 𝑧
ln ( 1 + 1) = −
𝑎 𝑏 𝑐𝒟12 𝑀1 𝑀2
and with 𝑚1 (𝐿) = 1, we find
𝑐𝒟12 𝑀1 𝑀2 1 𝑎
𝑐1 = − ln ( + 1)
𝐿 𝑎 𝑏
so that
𝑎𝑚1 (𝑧) 𝑧 𝑎
ln ( + 1) = ln ( + 1)
𝑏 𝐿 𝑏
and a little more algebra gives
1 𝑎 𝑧/𝐿 𝑏
𝑚1 (𝑧) =
( + 1) −
𝑎 𝑏 𝑎
and finally substituting for 𝑎 and 𝑏 gives
(𝑀2 /𝑀1 )𝑧/𝐿 − 𝑀1 Answer
𝑚1 (𝑧) = ⟵−−−−−−−
𝑀2 − 𝑀 1
c) With Fick’s law
𝑑𝑚1 𝑐𝑀𝒟12 𝑀
𝑗1 = −𝑐𝑀𝒟12 =− ln ( 2 )
𝑑𝑧 𝐿(𝑀2 − 𝑀1 ) 𝑀1
and using eqn. (*) to eliminate 𝑀
𝑐𝒟12 𝑀1 𝑀2 𝑀 Answer
𝑗1 = − ln ( 2 ) ⟵−−−−−−−
𝐿(𝑀2 − 𝑀1 ) 𝑀1
d) 𝑁𝑖 = 𝑛𝑖 /𝑀𝑖 = 𝑗𝑖 /𝑀𝑖 and 𝑁 = 𝑛/𝑀 = 0. With eqn. (11.22)
𝒟12 𝑀
𝑣∗ = 𝑐−1 (𝑗1 /𝑀1 + 𝑗2 /𝑀2 ) = 𝑐−1 𝑗1 (1/𝑀1 − 1/𝑀2 ) = − ln ( 2 )
𝐿 𝑀1
The mole-average velocity is only zero if 𝑀1 = 𝑀2 .
e) From Table 11.1, 𝒟He−air = 6.24 × 10−5 m2 /s, and 𝑀He = 4.003 and 𝑀air = 28.96:
(6.24 × 10−5 ) 28.96 Answer
𝑣∗ = − ln ( ) = −1.235 × 10−4 m/s ⟵−−−−−−−
1 4.003
With 𝑐 = (101325)/[(8314.5)(276)] = 0.0442 kmol/m2 ,
(0.0442)(6.24 × 10−5 )(4.003)(28.96) 28.96
𝑗1 = −𝑗2 = − ln ( )
(1)(28.96 − 4.003) 4.003
Answer
= −2.54 × 10−5 kg/m2 s ⟵−−−−−−−

𝑛1 𝑗 2.54 × 10−5 Answer

𝐽1 = = 1 =− = −6.34 × 10−6 kmol/m2 s ⟵−−−−−−−
𝑀1 𝑀1 4.003
𝑛2 𝑗 2.54 × 10−5 Answer
𝐽2 = = 2 = = 8.77 × 10−7 kmol/m2 s ⟵−−−−−−−
𝑀2 𝑀2 28.96

328
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Comment: What should be very clear is that the mole fluxes are entirely unequal. Even so,
many textbooks call this problem “equimolar counterdiffusion” and begin with the assumption that
𝐽1 = −𝐽2 if 𝑝 = constant! Mills [1] discusses this common error in more detail.

Reference:
[1] A.F Mills, On steady one-dimensional diffusion in binary ideal gas mixtures, Int. J. Heat
Mass Transfer, 46(13):2495–2497, 2003, https://doi.org/10.1016/S0017-9310(02)00537-9.

Problem 11.18 Suppose that a steel fitting with a carbon mass fraction of 0.2% is put into
contact with carburizing gases at 910°C, and that these gases produce a steady mass fraction, 𝑚C,ᵆ ,
of 1.0% carbon just within the surface of the metal. The diffusion coefficient of carbon in this
steel is [1]
𝒟C,Fe = (1.2 × 10−5 m2/s) exp[−(1.34 × 108 J/kmol)/(𝑅∘𝑇)]
for 𝑇 in kelvin. How long does it take to produce a carbon concentration of 0.6% by mass at a depth
of 0.5 mm? How much less time would it take if the temperature were 950°C?

Solution
We can refer to Example 11.9 and Fig. 11.12. As shown in the example, the solution for mass
fraction of carbon in the iron is
𝑚C (𝑥, 𝑡) − 𝑚C,ᵆ 𝑥
= erf ( )
𝑚C,0 − 𝑚C,ᵆ 2√𝒟C,Fe 𝑡
The boundary conditions are:
𝑚C,ᵆ = 0.010 𝑚C,0 = 0.002
To find the time for 𝑚C (𝑥 = 0.5 mm, 𝑡) = 0.006, we need the argument of erf to satisfy
𝑚C (𝑥, 𝑡) − 𝑚C,ᵆ 0.006 − 0.010
= = 0.500
𝑚C,0 − 𝑚C,ᵆ 0.002 − 0.010
From a table of the error function or an online error function calculator, we find erf(0.47694) =
0.50000. For each temperature, we must find the time that makes the argument equal 0.47694:
2.748 × 10−7 m2
2
1 0.0005
𝑡= ( ) = (*)
𝒟C,Fe 2(0.47694) 𝒟C,Fe m2/s
We can evaluate the diffusion coefficient at the two temperatures given, 910°C = 1183 K and
950°C = 1223 K:
𝒟C,Fe = (1.2 × 10−5 m2/s) exp[−(1.34 × 108 )/(8314.5)𝑇]
1.454 × 10−11 m2/s at 1183 K
={
2.271 × 10−11 m2/s at 1223 K
Putting these numbers into eqn. (*) we find
1.890 × 104 s = 5.250 h = 5 h 15 min at 1183 K = 910°C Answer
𝑡={ ⟵−−−−−−−
1.210 × 104 s = 3.362 h = 3 h 22 min at 1223 K = 950°C
A 40 K increase of temperature reduces the carburization time by 1⁄3.

329
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Comment: Carburization is widely used for hardening metal parts, and the process has been
widely studied. Analyses have addressed additional factors including convection, the temperature
dependence of physical properties, and base metal composition. Real processes may begin with
carburization as described here, but then be followed by an “annealing period” during which the
part is held at a somewhat lower temperature while the steep concentration profile in the metal
diffuses to a more uniform carbon concentration in the zone just under the surface [1].
[1] Goldstein, J.I., Moren, A.E., “Diffusion modeling of the carburization process,” Metallurgical
Transactions A, 9:1515–1525, 1978. doi:10.1007/BF02661934

Problem 11.19 (a) Derive eqn. (11.62) for the mole flux across a stagnant layer, working by
analogy to the mass-based analysis of Section 11.5 that led to eqn. (11.58b). Assume that 𝑐𝒟12
is constant, and use 𝑧 as the coordinate across the layer. (b) Show that the molar concentration
profile, analogous to eqn. (11.61), is
1 − 𝑥2 (𝑧) 1 − 𝑥2,𝐿 𝑥/𝐿
=( )
1 − 𝑥2,0 1 − 𝑥2,0

Solution
a) For one-dimensional steady mass transfer in the 𝑧-direction, the mole fluxes 𝑁1 and 𝑁2 satisfy
eqn. (11.50), with 𝑛𝑖 = 𝑀𝑖 𝑁𝑖 ,
𝑑𝑁1 𝑑𝑁2
= =0
𝑑𝑧 𝑑𝑧
so that 𝑁2 = constant and 𝑁1 = 0 throughout the layer.
With eqns. (11.24), (11.25), and (11.7)
𝑁1 = 𝑥1 𝑁 + 𝐽1 = (1 − 𝑥2 )𝑁 − 𝐽2 = 0 (*)
𝐽2
𝑁2 = 𝑥2 𝑁 + 𝐽2 = 𝑥2 + 𝐽2
(1 − 𝑥2 )
and eliminating 𝑁 with eqn. (*)
1
𝑁2 = 𝐽2 ( ) = constant in 𝑧
1 − 𝑥2
Now substituting the molar form of Fick’s law, eqn. (11.30),
𝑑𝑥 1
𝑁2 = (−𝑐𝒟12 2 ) ( ) = constant in 𝑧
𝑑𝑧 1 − 𝑥2
which is the molar analog of eqn. (11.58a).
The integration proceeds as in the mass-based case that gave eqn. (11.58b), but now
assuming that 𝑐𝒟12 is constant:
𝐿 2,𝐿𝑥
𝑁2 𝑑𝑥2
∫ 𝑑𝑧 = − ∫ (**)
𝑐𝒟12 0 𝑥
1 − 𝑥2
2,0

from which
𝑐𝒟12 1 − 𝑥2,𝐿 Answer
𝑁2 = ln( ) ⟵−−−−−−−
𝐿 1 − 𝑥2,0
which is eqn. (11.62).

330
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 b) Use the integral in eqn. (**), but set the upper limit to 𝑧 rather than 𝐿:
𝑧 2 𝑥 (𝑧)
𝑁2 𝑑𝑥2
∫ 𝑑𝑧 = − ∫
𝑐𝒟12 0 𝑥
1 − 𝑥2
2,0

from which
𝑁2 𝑧 1 − 𝑥2 (𝑧)
= ln( )
𝑐𝒟12 1 − 𝑥2,0
Now eliminate 𝑁2 with the result of part (a)
1 − 𝑥2 (𝑧) 𝑥 1 − 𝑥2,𝐿
ln( ) = ln( )
1 − 𝑥2,0 𝐿 1 − 𝑥2,0
from which
1 − 𝑥2 (𝑧) 1 − 𝑥2,𝐿 𝑥/𝐿 Answer
=( ) ⟵−−−−−−−
1 − 𝑥2,0 1 − 𝑥2,0

Comment: 𝑐𝒟12 is independent of pressure, so the assumption of constant 𝑐𝒟12 is appropriate

for a constant temperature gas mixture. Further, 𝑐 is independent of the mole fraction, so it will
not change across the layer even though a concentration gradient is present. This fact makes the
molar result more accurate than the mass-based result, which assumes that 𝜌𝒟12 is constant: the
variation in composition will affect 𝜌, since 𝜌 = 𝑀𝑐.

Problem 11.20 A Stefan tube 1 cm in diameter initially has a pool of liquid carbon tetra-
chloride 200.0 mm below the top. Pure argon flows over the tube. The system is held at 60°C and
8.0 × 104 Pa. During a 12 hr experiment, the pool level drops by 6.1 mm. What is the diffusivity
of CCl4 in Ar? The vapor pressure of CCl4 is log10 𝑝𝑣 = 4.023 − 1222/(𝑇 − 45.74), where 𝑝𝑣 is
in bar and 𝑇 in K. The specific gravity of liquid CCl4 is 1.59.

Solution We may use eqn. (11.64). The specific gravity tells us that
𝜌CCl4 = SG(𝜌H2O ) = (1.59)(982.9) = 1563 kg/m3
From Table 11.3, 𝑀CCl4 = 153.82 kmol/kg. The total pressure is 𝑝 = 8 × 104 Pa, and the saturation
pressure of CCl4 , or vapor pressure, is
1222
log10 𝑝𝑣 = 4.023 − = −0.2288 so 𝑝𝑣 = 0.5905 bar
333.15 − 45.74
Then
−1
[𝐿2 (𝑡) − 𝐿20 ] 𝜌CCl4 𝑅∘ 𝑇 𝑝
𝒟12 = ( )( ) [ln( )] (11.64)
2(𝑡 − 𝑡0 ) 𝑀CCl4 𝑝 𝑝 − 𝑝sat,2
−1
[(0.2061)2 − (0.2000)2 ] 1563 (8314.5)(333.15) 8 × 104
= ( )( ) [ln( )]
2(12)(3600) 153.82 8 × 104 8 × 104 − 59050
Answer
= 7.53 × 10−6 m2 /s ⟵−−−−−−−

331
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Comment: The Antoine equation used for vapor pressure is from the NIST Chemistry Webbook,
https://webbook.nist.gov, accessed 17 January 2024.

Problem 11.21 A Stefan tube at 60°C contains a pool of liquid ethanol 15 cm below the
top. Pure nitrogen gas flows across the top. The total pressure is 1.2 bar. Plot the concentration
profiles of ethanol and nitrogen in the tube, in terms of both mass fraction and mole fraction
(see Problem 11.19b). The vapor pressure of ethanol (C2 H5 OH) is given by log10 𝑝𝑣 = 5.247 −
1599/(𝑇 − 46.42) for 𝑝𝑣 is in bar and 𝑇 in K.

Solution At 60°C = 333.15 K,

1599
log10 𝑝𝑣 = 5.247 − = −0.3297 so 𝑝𝑣 = 0.4681 bar
333.15 − 46.42
From eqn. (11.15),
0.4681
𝑥C2H5OH,0 = = 0.3901
1.2
and at the top 𝑥C2H5OH,𝐿 = 0. With eqn. (11.8), molar mass above the liquid surface is
1 𝑥C H OH,0 𝑥N2,0
= 2 5 +
𝑀0 𝑀C2H5OH 𝑀N2
0.3901 1 − 0.3901
= +
46.07 28.01
𝑀0 = 33.07 kg/kmol
and with eqn. (11.9), the mass fractions are
(0.3901)(46.07)
𝑚C2H5OH,0 = = 0.5435
33.07
and 𝑚C2H5OH,𝐿 = 0.
Now we can use the result of Problem 11.19b and eqn. (11.61)
1 − 𝑥C2H5OH (𝑧) 1 − 𝑥C2H5OH,𝐿 𝑧/𝐿
=( )
1 − 𝑥C2H5OH,0 1 − 𝑥C2H5OH,0

1 − 𝑚C2H5OH (𝑧) 1 − 𝑚C2H5OH,𝐿 𝑧/𝐿

=( ) (11.61)
1 − 𝑚C2H5OH,0 1 − 𝑚C2H5OH,0
which rearrange to
1 − 𝑥C2H5OH,𝐿 𝑧/𝐿
𝑥C2H5OH (𝑧) = 1 − (1 − 𝑥C2H5OH,0 ) ( )
1 − 𝑥C2H5OH,0
= 1 − (0.6099)1−𝑧/(0.15)
with 𝑥N2 (𝑧) = 1 − 𝑥C2H5OH (𝑧), and
𝑚C2H5OH (𝑧) = 1 − (0.4565)1−𝑧/(0.015)
with 𝑚N2 (𝑧) = 1 − 𝑚C2H5OH (𝑧). The plots are below.

332
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 0.15
xC2 H5 OH
xN2
mC2 H5 OH
mN2

0.10
z [m]

0.05

0.00
0 0.2 0.4 0.6 0.8 1
mi (z) or xi (z)

Comment: The Antoine equation used for vapor pressure is from the NIST Chemistry Webbook,
https://webbook.nist.gov, accessed 19 January 2024.

Problem 11.22 Consider mass convection in a binary mixture, in which only species 1 is
transferred through the 𝑠-surface. Show that 𝑔𝑚,1 = 𝑔𝑚,2 . How does 𝑗2,𝑠 relate to 𝑛1,𝑠 ?

Solution From eqns. (11.66), (11.3), and (11.21) we have

𝑗1,𝑠 −𝑗2,𝑠 𝑗2,𝑠 Answer
𝑔𝑚,1 = = = = 𝑔𝑚,2 ⟵−−−−−−−
𝑚1,𝑠 − 𝑚1,𝑒 1 − 𝑚2,𝑠 − 1 + 𝑚2,𝑒 𝑚2,𝑠 − 𝑚2,𝑒
This applies for either low rate or high rate mass transfer.
Although 𝑛2,𝑠 = 0, species 2 may have a non-zero diffusive flux at the wall
𝑛2,𝑠 = 0 = 𝑗2,𝑠 + 𝑚2,𝑠 𝑛 = 𝑗2,𝑠 + 𝑚2,𝑠 𝑛1,𝑠
Answer
𝑗2,𝑠 = −𝑚2,𝑠 𝑛1,𝑠 ⟵−−−−−−−
The reason that 𝑗2,𝑠 is not zero is that species 2 diffuses against the convective velocity caused by
the transport of species 1 away from the wall.

Problem 11.23 A small sphere in a gas at rest has a low vapor pressure of species 1, so
that species 1 is dilute in the gas phase. When natural convection around the sphere is negligible,
the steady mass flux of species 1 in the radial direction is 𝑛1,𝑟 ≅ 𝑗1,𝑟 = −𝜌𝒟12 𝑑𝑚1 /𝑑𝑟. Use a
mass balance to obtain the 𝑠-surface mass flux in terms of the difference between the concentration

333
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 far from the sphere, 𝑚1,∞ , and near the surface, 𝑚1,𝑠 . Approximate 𝜌𝒟12 as constant, which is
accurate if species 1 is dilute. Then use eqns. (11.66) and (11.69) to show that Nu𝑚,𝐷 = 2. What
condition must apply for convection to be negligible?

Solution Let the small sphere have a radius 𝑅, and consider a control surface of radius 𝑟 > 𝑅
outside the small sphere. In steady state, the mass flow leaving the small sphere must equal the
mass leaving the outer sphere. Therefore
4𝜋𝑅2 𝑗1,𝑟 || = 4𝜋𝑟2 𝑗1,𝑟 || = constant in 𝑟 = ℂ
𝑅 𝑟
𝑑𝑚1 |
−4𝜋𝑟2 𝜌𝒟12 | =ℂ
𝑑𝑟 |𝑟
Rearrange and integrate
𝑚1,∞ ∞
ℂ 𝑑𝑟
∫ 𝑑𝑚1 = − ∫
𝑚⎵⏟⎵
4𝜋𝜌𝒟 12 ⏟⎵ 𝑟2
𝑅⏟⎵⏟
⏟⎵1,𝑠 ⎵⏟
=𝑚1,∞ −𝑚1,𝑠 =1/𝑅

Hence, with a bit of algebra,

ℂ 4𝜋𝜌𝒟12 𝑅(𝑚1,𝑠 − 𝑚1,∞ ) 𝜌𝒟12 Answer
𝑗1,𝑠 = 𝑗1,𝑟 || =
2
= 2
= (𝑚1,𝑠 − 𝑚1,∞ ) ⟵−−−−−−−
𝑅 4𝜋𝑅 4𝜋𝑅 𝑅
With eqns. (11.69) and (11.66), and putting 𝐷 = 2𝑅 for the diameter,
𝑔𝑚,1 𝐷 𝑗1,𝑠 (2𝑅) Answer
Nu𝑚,𝐷 = = = 2 ⟵−−−−−−−
𝜌𝒟12 𝜌𝒟12 (𝑚1,𝑠 − 𝑚1,∞ )
This result can be compared to eqn. (8.33) for natural convection from a sphere
0.589 Ra1/4
Nu𝐷 = 2 + 𝐷
4/9
(8.33)
[1 + (0.492/Pr)9/16 ]
which limits to Nu𝐷 = 2 when Ra𝐷 ≪ 1. From the analogy of heat and mass transfer, or
eqn. (11.71), we must have
𝑔Δ𝜌𝐷3 Answer
Ra𝐷 = Gr𝐷 Sc = ≪ 1 ⟵−−−−−−−
𝜇𝒟12

334
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 11.24

(8.51).

m1,o.

(11.44)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 335

 eqn.(11.57)

11.9

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 336

 11.25

(11.115)

(8.53).

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 337

 11.26 (11.44)

(6.42)

Like (6.21), (6.41) & (6.43),

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 338

 11.27

(6.50)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 339

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 340
 11.28

(11.108) 0

(7.65)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 341

 Problem 11.29 Helium is bled through a porous vertical wall, 40 cm high, into surrounding
air at a rate of 87.0 mg/m2 s. Both the helium and the air are at 300 K, the environment is at 1 atm,
and 𝒟He,air = 7.12 × 10−5 m2 /s. What is the average concentration of helium in the air along the
wall, 𝑚He,𝑠 ?

Solution This is a uniform flux natural convection problem. Here 𝑔𝑚,He and Δ𝜌 depend on
𝑚He,𝑠 , so the calculation is not as straightforward as it was for thermally driven natural convection.
To begin, let us assume that the concentration of helium at the wall is very small. Since 𝑚He,𝑒 = 0,
if 𝑚He,𝑠 ≪ 1, then 𝑚He,𝑠 − 𝑚He,𝑒 ≪ 1 as well. The mass transfer driving force will be small, and
the analogy to heat transfer can be used.
The mass flux of helium at the wall, 𝑛He,𝑠 , is given; and because the mass transfer rate is low,
𝑛He,𝑠 ≈ 𝑗He,𝑠 = 𝑔𝑚,He (𝑚He,𝑠 − 𝑚He,𝑒 )
Hence,
𝑔𝑚,He 𝐿 𝑛He,𝑠 𝐿
Nu𝑚,𝐿 = =
𝜌𝒟He,air 𝜌𝒟He,air (𝑚He,𝑠 − 𝑚He,𝑒 )
The appropriate Nusselt number is obtained from the mass transfer analog of eqn. (8.44b) for a
vertical plate with uniform flux
1/5
6 Ra∗𝐿 Sc
Nu𝑚,𝐿 = ( )
5 4 + 9√Sc + 10 Sc
where
𝑔Δ𝜌 𝑛He,𝑠 𝐿4
Ra∗𝐿 = Ra𝐿 Nu𝑚,𝐿 =2
𝜇𝜌𝒟He ,air (𝑚He,𝑠 − 𝑚He,𝑒 )
The Rayleigh number cannot be evaluated without assuming a value of the mass fraction of helium
at the wall. As a first guess, we pick 𝑚He,𝑠 = 0.010. Then the film composition is
𝑚He,𝑓 = (0.010 + 0)/2 = 0.005
From eqn. (12.8) and the ideal gas law, we obtain estimates for the film density (at the film
composition) and the wall density
𝜌𝑓 = 1.141 kg/m3 and 𝜌𝑠 = 1.107 kg/m3
At this low concentration of helium, we expect the film viscosity to be close to that of pure air:
𝜇𝑓 ≅ 𝜇air = 1.857 × 10−5 kg/m⋅s. The corresponding Schmidt number is Sc = (𝜇𝑓 /𝜌𝑓 )/𝒟He,air =
0.2286. Outside the boundary layer, we have pure air with density
𝜌𝑒 = 𝜌air = 1.177 kg/m3
From these values,
9.806(1.177 − 1.107)(87.0 × 10−6 )(0.40)4
Ra∗𝐿 =
(1.857 × 10−5 )(1.141)(7.119 × 10−5 )2 (0.010)
= 1.424 × 109
We may now evaluate the mass transfer Nusselt number
1/5
6[(1.424 × 109 )(0.2286)]
Nu𝑚,𝐿 = 1/5
= 37.73
5[4 + 9√0.2286 + 10(0.2286)]

342
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 From this value, we calculate
𝑛He,𝑠 𝐿
(𝑚He,𝑠 − 𝑚He,𝑒 ) =
𝜌𝒟He,air Nu𝑚,𝐿
(87.0 × 10−6 )(0.40)
=
(1.141)(7.12 × 10−5 )(37.73)
= 0.01136
But (𝑚He,𝑠 − 𝑚He,𝑒 ) = 𝑚He,𝑠 . This value of the average wall concentration is only 13.6% higher
than our initial guess of 0.010.
Using 𝑚He,𝑠 = 0.01136 as our second guess, we repeat the preceding calculations with revised
values of the densities to obtain
Answer
𝑚He,𝑠 = 0.01142 ⟵−−−−−−−
This result is within 0.5% of our second guess, so no further iteration is needed.

343
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.30 We’re off on a drive across West Texas. It’s going to be hot today—40°C—but
we’re unsure of the humidity. We attach a “desert water bag” to the shaded side of our pickup truck.
It’s made of canvas, and it holds a liter and a half of drinking water. When we fill it, we make sure
to saturate the canvas inside and out. Water will continue to permeate the fabric, but the weave
is tight enough that no water drips from it. Plot the temperature of the water inside the bag as a
function of the outdoor humidity. Hint: These bags were once widely used in the Western US, but
they never found much use along the US Gulf coast.

Solution
Water will evaporate from the wet canvas, cooling the water inside. In steady state, the bag
should cool to the wet bulb temperature appropriate to the outdoor humidity, assuming that it is
kept in the shade. This temperature can be found by solving eqn. (11.74), but a student familiar
with psychrometry (from a thermodynamics class) can read the wet bulb temperature from a
psychrometric chart as a function of RH. In addition, wet bulb temperature calculators are available
online, e.g., www.weather.gov/epz/wxcalc_rh.
For 𝑇dry bulb = 𝑇𝑒 = 40°C and an atmospheric pressure of 1 atm = 1013.25 hPa = 1013.25 mbar,
the results are as follow.

RH [%] 𝑇wet bulb [°C ] RH [%] 𝑇wet bulb [°C ]

10 18.9 60 32.7
20 22.3 70 34.7
30 25.3 80 36.6
40 28.0 90 38.4
50 30.5 100 40.0

45.0

40.0

35.0
𝖳wet bulb [°C ]

30.0

25.0
𝖳dry bulb = 𝟦𝟢 °C
20.0
𝗉 = 𝟣 atm

15.0
0 10 20 30 40 50 60 70 80 90 100
Relative Humidity, RH [%]

Comment: The water bag will be most effective in arid climates, such as the US Southwest. In
humid areas, such as the US Gulf coast, the cooling effect is minimal.

344
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.31 The following data were taken at a weather station over several months:
Date 𝑇dry-bulb 𝑇wet-bulb
3/15 15.5°C 11.0°C
4/21 22.0 16.8
5/13 27.3 25.8
5/31 32.7 20.0
7/4 39.0 31.2
Use eqn. (11.74) to find the mass fraction of water in the air at each date. Compare to values
from a psychrometric chart.

Solution By rearranging eqn. (11.74)

𝑐𝑝air Le2/3 (𝑇𝑒 − 𝑇wet-bulb )
𝑚H2O,𝑒 = 𝑚H2O,𝑠 − ( )
ℎ𝑓𝑔 |𝑇wet-bulb
To solve, we take Le = 0.847, and note from Table A.6 that 𝑐𝑝air = 1006 J/kg⋅K in the range of
interest, going up to 1007 for the hottest point. For better accuracy, optionally, we may replace 𝑐𝑝air
by 𝑐𝑝 = ∑𝑖 𝑐𝑝,𝑖 𝑚𝑖,𝑠 with 𝑐𝑝,water vap. ≅ 1920 J/kg⋅K. The latent heat of vaporization is available in
Table A.4 or any steam table. We may find 𝑚H2O,𝑠 as in Example 11.3 using a steam table.
𝑇wet-bulb ℎ𝑓𝑔 𝑝sat 𝑚H2O,𝑠 𝑐𝑝 𝑚H2O,𝑒
°C kJ/kg⋅K Pa [–] J/kg⋅K [–]
11.0 2475.4 1312 0.00810 1013 0.00645
16.8 2461.7 1914 0.01184 1017 0.00992
25.8 2440.3 3324 0.02067 1025 0.02061
20.0 2454.1 2339 0.01449 1019 0.00977
31.2 2558.5 4549 0.02842 1033 0.02560
Psychrometric charts provide moisture as 𝛾 in units of kg H2 O per kg dry air. To convert that to
mass fraction, note that
𝑚H2O = 𝜌H2O /𝜌 = 1/(1 + 𝜌dry air /𝜌H2O ) = 1/(1 + 1/𝛾)
Values read from a chart are used in the table below.
Date 𝛾 (chart) 𝑚H2O,𝑒 (chart) 𝑚H2O,𝑒 (computed) %-difference
3/15 0.0064 0.00636 0.00645 1.4%
4/21 0.0098 0.00970 0.00992 2.2
5/13 0.0206 0.0202 0.02061 2.0
5/31 0.0095 0.00941 0.00977 3.8
7/4 0.0260 0.0253 0.02560 1.2

Comment 1: The values read from a psychrometric chart are within 1.2 to 0.5% values obtained
from a reliable online calculator, with the largest difference found at the smallest 𝛾. Not all such
calculators give the same values, however—Caveat emptor!
Comment 2: The pyschrometric chart is based on thermodynamic relationships, such as the adi-
abatic saturation temperature, rather than transport relationships, as for the wet-bulb temperature.
As a result, we don’t expect exact agreement between these different calculations.

345
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.32 During a coating process, a thin film of ethanol is wiped onto a thick flat
plate, 0.1 m by 0.1 m. The initial thickness of the liquid film is 0.1 mm, and the initial temperature
of both the plate and the film is 303 K. The air above the film is at 303 K, flows at 10 m/s, and
contains no ethanol. (a) Assume that the plate is a poor conductor, so that heat transfer from it is
negligible. After a short initial transient, the liquid film reaches a steady temperature. Find this
temperature and calculate the time required for the film to completely evaporate. (b) Discuss what
happens when the plate is a very good conductor of heat, and calculate a lower bound on the time
to evaporate. Properties of ethanol are as follow: log10 (𝑝𝑣 mmHg) = 9.4432 − 2287.8/(𝑇 K);
ℎ𝑓𝑔 = 9.3 × 105 J/kg; liquid density, 𝜌eth = 789 kg/m3 ; Sc = 1.30 for ethanol vapor in air; vapor
specific heat capacity, 𝑐𝑝eth = 1420 J/kg⋅K.

Solution
a) When the plate is a poor conductor, the thin film will cool to the wet bulb temperature
appropriate for ethanol evaporating into air.
We will use values for pure air, assuming that the ethanol concentration in the boundary
layer remains low. The relevant properties at 300 K (from Table A.6) are

𝜈 = 1.575 × 10−5 m2/s 𝜌 = 1.177 kg/m3 𝑐𝑝 = 1006 J/kg-K

and the Lewis number is

Le = Sc ∕ Pr = (1.30) ∕ (0.707) = 1.84

The air away from the plate contains no ethanol, so 𝑚eth,𝑒 = 0. At the surface of the
film, the mass fraction will depend on the vapor pressure, which depends on the liquid
surface temperature (see Examples 11.3 and 11.6). The relevant equations are (with 𝑝atm =
760 mmHg and 𝑀eth = 46.07 kg/kmol, since ethanol is C2 H5 OH, cf. Table 11.2):

log10 (𝑝𝑣,𝑠 mmHg) = 9.4432 − 2287.8/(𝑇𝑠 K) (*)

𝑥eth,𝑠 = 𝑝𝑣,𝑠 ∕ 𝑝atm = 𝑝𝑣,𝑠 ∕ (760)
(𝑥eth,𝑠 )(46.07)
𝑚eth,𝑠 =
[(𝑥eth,𝑠 )(46.07) + (1 − 𝑥eth,𝑠 )(28.96)]
We need to iterate this set of equations with the energy balance for the wet bulb tempera-
ture, 𝑇𝑠 , which follows eqn. (11.74):
ℎ𝑠𝑓
𝑇𝑠 = 𝑇𝑒 − ( )(𝑚eth,𝑠 − 𝑚eth,𝑒 )
𝑐𝑝 Le2/3
9.3 × 105
𝑇𝑠 = 303 − ( )𝑚eth,𝑠
(1006)(1.84)2/3
𝑇𝑠 = [303 − 615.7 𝑚eth,𝑠 ] K (**)

In the table that follows, we guess 𝑇𝑠 in eqn. (*), and evaluate sequentially until we get a
revised value from eqn. (**). Since the revised values are sensitive to the guessed values,
we can guess a new value between the original and revised values (in numerical methods
parlance, this approach is called “under-relaxation”).

346
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Guess 𝑇𝑠 𝑝𝑣,𝑠 𝑥eth,𝑠 𝑚eth,𝑠 Get 𝑇𝑠
[K] [mmHg] [K]
303.0 78.11 0.1028 0.1542 208.7
280.0 18.73 0.02463 0.03864 279.2
279.7 18.35 0.02415 0.03788 279.68
Answer
The liquid film temperature is 𝑇𝑠 = 279.7 K ⟵−−−−−−−
The final value of 𝑚eth,𝑠 is low enough (3.8 wt%) that we may neglect the effect of ethanol
vapor on 𝑐𝑝 and 𝑇𝑠 (the difference would be on the order of +0.1 K).
To find the evaporation time, we need the mass transfer coefficient for the air flowing over
the plate. If we assume the air flows parallel to a 0.1 m dimension
(10)(0.1)
Re𝐿 = = 6.349 × 104 ⟹ laminar flow
1.575 × 10−5
Then, with eqn. (6.68) and the analogy between heat transfer and low rate mass transfer
(eqn. 11.69)
Nu𝐿 = 0.664 Re1/2
𝐿 Sc
1/3
= 182.6 (6.68)
The diffusion coefficient is
𝜈 1.575 × 10−5
𝒟eth,air = = = 1.212 × 10−5 m2/s
Sc 1.30
and, with eqn. (11.69), the mass transfer coefficient is
𝜌𝒟eth,air (1.177)(1.212 × 10−5 )
𝑔𝑚 = Nu𝐿 = (182.6) = 0.0260 kg/m2 s
𝐿 0.1
Since the conditions are steady, the time 𝑇 to complete evaporation is found by a mass
balance, per unit area, where 𝑡 is the film thickness:
𝜌eth 𝑡 = 𝑔𝑚 (𝑚eth,𝑠 − 𝑚eth,𝑒 ) 𝑇
(789)(0.0001) = (0.0260)(0.03788) 𝑇
from which
Answer
𝑇 = 80.1 sec ⟵−−−−−−−
The film will completely evaporate in less than a minute and half.

b) In this case, the thick plate will conduct heat into the ethanol film, holding its base temper-
ature at 303 K. The top of the film will be cooler as a result of heat conduction through the
film and heat loss to evaporation, but the two temperatures will grow closer as the film thins
with ongoing evaporation. A lower bound on the evaporation time would be obtained for
𝑇𝑠 = 303 K, which we can quickly calculate using 𝑚eth,𝑠 = 0.1542 from the first row of the
table in Part (a):
𝜌eth 𝑡 = 𝑔𝑚 (𝑚eth,𝑠 − 𝑚eth,𝑒 ) 𝑇
(789)(0.0001) = (0.0260)(0.1542) 𝑇
from which
Answer
𝑇 = 19.7 sec ⟵−−−−−−−

347
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 This time is 75% less than the wet-bulb time. The reason is that the vapor pressure rises
exponentially with the liquid surface temperature.
Comment 1: For part (b), one might worry about the conduction temperature drop from
the plate to the surface of the ethanol film. With 𝑘 = 0.17 W/m⋅K for ethanol, the thermal
resistance of the film is 𝑡/𝑘 = 5.9 × 10−4 m2 K/W, initially. The heat flux at liquid surface
𝑞 = ℎ𝑓𝑔 𝑔𝑚 𝑚eth,𝑠 , and for a surface temperature of 303 K, our values show 𝑞 = 3729 W/m2 .
Combining 𝑞 with the thermal resistance gives an initial temperature drop through the film of
2.2 K. So, the film surface temperature is initially slightly less than 301 K and the evaporation
rate is slightly lower. This difference diminishes to zero, however, as the film gets thinner
with continuing evaporation.
Comment 2: Further, one might worry about the conductive plate’s surface cooling down
through a transient heat conduction process. Using eqn (5.56) and the flux given in Com-
ment 1, a copper plate’s surface can be estimated to cool by about 0.7 K. Between this effect
and the effect in Comment 1, the time to complete evaporation is probably about 10% longer
than our lower bound calculation.

348
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.33 Ice cubes left in a freezer will slowly sublime into the air. Suppose that a
tray of ice cubes is left in a freezer with air at −10°C and a relative humidity of 50%. The air in the
freezer is circulated by a small fan, creating a heat transfer coefficient from the top of the ice cube
tray of 5 W/m2 K. If a 20 g ice cube is rectangular and has an exposed top surface area of 8 cm2 ,
find the temperature of the ice cube and estimate the time required for it to sublime completely.
Assume that no heat is transferred through the ice cube tray. For ice, take ℎsg = 2.837 MJ/kg, and
for water vapor in air, take Sc = 0.63. The vapor pressure of ice is given in Example 11.6.

Solution
Sublimation absorbs the latent heat, and so heat must be convected from the air to the ice. As
a result, the ice will become cooler than the surrounding air. Further, the ice will cool to a “wet-
bulb” temperature that remains fixes while the ice cube sublimes away. In that condition, no heat
is conducted into the ice. The calculation needed then follows Sect. 11.7.
As mentioned, Example 11.6 provides the saturation vapor pressure at −10°C
𝑝sat (−10°C) = 0.260 kPa
so that the partial pressure at 50% relative humidity is
𝑝𝑣 = (0.50)(0.0260) = 0.130 kPa
Using the method of Example 11.3:
𝑥H2O,𝑒 = 𝑝𝑣 /𝑝atm = (0.130) ∕ (101.325) = 1.283 × 10−3
and
(𝑥H2O,𝑒 )(18.02)
𝑚H2O,𝑒 = (11.48)
[(𝑥H2O,𝑒 )(18.02) + (1 − 𝑥H2O,𝑒 )(28.96)]
= 7.99 × 10−4
The mole fraction of water vapor above the ice will be smaller than 𝑝sat (−10°C)/𝑝atm = 2.6 ×
10−3 and the mass fraction will be similarly small, so we can be certain that 𝐵𝑚,H2O ≪ 1 and that
the mass transfer rate is very low.
We may apply eqn. (11.74) directly by replacing the wet-bulb temperature by the ice temperature
and ℎ𝑓𝑔 by ℎ𝑠𝑓
ℎ𝑠𝑓 |𝑇
𝑇𝑒 − 𝑇ice = ( ice
)(𝑚H2O,𝑠 − 𝑚H2O,𝑒 ) (11.74)
𝑐𝑝air Le2/3
From Table A.6, 𝑐𝑝 = 1006 J/kg-K and Pr = 0.713, for air at 260 K = −13 °C. We are given
ℎsg = 2.837 MJ/kg. The Lewis number is
Sc 0.63
Le =
= = 0.884
Pr 0.713
The rest of the solution is an iteration on the following equation:
ℎ𝑠𝑓
𝑇ice = 𝑇𝑒 − ( )(𝑚H2O,𝑠 − 𝑚H2O,𝑒 )
𝑐𝑝air Le2/3
2.837 × 106
= −10 − ( )(𝑚H2O,𝑠 − 7.99 × 10−4 )
(1006)(0.884)2/3

349
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 We may guess 𝑇ice , use that value to compute 𝑚H2O,𝑠 (following Examples 11.3 and 11.6), and then
get a new value of 𝑇ice by evaluating the right-hand side of the equation.
Guess 𝑇ice [°C] Guess 𝑇ice [K] 𝑚H2O,𝑠 Get 𝑇𝑠 [°C]
−14.0 259.15 0.00112 −10.98
−10.0 263.15 0.00160 −12.45
−12.0 261.15 0.00134 −11.66
−11.8 261.35 0.00136 −11.72
−11.7 261.45 0.00135 −11.75

Answer
So, the ice has a temperature of 𝑇ice ≅ 11.7°C. ⟵−−−−−−−
We find the mass transfer coefficient from eqn. (11.73)
ℎ
≅ Le2/3 (11.73)
𝑔𝑚 𝑐𝑝
so that
ℎ 5
𝑔𝑚 = = = 5.40 × 10−3 kg/m2 -s
𝑐𝑝 Le 2/3 (1006)(0.884) 2/3

The mass transfer rate is steady, so the time 𝑡 to sublime the entire ice cube is found by solving
20 g = 𝐴𝑔𝑚 (𝑚H2O,𝑠 − 𝑚H2O,𝑒 ) 𝑡
0.020 kg = (8 × 10−4 )(5.40 × 10−3 )(0.00135 − 7.99 × 10−4 ) 𝑡
from which
Answer
𝑡 = 8.40 × 106 sec = 2334 h = 97.2 days ⟵−−−−−−−
The ice cube will fully sublime in about 3 months.
Comment: The rate of sublimation depends on the temperature, humidity, and heat transfer
coefficient present in any specific freezer. Warmer, drier freezers will tend to reduce the ice cube’s
lifetime.

350
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.34 Bikram Yoga was a strenuous style of yoga done in a room at 38 to 41°C
with relative humidity from 20 to 50%. People doing this yoga will generate body heat 𝑄𝑏̇ of 300
to 600 W, which must be removed to avoid heat stroke. Calculate the rate at which one’s body can
cool under these conditions and compare it to the rate of heat generation.
The body sweats more as its need to cool increases, but the amount of sweat evaporated on the
skin depends on air temperature and humidity. Sweating cannot exceed about 2 liters per hour, of
which only about half evaporates (the rest will simply drip).
Assume that sweating skin has a temperature of 36°C and an emittance of 0.95, and that an
average body surface area is 𝐴𝑏 = 1.8 m2 . Assume that the walls in the yoga studio are at the
air temperature. Assume that the lightweight yoga clothing has no thermal effect. Water’s vapor
pressure can be taken from a steam table or other database. Convection to a person active in still
air can be estimated from the following equation, see [11.79]:
0.39
𝑄𝑏̇
ℎ = (5.7 W/m K)( 2
− 0.8)
(58.1 W/m2 ) 𝐴𝑏
Note that at high humidity and temperature, some people become overheated and must stop exer-
cising.

Solution
The body exchanges heat with the room by convection, radiation, and evaporation. If the rate of
internal generation by exercise, 𝑄𝑏̇ , exceeds the rate of cooling, the body’s temperature will rise
and potentially cause heat stroke. The three modes of heat transfer may be calculated as

𝑄conv = 𝐴𝑏 ℎ (𝑇skin − 𝑇air )

𝑄rad = 𝐴𝑏 𝜀skin 𝜎(𝑇𝑏4 − 𝑇walls
4
)
𝑄evap = 𝐴𝑏 𝑔𝑚 (𝑚H2O,skin − 𝑚H2O,air )ℎ𝑓𝑔
where
𝑇skin = 36°C = 309 K 𝑇walls = 𝑇air
𝜀skin = 0.95 𝜎 = 5.670 × 10−8 W/m2 K4
𝐴𝑏 = 1.8 m2 ℎ𝑓𝑔 = 2.414 × 106 J/kg-K
The mass transfer coefficient can be estimated from the heat transfer coefficient using eqn. (11.73)

ℎ Le = 0.847 see Sect. 11.7

𝑔𝑚 = for {
𝑐𝑝,air , Le2/3 𝑐𝑝,air = 1007 J/kg-K at 309 K

For the two values of 𝑄𝑏̇ , we calculate ℎ and 𝑔𝑚 from the given equation

ℎ = 7.57 W/m2 K
𝑄𝑏̇ = 300 W {
𝑔𝑚 = 0.008397 kg/m2 s
ℎ = 10.6 W/m2 K
𝑄𝑏̇ = 600 W {
𝑔𝑚 = 0.01178 kg/m2 s

351
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Assuming wet skin, we can find the saturation vapor pressure at the skin temperature (36°C)
from a steam table or from the NIST Webbook:
𝑝sat = 0.0059479 MPa = 5948 Pa
Then, by following Example 11.3, we find
𝑚H2O,skin = 0.03736
For the mass fraction of water in the air at a relative humidity (RH) less than 1, the mole fraction is
𝑝H O,room RH ⋅ 𝑝sat (𝑇room )
𝑥H2O,room = 2 = (*)
𝑝atm 𝑝atm
Referring to a steam table
6633 Pa 38°C
𝑝sat = {
7788 Pa 41°C
and with eqn. (*) and Example 11.3
0.00819 RH 20%, 38°C
𝑚H2O,room = {
0.0243 RH 50%, 41°C
N.B.: The amount of water vapor in the air becomes much higher with seemingly small increases
in the temperature and RH.
We must also consider the maximum rate of sweat evaporation, 1 L/h. The associated heat
removal is
𝑄evap,max = 𝜌water ℎ𝑓𝑔 (1 L/h)(0.001 m3 /L)(1 h/3600 s)
= (993.6)(2.414 × 106 )(0.001)(3600)−1 = 666.3 W
We may now calculate all heat transfers for each of the four conditions, as in the table below.
When the evaporative heat transfer exceeded 𝑄evap,max , we replaced it by 666.3 W.

20%, 38°C 50%, 41°C

Mode 300 W 600 W 300 W 600 W
𝑄rad −23.1 −23.1 −58.7 −58.7
𝑄conv −27.3 −38.2 −68.1 −95.4
𝑄evap 666.3 666.3 476.5 666.3
𝑄total 616.0 604.9 350.9 512.0

Can one stay cool? Reading from left to right on the bottom row, in the first case a person
generating 300 W has twice the needed cooling power available and will not overheat. In the
second case, the person can barely stay cool enough (and note that these estimates have significant
uncertainty!). In the third case, the person can stay cooling, but again only barely. In the fourth
case, the person will not be able to remain cool and would likely need to stop and lie down.

352
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 11.35

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 353

 do apply

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 354

 Problem 11.36 Small water droplets evaporate rapidly dry air. Suppose that a droplet is
placed in a room at 30°C, 40% RH, and 1 atm. Calculate the lifetime of the droplet if its initial
diameter is 0.2, 0.5, 1, 2, and 5 µm.

Solution From Example 11.14, we know that the droplet cools rapidly to its wet bulb
temperature and then evaporates until it vanishes, following the 𝐷2 law, eqn. (11.75). In particular,
𝑚H2O,𝑠 is the mass fraction of water vapor adjacent to the droplet at the wet bulb temperature.
Therefore, the first step in this problem is to find 𝑇wet bulb . We can do this the hard way, by solving
eqn. (11.74), while calculating the mass fraction as in Example 11.3 for RH = 𝑝H2O,𝑒 /𝑝sat (𝑇𝑒 )—or
we can do this the easy way, by using a psychrometric chart or psychrometric calculator. The author
has written many problem solutions already, so he chooses the easy way:
𝑇wet bulb = 20.0°C and 𝛾 = 0.0106 kg H2 O/kg dry air
Note that the moisture ratio on a psychrometric chart is for the dry bulb condition, not the wet bulb
condition. Some algebra relates 𝛾 to 𝑚H2O,𝑒
𝑚H2O,𝑒 = 𝜌H2O /𝜌 = 1/(1 + 𝜌dry air /𝜌H2O ) = 1/(1 + 1/𝛾) = 0.01049
We are not completely off the hook, however, as we still need to find 𝑚H2O,𝑠 , which we do
as in Example 11.3. With a steam table, 𝑝sat (20.0°C) = 2339.3 Pa; then eqn. (11.47) gives
𝑥H2O,𝑠 = 0.02309, and eqn. (11.48) gives 𝑚H2O,𝑠 = 0.01449.
Now we may turn our attention to eqn. (11.75). Call the droplet’s lifetime 𝜏, so that 𝐷(𝜏) = 0.
Then, by rearranging eqn. (11.75):
𝜌𝑙 𝐷02
𝜏=
8𝜌𝒟H2O,gas (𝑚H2O,𝑠 − 𝑚H2O,𝑒 )
We need two densities and a diffusion coefficient. Liquid water at 20°C has 𝜌𝑙 = 998.2 kg/m2 . At
water mass fractions near 1%, the density of air is close to that of dry air. We can easily be precise,
however, by using eqn. (11.8) and the ideal gas law:
𝑝
𝜌 = ∘ 𝑀mixture
𝑅𝑇
(101325)
= [(0.02309)(18.02) + (1 − 0.02309)(28.96)] = 1.193 kg/m3
(8314.5)(293.15)
The diffusion coefficient may be computed from eqn. (11.36):
𝑇 2.072 (293.15)2.072
𝒟H2O,air = 1.87 × 10−10 ( ) = 1.87 × 10−10 ( ) = 2.42 × 10−5 m2 /s
𝑝 1
Putting all this together, for 𝜏 in seconds and 𝐷0 in meters,
998.2
𝜏=[ ] 𝐷2
8(1.193)(2.42 × 10−5 )(0.01449 − 0.01049) 0
= 1.081 × 109 𝐷02
which looks very large until we notice that 𝐷02 = 𝒪(10−12 m2 )! The results are summarized below.

𝐷0 [µm] 0.2 0.5 1.0 2.0 5.0

𝜏 [ms] 0.0432 0.270 1.08 4.32 27.0

Comment: Small droplets do not last very long! (Unless the air is quite humid!)

355
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.37 A Couette flow (or stagnant film) model of a laminar boundary layer neglects
streamwise derivatives locally, so that the velocity varies in 𝑦, but not 𝑥, from 𝑢 = 0 at the wall to
𝑢∞ at the edge of the b.l., 𝑦 = 𝛿. The b.l. thickness 𝛿 is assumed increase only slowly with 𝑥. (a)
Show that 𝑢/𝑢∞ = 𝑦/𝛿 in laminar flow. (b) Calculate skin friction coefficient, 𝐶𝑓 , the temperature
profile, 𝑇(𝑦), and the Nusselt number, Nu𝑥 , in terms of 𝛿 and 𝛿𝑡 . (c) Using eqns. (6.31b) and
(6.55), show that the laminar Couette flow model results in estimates of 𝐶𝑓 (𝑥) and Nu𝑥 that differ
from eqns. (6.33) and (6.58) by a constant on the order of one.

Solution
a) We start with the continuity equation, eqn. (6.11a), neglecting the 𝑥-derivative:
0
𝜕𝑢 𝜕𝑣




+ =0

𝜕𝑥 𝜕𝑦
This equation gives 𝑣(𝑦) = constant = 0, since 𝑣 = 0 at the wall. The momentum equation,
eqn. (6.15), then becomes
0 0
𝜕𝑢 

 𝜕𝑢 𝜕2 𝑢
𝑢 +𝑣

 =𝜈 2 =0

𝜕𝑥  𝜕𝑦 𝜕𝑦
which is easily solved for
0 0 𝑦 𝑦 Answer
𝑢(𝑦) = 𝑎 + 𝑏𝑦 = 
𝑢wall + (𝑢∞ − 
𝑢wall ) = 𝑢∞ ⟵−−−−−−−
𝛿 𝛿
b) The wall shear stress is
𝜕𝑢 | 𝜇𝑢∞
𝜏𝑤 = 𝜇 | =
|
𝜕𝑦 𝑦=0 𝛿
so that
𝜏𝑤 2𝑥 Answer
𝐶𝑓 (𝑥) = = ⟵−−−−−−−
𝜌𝑢∞ /2 𝛿Re𝑥
2

The temperature profile satisfies eqn. (6.40), which simplifies like the momentum equation:
0 0
𝜕𝑇
 𝜕𝑇 𝜕2 𝑇
𝑢
+ 𝑣 =𝛼 2 =0

𝜕𝑥  𝜕𝑦 𝜕𝑦
This equation can easily be solved for
𝑦
𝑇(𝑦) = 𝑎 + 𝑏𝑦 = 𝑇𝑤 + (𝑇∞ − 𝑇𝑤 )
𝛿𝑡
in which 𝛿𝑡 the thermal boundary layer thickness. Then
𝜕𝑇 |
−𝑘 |
𝜕𝑦 |𝑦=0 𝑘
ℎ= =
(𝑇𝑤 − 𝑇∞ ) 𝛿𝑡
and
ℎ𝑥 𝑥 Answer
Nu𝑥 = = ⟵−−−−−−−
𝑘 𝛿𝑡

356
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 c) Now we can substitute the expressions for the boundary layer thicknesses from eqns. (6.31b)
and (6.55):

2𝑥 2√Re𝑥 0.431 Answer

𝐶𝑓 (𝑥) = = = ⟵−−−−−−−
𝛿Re𝑥 4.64 Re𝑥 √Re𝑥

𝑥 𝑥𝛿 √Re𝑥 𝛿 Answer
Nu𝑥 = = = = 0.216 Re1/2
𝑥 Pr
1/3
⟵−−−−−−−
𝛿𝑡 𝛿 𝛿𝑡 4.64 𝛿𝑡

Comment: The coefficients are low by about 2/3 relative to eqns. (6.33) and (6.58) because the
effective boundary thickness of the model is about 1/3 too large. In Section 11.8, this difference is
accounted for by eqn. (11.80), which may be regarded as defining 𝛿𝑐 to provide the correct 𝑔∗𝑚 for
either laminar or turbulent flow.

Problem 11.38 (a) What are the largest and smallest values of the mass transfer driving
force, 𝐵𝑚,2 ? (b) Plot the blowing factor as a function of 𝐵𝑚,2 . Indicate on your graph the regions
of blowing, suction, and low-rate mass transfer.

Solution
a) The mass transfer driving force, 𝐵𝑚,2 , is

𝑚2,𝑠 − 𝑚2,𝑒
𝐵𝑚,2 =
1 − 𝑚2,𝑠

Let us first assume that 𝑚2,𝑠 ⩾ 𝑚2,𝑒 . Then, as 𝑚2,𝑠 → 1, 𝐵𝑚,2 → ∞. If 𝑚2,𝑠 → 𝑚2,𝑒 , then
𝐵𝑚,2 → 0. Now assume that 𝑚2,𝑠 < 𝑚2,𝑒 . If 𝑚2,𝑠 → 0, 𝐵𝑚,2 → −𝑚2,𝑒 ⩾ −1. Therefore,

Answer
−1 ⩽ 𝐵𝑚,2 ⩽ ∞ ⟵−−−−−−−

b) The blowing factor is BF = ln(1 + 𝐵𝑚,2 )/𝐵𝑚,2 . As 𝐵𝑚,2 → ∞, BF → 0. As 𝐵𝑚,2 → −1+ ,

BF → +∞. For |𝐵𝑚,2 | small, ln(1 + 𝐵𝑚,2 ) ∼ 𝐵𝑚,2 − 𝐵𝑚,2
2
/2 + ⋯, so BF ∼ 1 − 𝐵𝑚,2 /2 + ⋯,
or succinctly

ln(1 + 𝐵𝑚,2 ) ⎧0, 𝐵𝑚,2 → ∞

⟶ 1 − 𝐵𝑚,2 /2 + ⋯ , |𝐵𝑚,2 | → 0
𝐵𝑚,2 ⎨
⎩∞, 𝐵𝑚,2 → −1+

The plot is as shown below.

357
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 ln(1 + Bm,2 )/Bm,2

Suction

Low rates

1
Blowing

−1 −0.2 0.2 Bm,2

358
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 11.39 (11.85)
(11.86) and (11.87).

(11.85)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 359

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 360
 Problem 11.40 (a) Derive eqn. (11.98) from eqn. (11.97). (b) In a test of a scramjet com-
bustor, gas is injected through a porous section of the combustor wall [1]. The gas velocity is
980 m/s. As a result of viscous friction and gas compression, the temperature of an adiabatic
section of the wall would be 1048 K: use this temperature as if it were the freestream temperature.
The heat transfer coefficient without injection is ℎ∗ = 720 W/m2 K. The injected gas is drawn from
a reservoir at 350 K at a mass flux of 𝑛𝑖,𝑠 = 0.3 kg/m2 s. What is the wall temperature 𝑇𝑠 if the gas
is He, Ar, or N2 ? (c) Plot 𝑇𝑠 for each gas if 0 ⩽ 𝑛𝑖,𝑠 ⩽ 1.5 kg/m2 s.

Solution
a) Substitute for ℎ in eqn. (11.98) with eqn. (11.88) and divide through by 𝑛𝑖,𝑠 𝑐𝑝,𝑖 :
𝑇𝑒 + 𝑇𝑟 [exp(𝑛𝑖,𝑠 𝑐𝑝,𝑖 /ℎ∗ ) − 1]
𝑇𝑠 =
1 + [exp(𝑛𝑖,𝑠 𝑐𝑝,𝑖 /ℎ∗ ) − 1]
Answer
= 𝑇𝑟 + (𝑇𝑒 − 𝑇𝑟 ) exp(−𝑛𝑖,𝑠 𝑐𝑝,𝑖 /ℎ∗ ) ⟵−−−−−−−
b) The specific heat capacity does not depend strongly on either 𝑝 or 𝑇, so we will overlook
variations in the thermodynamic state, using 1 atm values from Table A.6 at 350 K:

⎧521 J/kgK Ar 𝑛𝑖,𝑠 𝑐𝑝,𝑖 ⎧

0.217 Ar ⎧912 K Ar Answer
𝑐𝑝,𝑖 = 1043 J/kgK N2 = 0.435 N2 𝑇𝑠 = 802 K N2 ⟵−−−−−−−
⎨ ℎ∗ ⎨ ⎨
⎩5193 J/kgK He ⎩2.164 He ⎩430 K He
c)
1,000

900

800
Ar
700
Ts [K]

N2
600

500 He

400

300
0 0.2 0.4 0.6 0.8 1 1.2 1.4

Mass flux, ni,s [kg/m2 s]

Reference:
[1] T. Langener, J. von Wolfersdorf, and J. Steelant, “Experimental Investigations on Transpi-
ration Cooling for Scramjet Applications Using Different Coolants,” AIAA J., 49:7:1409–1419,
2011.

361
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.41 Dry ice (solid CO2 ) is used to cool medical supplies transported by a small
plane to a remote village in Alaska. A roughly spherical chunk of dry ice, 5 cm in diameter, falls
from the plane through air at 5°C. It reaches a terminal velocity of 40 m/s. What are the temperature
and sublimation rate of the dry ice, assuming steady conditions? The latent heat of sublimation is
about 590 kJ/kg and log10 (𝑝𝑣 bar) = 6.81228 − 1301.679/(𝑇 K − 3.494). The surface temperature
will be well below the solid-vapor equilibrium temperature of CO2 at 1 atm, which is −78.5°C.
Use this correlation for forced convection over a sphere in air at room temperature

𝐷 + 0.0011 Re𝐷
Nu𝐷 = 2 + 0.493 Re1/2
for 7800 ⩽ Re𝐷 ⩽ 2.9 × 105 [1], and approximate the Prandtl number dependence. Neglect heat
conduction into the ice. Hint: First use the properties of pure air, and then correct the properties if
necessary.

Solution This calculation is analogous to a wet-bulb temperature calculation if heat con-

duction, 𝑞ᵆ , into the dry ice is neglected. Heat conduction would reduce the amount of energy
available for sublimation at the surface, so that when we neglect conduction we obtain an upper
bound on the sublimation rate and a lower bound on the surface temperature.
For a property reference state, we may first use the properties of pure
air. We don’t know the surface temperature, although it will be below
−78.5°C. We can pick a film temperature that is easy to interpolate in
Table A.6, say 𝑇𝑓 = 225 K (which corresponds to 𝑇𝑠 = −101.3°C).
From Table A.6, for air: 𝑘 = 0.0206 W/m⋅K, 𝜈 = 0.944 × 10−5 kg/m-
s, Pr = 0.721, 𝜌 = 1.57 kg/m3 . We find the diffusivity from eqn. (11.108)
with 𝜎𝐴𝐵 = 3.785 Å, 𝜀𝐴𝐵 /𝑘𝐵 = 129.4, Ω𝐷 (1.739) = 1.132:
𝒟CO2,air = 0.9254 × 10−5 m2 /s so Sc = 1.02
In the absence of conduction, the heat convected to the surface will balance the latent heat of
sublimation. Our energy balance is
𝑞conv = ℎ𝑠𝑓, CO2 𝑛CO2,𝑠
ℎ(𝑇air − 𝑇𝑠 ) = ℎ𝑠𝑓, CO2 𝑔∗𝑚 ln(1 + 𝐵𝑚,CO2 ) (*)
The Reynolds number is
(40)(0.05)
Re𝐷 = = 2.12 × 105
0.944 × 10−5
and so Nu𝐷 = 462, from which
ℎ∗ = (462)(0.0206)/(0.05) = 190.3 W/m⋅K
Now we face a minor problem: the correlation is for heat transfer in air at Pr = 0.707 or so.
The Prandtl number dependence is likely to be Pr1/3 or so. Thus, the difference for Pr = 0.721 is
negligible. On the other hand, to use the analogy of heat and mass transfer, we must correct to Sc
= 1.02, so Copyright
we multiply the
2023, John NusseltIVnumber
H. Lienhard, and John H.by (1.02/0.707)
Lienhard, V
1/3
= 1.13 to find Nu𝑚,𝐷 = 522 from
which
𝑔∗𝑚 = (522)(1.57)(0.9254 × 10−5 )/(0.05) = 0.1517 kg/m2 s
Now, for the guessed surface temperature 𝑇𝑠 = −101.3°C = 171.9 K, the vapor pressure is
log10 (𝑝𝑣 bar) = 6.81228 − 1301.679/(171.9 − 3.494) = −0.9171 or 𝑝𝑣 = 0.1210 bar

362
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 which gives 𝑥CO2,𝑠 = 0.1194, with eqn. (11.15) and then 𝑚CO2,𝑠 = 0.1708, with eqn. (11.9). The
mass transfer driving force is
0.1708 − 0
𝐵𝑚,CO2 = = 0.206
1 − 0.1708
This value is at the edge of the low-rate mass transfer regime, but we can still apply the high rate
equations. The right-hand side of eqn. (*) provides
𝑛CO2,𝑠 = 𝑔∗𝑚 ln(1 + 𝐵𝑚,CO2 ) = 0.02842 kg/m2 s
and eqn. (11.88) with 𝑐𝑝,CO2 ≅ 780 J/kgK leads to ℎ = 179.4 W/m2 K. Then we use eqn. (*) to find
ℎ𝑠𝑓, CO2 𝑛CO2,𝑠
(5.90 × 105 )(0.02842)
𝑇𝑠 = 𝑇air − =5+ = −88.5°C
ℎ 179.4
(It would make sense to put these equations into a spreadsheet to do the [tedious] iterations
needs to find the solution. However, as the author is old school, a hand iteration follows.)
Some caution is needed in making the next guess because the vapor pressure rises exponentially
with temperature: a big temperature change will cause a massive increase in vapor pressure, but our
two temperature estimates are close enough that a big change does not seem appropriate. Instead,
we will repeat the calculation with 𝑇𝑠 = −97°C = 176.2 K. Then
log10 (𝑝𝑣 bar) = 6.81228 − 1301.679/(176.2 − 3.494) = −0.7247 or 𝑝𝑣 = 0.1885 bar
which gives 𝑥CO2,𝑠 = 0.1860 and 𝑚CO2,𝑠 = 0.2578. The mass transfer driving force is
0.2578 − 0
𝐵𝑚,CO2 = = 0.3473
1 − 0.2578
Then, 𝑛CO2,𝑠 = 0.04522 kg/m2 s, ℎ = 173.2 W/m2 K, and we solve as before:
(5.90 × 105 )(0.04522)
𝑇𝑠 = 5 + = −149.0°C
173.2
The calculation is obviously very unstable. So we need to guess a lower value, say 𝑇𝑠 = −99°C.
Proceeding as before, but with 𝑇𝑠 = −99°C, we find 𝑥CO2,𝑠 = 0.1510, 𝑚CO2,𝑠 = 0.2128, 𝐵𝑚,CO2 =
0.2703, 𝑛CO2,𝑠 = 0.03630 kg/m2 s, ℎ = 176.5 W/m2 K, and we solve as before:
(5.90 × 105 )(0.03630)
𝑇𝑠 = 5 + = −116.3°C
176.5
At this point, we may reasonably conclude that −99°C < 𝑇𝑠 < −101.3°C. Further iteration does
not make sense, given other approximations and uncertainties. We may estimate the sublimation
rate as the average of the values computed at these two temperatures:
Answer
𝑇𝑠 ≅ −100°C ⟵−−−−−−−
Answer
𝑛CO2,𝑠 = (0.02842 + 0.03630)/2 ≅ 0.032 kg/m2 s ⟵−−−−−−−
The final temperature is very close to our guessed film temperature. The film composition is
about 10 wt% CO2 . Given this low amount and the relative similar of many air and CO2 properties,
adjustment of the property reference will not significantly improve the accuracy of the solution.

Comment 1: The ambiguity in the Prandtl number dependence of the Nusselt number is a real-
world problem. Available correlations that include the Prandtl number do not extend to such a high
Reynolds number [1]; and those correlations, when extrapolated for air, give a value well below the

363
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 very accurate correlation for air from Ref. [1]. Extrapolating the Prandtl or Schmidt number from
0.7 to 1 is likely to provide better accuracy than extrapolating the Reynolds number.
Comment 2: The conduction heat flux at the surface of the ice may be small compared to the
convective heat flux outside, as a result of the high temperature difference and the very high value
of ℎ. We can calculate the conduction flux for Fo > 0.2 by differentiating eqn. (5.42) with respect
to 𝑟 (in 𝑓1 ) for Fo > 0.2. Skipping the details, this calculation leads to
𝑑𝑇 | 𝑘Δ𝑇𝐴1
𝑞ᵆ = −𝑘 | = − [ cos(𝜆1̂ ) − sin(𝜆1̂ )/𝜆1̂ ] exp(−𝜆21 Fo)
𝑑𝑟 |𝑟 𝑟𝑜
𝑜

We lack property data for dry ice, but can crudely approximate 𝑘 ≈ 1 W/m⋅K and 𝛼 ≈ 10−6 m2 /s.
This makes Bi ≈ 4.5, so that 𝜆1̂ ≈ 2.5 and 𝐴1 ≈ 1.75. Hence,
𝑞ᵆ ≈ 7315 exp(−6.25 Fo) W/m2
Since the ice is falling fast, let’s consider only Fo = 0.2. That’s 𝑡 = (0.2)𝑟𝑜2 /𝛼 ≈ 125 s. The ice
will have fallen 5000 m at terminal speed by then, so it’s very likely to reach the ground within this
time frame. In any case, the flux at this time is:
𝑞ᵆ ≈ 2100 W/m2
The convective heat flux is ℎΔ𝑇 = 178(5 + 100) = 18.7 kW/m, which is almost 9 times larger.
We conclude that conduction is negligible by the end of a long fall, but that it will be larger and
perhaps significant in the early stages of the fall.
References:
[1] J. B. Will, N. P. Kruyt, C. H. Venner, “An experimental study of forced convective heat transfer
from smooth, solid spheres,” Int. J. Heat Mass Transfer, 109:1059–1067, 2017, doi:10.1016/j.ijheat-
masstransfer.2017.02.018.
[2] Sublimation data are from NIST, https://webbook.nist.gov, accessed 9 February 2024.

364
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 11.42

Use
315

(close to saturation)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 365

 From Table A.5:

ρ = 0.05766 kg/m3
cp = 1935 J/kg-K µ = 1.037x10-5 kg/m-s
k = 0.01975 W/m-K Pr = 1.02

11,126

(6.68) 5.570

Use the same property ref. state.

From (11.36), DH2O-air = 3.556x10-4 m2/s. Calculate Sc = 0.506.

Using (6.68) with the analogy of heat and mass transfer, Num

= 121.6. Solving: gm* = 4.577x10-3 kg/m2-s.

=(4.577x10-3)

5.570

Use eqn. (11.48) for the mass fraction. The overall

calculation is best done with a spreadsheet.

29 24.13 45
28.5 0.3712 -0.9364 -0.01261 24.71 28.58
28.6 0.3775 -0.9357 -0.01256 24.62 28.56

It appears that Ts is about 28.5 degC. This implies a film

temperature of 310 K and film composition of 66.9 wt% H2O,
compared to 315 K and 100 wt% for our property ref. state.
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 366
 An additional iteration on the property reference state
increases gm* by about 5%, but has little effect on h*. The
net result is to raise Ts to 28.8 degC (negligible), but
also to increase the mass flux to about -0.0131 kg/m2s.

0.013

69

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 367

 Problem 11.43 (a) Compute the binary diffusivity of each of the noble gases when they
are individually mixed with nitrogen gas at 1 atm and 300 K. Plot the results as a function of the
molar mass of the noble gas. What do you conclude? (b) Consider the addition of a small amount
of helium, 𝑥He = 0.04, to a mixture of nitrogen, 𝑥N2 = 0.48, and argon, 𝑥Ar = 0.48. Compute
𝒟He,𝑚 and compare it with 𝒟Ar,𝑚 . Note that the increased concentration of argon does not raise
its diffusivity in the mixture.
Solution
a) We can make a table, with reference to Tables 11.3, 11.4, and eqn. (11.108):

Ar He Kr Ne Xe
𝜎𝑎,N2 [Å] 3.516 3.115 3.608 3.253 3.804
𝜀𝑎,N2 /𝑘𝐵 [K] 110.5 28.95 132.7 53.68 149.4
𝑘𝐵 𝑇/𝜀𝑎,N2 2.714 10.36 2.261 5.589 2.008
Ω𝐷 0.9769 0.7382 1.0331 0.8241 1.0746
𝑀 [kg/kmol] 39.95 4.003 83.80 20.18 131.29
𝒟𝑎,N2 × 10 [m /s] 1.970
5 2
7.203 1.567 3.233 1.292

8.0
He

6.0
Da,N2 × 105 [m2/s]

4.0

Ne
2.0
Ar
Kr
Xe
0.0
𝟣𝟢𝟣 𝟣𝟢𝟤
Molar mass [kg/kmol]
Answer
We see that the lighter species diffuse more readily than the heavier ones. ⟵−−−−−−−
b) Use Blanc’s law, eqn. (11.109). With eqn. (11.108), we compute 𝒟He,Ar = 7.664 × 105 m2 /s.
Then
𝑥 𝑥N2 −1 0.48 0.48 −1 Answer
𝒟He,𝑚 = ( Ar + ) =( + ) × 105 = 7.736 × 105 m2 /s ⟵−−−−−−−
𝒟He,Ar 𝒟He,N2 7.664 7.203
Since helium is a trace gas, it has little effect on the interaction of argon and nitrogen, so
Answer
𝒟Ar,𝑚 ≅ 𝒟Ar,N2 = 1.970 × 105 m2 /s ⟵−−−−−−−
The helium diffuses more readily than the heavier argon.

368
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.44 A mothball consists of a 2.5 cm diameter sphere of naphthalene (C10 H8 ) that
is hung by a wire in a closet. The solid naphthalene slowly sublimes to vapor, which drives off the
moths. Estimate the lifetime of this mothball in a closet with a mean temperature of 20°C. Use the
following data for napthalene
𝜎 = 6.18 Å, 𝜀/𝑘B = 561.5 K,
and, for the solid, 𝜌C10H8 = 1145 kg/m3 at 20°C. The vapor pressure of naphthalene near room
temperature is approximated by
log10 (𝑝𝑣 mmHg) = 11.450 − 3729.3/(𝑇 K)
The integral you will obtain can be evaluated numerically. The latent heat of sublimation and evap-
oration rate are low enough that the wet-bulb temperature is essentially the ambient temperature.

Solution
Natural convection is driven by a higher concentration of naphthalene near the mothball. At
20°C (293.15 K) for 1 atm = 760 mmHg, with 𝑀C10H8 = 128.17 kg/kmol,
𝑝𝑣 = 0.05352 mmHg (7.136 Pa)
𝑥C10H8,𝑠 = 𝑝𝑣 ∕ 𝑝atm = 0.05352 ∕ 760 = 7.042 × 10−5
(𝑥C10H8,𝑠 )(128.17)
𝑚C10H8,𝑠 = = 3.116 × 10−4
[(𝑥C10H8,𝑠 )(128.17) + (1 − 𝑥C10H8,𝑠 )(28.96)]
The closet should maintain only a very low concentration of naphthalene, so 𝑚C10H8,𝑠 ≅ 0. Further,
the concentration of naphthalene is low even close to the mothball, so we can use the properties of
air for all properties except for the gas density near the mothball.
To find a mass transfer coefficient, we must estimate the diffusion coefficient with eqn. (11.108)
(1.8583 × 10−7 )𝑇 3/2 1 1
𝒟𝐴𝐵 = + (11.108)
2
𝑝𝜎𝐴𝐵 Ω𝐷 √ 𝑀𝐴 𝑀𝐵

for 𝑝 in atm, 𝑇 in K, and 𝒟𝐴𝐵 in m2 /s. We use eqns. (11.106) and (11.107) and Tables 11.3 and
11.4 to find the necessary constants at 𝑇 = 293.15 K:
𝜎𝐴𝐵 = (𝜎C10H8 + 𝜎air )/2 = (6.18 + 3.711)/2 = 4.95 Å
𝜀𝐴𝐵 ∕ 𝑘B = √(561.5)(78.6) = 210 K, 𝑘B 𝑇 ∕ 𝜀𝐴𝐵 = 1.40
Ω𝐷 (1.40) = 1.2335
so that
(1.8583 × 10−7 )(293.15)3/2 1 1
√ 128.17 28.96 = 6.35 × 10 m /s
−6 2
𝒟C10H8,air = +
(1)(4.95)2 (1.2335)
and with 𝜈air = 1.51 × 10−5 m2/s, the Schmidt number is
Sc = 𝜈 ∕ 𝒟C10H8,air = 1.51 × 10−5 ∕ 6.35 × 10−6 = 2.38
Now, the naphthalene-containing air near the mothball will be denser than the fresh air farther
away, so a natural convection boundary layer will form. For natural convection on a sphere, we use

369
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 eqn. (8.33) with the analogy between heat transfer and low rate mass transfer
0.589 Ra1/4
Nu𝑚,𝐷 = 2 + 𝐷
4/9
(8.33)
[1 + (0.492/Sc)9/16 ]
𝑔𝑚 𝐷
= 2 + 0.5053 Ra1/4
𝐷 (*)
𝜌𝒟C10H8,air
𝜌𝒟C10H8,air
𝑔𝑚 = [2 + 0.5053 Ra1/4
𝐷 ] (**)
𝐷
where the diameter of the mothball, 𝐷, decreases in time. The Rayleigh number is
𝑔Δ𝜌 𝐷3
Ra𝐷 =
𝜌𝜈𝒟C10H8,air
We may find the densities using the ideal gas law. The mixture molar mass at the surface of the
mothball is, with eqn. (11.8),
3.116 × 10−4 1 − 3.116 × 10−4
−1
𝑀mix ,𝑠 = + ⟹ 𝑀mix,𝑠 = 28.967 kg/kmol
128.17 28.96
and
𝑝𝑀mix,𝑠 (101325)(28.967)
𝜌mix,𝑠 =
∘
= = 1.20419 kg/m3
𝑅𝑇 (8314.46)(293.15)
𝑝𝑀 (101325)(28.96)
𝜌mix,𝑒 = ∘ air = = 1.20390 kg/m3
𝑅𝑇 (8314.46)(293.15)
These densities differ by a very small amount, so a large number of digits must be carried. Then,
(9.806)(1.20419 − 1.20390)𝐷3
Ra𝐷 = −5 −6
= 2.49 × 107 𝐷3
(1.204)(1.51 × 10 )(6.35 × 10 )
Because the density difference is very small, so is the Rayleigh number: the initial value (for
𝐷 = 2.5 cm) is only 388.5. However, from eqn. (*), we find that the Nusselt number varies between
2 and 4.24, so that both terms in eqn. (*) remain important.
A mass balance on the mothball gives us
𝑑 𝜋
(𝜌C10H8,solid 𝐷3 ) = −𝜋𝐷2 𝑔𝑚 Δ𝑚C10H8 = 𝜋𝐷2 𝑔𝑚 𝑚C10H8,𝑠
𝑑𝑡 6
𝜋 𝑑𝐷
(𝜌C10H8,solid 𝐷2 ) = −𝜋𝐷2 𝑔𝑚 𝑚C10H8,𝑠
2 𝑑𝑡
and with eqn. (**)
𝑑𝐷 2𝑚C10H8,𝑠 𝜌𝒟C10H8,air 2𝑚C10H8,𝑠
= −𝑔𝑚 = [2 + 0.5053 (2.49 × 107 𝐷3 )1/4 ]
𝑑𝑡 𝜌C10H8,solid 𝐷 𝜌C10H8,solid
2 + 35.69 𝐷3/4 2(1.204)(6.35 × 10−6 )(3.116 × 10−4 )
= −( )( )
𝐷 1145
2 + 35.69 𝐷3/4
= (−4.161 × 10−12 )( )
𝐷
This equation can be integrated from the initial diameter 𝐷0 to zero:
𝑇 𝐷0 =0.025 m
𝐷 𝑑𝐷
(4.161 × 10−12 ) ∫ 𝑑𝑡 = ∫ = 8.766 × 10−5
0 0 (2 + 35.69 𝐷3/4 )

370
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 where the integral was evaluated numerically (in this case, using wolframalpha.com). Solving for
the time, 𝑇 :
8.766 × 10−5 Answer
𝑇= = 2.107 × 10 7
sec = 243.8 days ⟵ −−−−−−−
4.161 × 10−12
The mothball lasts about 8 months.
Comment 1: The vapor pressure rises very rapidly with temperature. If the closet is warmer, the
mothball lifetime will be much shorter.
Comment 2: Air circulation in the closet (as from drafts, HVAC systems, or room scale natural
convection) would increase the mass transfer coefficient and shorten the lifetime of the mothball.
Such circulation is more likely than not.
Comment 3: If the mothball is placed in a small closed container, such as a drawer or box, the
concentration of naphthalene may eventually rise to match the vapor pressure of naphthalene at the
surface of the mothball. In this equilibrium condition, the sublimation rate would drop to zero.
Comment 4: Naphthalene is considered to pose health hazards. Its use in mothballs has been
banned in the EU since 2008.
Comment 5: A fuller equation for the vapor pressure of solid naphthalene from 230 K to
𝑇tp = 353.34 K is given by: D. Ambrose, I.J. Lawrenson, C.H.S. Sprake, “The vapour pres-
sure of naphthalene,” J. Chemical Thermodynamics, 7(12):1173–1176, 1975. doi:10.1016/0021-
9614(75)90038-5

371
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.45 In contrast to the naphthalene mothball described in Problem 11.44, other
mothballs are made from paradichlorobenzene (PDB). Estimate the lifetime of a 2.5 cm diameter
PDB mothball using the following room temperature property data:
𝜎 = 5.76 Å 𝜀/𝑘𝐵 = 578.9 K 𝑀PDB = 147.0 kg/kmol
log10 (𝑝𝑣 mmHg) = 11.985 − 3570/(𝑇 K)
𝜌PDB = 1248 kg/m3

Solution Natural convection is driven by a higher concentration of paradichlorobenzene

near the mothball. At 20°C (293.15 K) for 1 atm = 760 mmHg, with 𝑀PDB = 147.0 kg/kmol,
𝑝𝑣 = 0.6411 mmHg (136.8 Pa)
𝑥PDB,𝑠 = 𝑝𝑣 ∕ 𝑝atm = 0.6411 ∕ 760 = 8.436 × 10−4
(𝑥PDB,𝑠 )(147.0)
𝑚PDB,𝑠 = = 4.267 × 10−3
[(𝑥PDB,𝑠 )(147.0) + (1 − 𝑥PDB,𝑠 )(28.96)]
The closet should maintain only a very low concentration of paradichlorobenzene, so 𝑚PDB,𝑠 ≅ 0.
Further, the concentration of paradichlorobenzene is low even close to the mothball, so we can use
the properties of air for all properties except for the gas density near the mothball.
To find a mass transfer coefficient, we must estimate the diffusion coefficient with eqn. (11.108)
(1.8583 × 10−7 )𝑇 3/2 1 1
𝒟𝐴𝐵 = + (11.108)
2
𝑝𝜎𝐴𝐵 Ω𝐷 √ 𝑀𝐴 𝑀𝐵
for 𝑝 in atm, 𝑇 in K, and 𝒟𝐴𝐵 in m2 /s. We use eqns. (11.14-15) and Tables 11.2 and 11.3 to find
the necessary constants at 𝑇 = 293.15 K:
𝜎𝐴𝐵 = (𝜎PDB + 𝜎air )/2 = (5.76 + 3.711)/2 = 4.736 Å
𝜀𝐴𝐵 ∕ 𝑘B = √(578.9)(78.6) = 213 K, 𝑘B 𝑇 ∕ 𝜀𝐴𝐵 = 1.374
Ω𝐷 (1.374) = 1.244
so that
(1.8583 × 10−7 )(293.15)3/2 1 1
√ 147.0 28.96 = 6.23 × 10 m /s
−6 2
𝒟PDB,air = 2
+
(1)(4.95) (1.244)
and with 𝜈air = 1.51 × 10−5 m2/s, the Schmidt number is
Sc = 𝜈 ∕ 𝒟PDB,air = 1.51 × 10−5 ∕ 6.23 × 10−6 = 2.42
Now, the 𝑝-dichlorobenzene containing air near the mothball will be denser than the fresh air
farther away, so a natural convection boundary layer will form. For natural convection on a sphere,
we use eqn. (8.33) with the analogy between heat transfer and low rate mass transfer
0.589 Ra1/4
Nu𝑚,𝐷 = 2 + 𝐷
4/9
(8.33)
[1 + (0.492/Sc)9/16 ]
𝑔𝑚 𝐷
= 2 + 0.5059 Ra1/4
𝐷 (*)
𝜌𝒟PDB,air

372
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 𝜌𝒟PDB,air
𝑔𝑚 = [2 + 0.5059 Ra1/4
𝐷 ] (**)
𝐷
where the diameter of the mothball, 𝐷, decreases in time. The Rayleigh number is
𝑔Δ𝜌 𝐷3
Ra𝐷 =
𝜌𝜈𝒟PDB,air
We may find the densities using the ideal gas law. The mixture molar mass at the surface of the
mothball is, with eqn. (11.8),
4.267 × 10−4 1 − 4.267 × 10−3
−1
𝑀mix ,𝑠 = + ⟹ 𝑀mix,𝑠 = 29.060 kg/kmol
147.0 28.96
and
𝑝𝑀mix,𝑠 (101325)(29.060)
𝜌mix,𝑠 =
∘
= = 1.20804 kg/m3
𝑅𝑇 (8314.46)(293.15)
𝑝𝑀 (101325)(28.96)
𝜌mix,𝑒 = ∘ air = = 1.20390 kg/m3
𝑅𝑇 (8314.46)(293.15)
These densities differ by a very small amount, so a large number of digits must be carried. Then,
(9.806)(1.20804 − 1.20390)𝐷3
Ra𝐷 = = 3.49 × 108 𝐷3
(1.206)(1.51 × 10−5 )(6.35 × 10−6 )
Because the density difference is small, so is the Rayleigh number: the initial value (for 𝐷 = 2.5 cm)
is only 5446. However, from eqn. (*), we find that the Nusselt number varies between 2 and 6.35,
so that both terms in eqn. (*) remain important.
A mass balance on the mothball gives us
𝑑 𝜋
(𝜌PDB,solid 𝐷3 ) = −𝜋𝐷2 𝑔𝑚 Δ𝑚PDB = 𝜋𝐷2 𝑔𝑚 𝑚PDB,𝑠
𝑑𝑡 6
𝜋 𝑑𝐷
(𝜌PDB,solid 𝐷2 ) = −𝜋𝐷2 𝑔𝑚 𝑚PDB,𝑠
2 𝑑𝑡
and with eqn. (**)
𝑑𝐷 2𝑚PDB,𝑠 𝜌𝒟PDB,air 2𝑚PDB,𝑠
= −𝑔𝑚 = [2 + 0.5058 (3.49 × 108 𝐷3 )1/4 ]
𝑑𝑡 𝜌PDB,solid 𝐷 𝜌PDB,solid
2 + 69.13 𝐷3/4 2(1.206)(6.23 × 10−6 )(4.267 × 10−3 )
= −( )( )
𝐷 1145
2 + 69.13 𝐷3/4
= (−5.600 × 10−11 )( )
𝐷
This equation can be integrated from the initial diameter 𝐷0 to zero:
𝑇 𝐷0 =0.025 m
−11 𝐷 𝑑𝐷
(5.600 × 10 ) ∫ 𝑑𝑡 = ∫ = 6.287 × 10−5
0 0 (2 + 69.13 𝐷3/4 )
where the integral was evaluated numerically (in this case, using wolframalpha.com). Solving for
the time, 𝑇 :
6.287 × 10−5 Answer
𝑇= = 1.123 × 106 sec = 13.0 days ⟵−−−−−−−
5.600 × 10−11
The mothball lasts about two weeks.

373
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Comment 1: The vapor pressure rises rapidly with temperature. If the closet is warmer, the
mothball lifetime will be much shorter.
Comment 2: Air circulation in the closet (as from drafts, HVAC systems, or room scale natural
convection) would increase the mass transfer coefficient and shorten the lifetime of the mothball.
Such circulation is more likely than not.
Comment 3: If the mothball is placed in a small closed container, such as a drawer or box,
the concentration of paradichlorobenzene may eventually rise to match the vapor pressure of
paradichlorobenzene at the surface of the mothball. In this equilibrium condition, the sublima-
tion rate would drop to zero.
Comment 4: Paradichlorobenzene is considered to pose health hazards. Its use in mothballs has
been banned in the EU since 2008.

374
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.47 In Section 11.5, 𝜌𝒟12 or 𝑐𝒟12 were at times assumed to be independent
of position. Consider this approximation for gases. (a) Do these two groups depend on pressure,
temperature, or the proportions of species 1 and 2? Are isobaric conditions necessary to hold
either group constant? (b) For what type of mixture is 𝜌𝒟12 most sensitive to composition? What
does this indicate about mole versus mass-based analysis? (c) Do Pr or Sc depend on composition,
temperature, or pressure?

Solution
a) For ideal gases
𝑝 𝑝
∘
𝑐= and 𝜌 = 𝑀 ∘
𝑅𝑇 𝑅𝑇
where 𝑀 depends on the mole or mass fractions of species 1 and 2 from eqns. (11.8). The
diffusion coefficient from eqn. (11.108),
(1.8583 × 10−7 )𝑇 3/2 1 1
𝒟12 = +
2
𝑝𝜎12 Ω𝐷 √ 𝑀1 𝑀2
does not depend on mass or mole fraction. Recall that Ω𝐷 depends on temperature. Thus,
𝑇 1/2
𝑐𝒟12 ∝ = function of 𝑇 only
Ω𝐷
𝑀𝑇 1/2
𝜌𝒟12 ∝ = function of 𝑇 and mole fraction
Ω𝐷
Neither group depends on pressure, so isobaric conditions are irrelevant. ⟵−−−−−−−
Answer
b) The molar mass is
𝑀 = 𝑥 1 𝑀1 + 𝑥 2 𝑀2
When the spatial variations of mole fraction are large, 𝑀 can change significantly with
position. However, if 𝑀1 ≈ 𝑀2 , the value of 𝑀 is not sensitive to changes in mole fraction.
The sensitivity to composition is greatest when 𝑀1 and 𝑀2 are very different. ⟵−−−−−−−
Answer
So, 𝜌𝒟12 is nearly constant in isothermal mixtures of gases with similar molar mass, or
when the composition does not vary much. On the hand, 𝑐𝒟12 stays constant in an isothermal
mixture of any pair of gases even if the concentration changes are large. ⟵−−−−−−−
Answer

c) We have
𝜈 𝜇𝑐𝑝
Pr =
=
𝛼 𝑘
and we know from Section 11.10 that 𝜇 and 𝑘 depend on 𝑇 and composition, but not on 𝑝.
Also, 𝑐𝑝 is constant for an ideal gas. So, Pr is a function of temperature and composition.
Likewise,
𝜈 𝜇
Sc = =
𝒟12 𝜌𝒟12
From the result of part (a), the denominator depends on 𝑇 and composition. Thus Sc is also
a function of 𝑇 and composition.
Pr and Sc depend on temperature and composition, but neither depends on pressure.⟵−−−−−
Answer

375
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.48 A dilute aqueous solution containing potassium ions is subjected to a 1 V/cm
electric field. A measurement suggests that the K+ ions move at 4 × 10−4 cm/s in response to the
field. Estimate the effective radius of K+ ions if the solution is at 300 K. The charge of an electron
is −1.609 × 10−19 C and 1 V/m = 1 N/C.

Solution The force on a K+ ion is

𝐹 = (1.609 × 10−19 C)(1V/cm) = 1.609 × 10−17 N
From the Einstein relation, eqn. (11.111), the diffusion coefficient is
𝑣 +
𝒟K+,H2O = 𝑘𝑏 𝑇 ( K )
𝐹K+
(4 × 10−4 cm/s)(0.01 m/cm)
= (1.38065 × 10−23 J/K)(300 K)
(1.609 × 10−17 N)
= 1.03 × 10−9 m2 /s
K+ ions have a radius similar to water molecules, so eqn. (11.114b) is applicable. Rearranging that
equation and taking 𝜇H2O from Table A.3,
𝑘𝑏 𝑇
𝑅K+ =
4𝜋𝜇H2O 𝒟K+,H2O
(1.38065 × 10−23 J/K)(300 K)
=
4𝜋(8.54 × 10−4 kg/m⋅s)(1.03 × 10−9 m2 /s)
Answer
= 3.75 × 10−10 m = 3.75 Å ⟵−−−−−−−

Comment 1: Ions in aqueous solution are surrounded by a shell of water molecules which are
attracted by hydrogen bonding. This hydration shell can be several times the size of an isolated ion.
For potassium, the hydration shell may consist of 5–7 water molecules [1]. The average reported
radius of a hydrated potassium ion is 3.32 Å [2]
Comment 2: The “experiment” described here is completely made up!

References:
[1] John Burgess, “Solvation numbers.” In John Burgess (ed.), Ions in Solution, Woodhead
Publishing, 1999, Chapter 2, pp. 28–35, doi:10.1533/9781782420569.28.
[2] Bruce Railsback, Some Fundamentals of Mineralogy and Geochemistry, http://railsback.org/Fun-
damentals/IonicChargeRadiusPlot06P.pdf

376
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 11.49

(11.115)

(11.115)

(11.115)

(11.115)

(11.114b) (11.115) (11.115)

(11.114b)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 377

 Problem 11.50 Use eqn. (11.116) to calculate the dynamic viscosity, 𝜇, of gaseous methane,
hydrogen sulfide, and nitrous oxide, under the following conditions: 250 K and 1 atm, 500 K and
1 atm, 250 K and 2 atm, 250 K and 12 atm, 500 K and 12 atm. Is the calculation possible in every
case?

Solution
The dynamic viscosity of ideal gases is independent of pressure. However, we must also consider
whether the fluid remains gaseous or liquefies as the pressure is increased. The six conditions
described involve a maximum pressure of 12 atm and a minimum temperature of 250 K. So, we can
start by checking whether each fluid is gaseous at 250 K and 12 atm, which requires us to obtain data
outside this textbook, as from the NIST Chemistry Webbook, https://webbook.nist.gov/chemistry/.
To check for phase transitions, we can look at the 12 atm isotherm for each fluid to see whether
the fluid remains gaseous at 250 K. Nitrous oxide remains gaseous at 12 atm. Hydrogen sulfide at
250 K liquefies at a pressure of 4.83 atm, so, we cannot evaluate gaseous 𝜇 for H2 S at 250 K and
12 atm. We must also check the 500 K isotherm, and we find that H2 S remains gaseous. Methane
also remains a gas under the given conditions (in fact, these temperatures exceed the methane’s
critical point temperature of 190.6 K, with pressures below the critical point pressure of 46.1 bar,
so methane is clearly gaseous).
Equation (11.25) is
√𝑀𝐴 𝑇
𝜇𝐴 = (2.6693 × 10−6 ) 2 (11.116)
𝜎𝐴 Ω𝜇
so we need to find 𝜎, 𝑀, and Ω𝜇 (𝑇) for each gas at each temperature.

𝑇 = 250 K 𝑇 = 500 K
Gas 𝜎 [Å] 𝜀/𝑘B [K] 𝑀 [kg/kmol] Ω𝜇 𝜇 [kg/m-s] Ω𝜇 𝜇 [kg/m-s]
CH4 3.758 148.6 16.04 1.254 9.545 × 10−6 1.010 1.676 × 10−5
H2 S 3.623 301.1 34.08 1.747 1.075 × 10−5 1.261 2.105 × 10−5
N2 O 3.828 232.4 44.01 1.531 1.248 × 10−5 1.145 2.360 × 10−5
Equation (11.125) does not depend on pressure, so these values apply at 1 atm, 2 atm, and 12 atm,
with the exception of H2 S for which the value does not apply at 12 atm.
Comment: These predictions can be compared to the more complex reference correlations in the
NIST Webbook (below, in kg/m-s). The predictions are within ±4% or better of the NIST values.

𝑇 = 250 K 𝑇 = 500 K
Gas 𝑝 = 1 atm 𝑝 = 12 atm 𝑝 = 1 atm 𝑝 = 12 atm
CH4 9.471 × 10−6 9.647 × 10−6 1.690 × 10−5 1.701 × 10−5
H2 S 1.030 × 10−5 liquid 2.108 × 10−5 2.121 × 10−5
N2 O 1.249 × 10 −5
1.299 × 10−6 2.357 × 10−5 2.375 × 10−5

378
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 11.51

eqns. (11.119)

(11.116) (11.116)

(11.118)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 379

 (11.119ac)

(7.21) 7.1
(7.21)

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 380

 11.52

(11.116) (11.117)

compare to
eqns.(6.45cd)

(11.118)

Pr = 16/21. H2O may also be more susceptible to vibrational modes.

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 381

 Ar N2 H2 O
T [K] Pr T [K] Pr T [K] Pr
Eucken - 0.666 - 0.737 - 0.762
Data 100 0.692 100 0.738 373 1.018
200 0.669 300 0.721 473 0.961
300 0.667 500 0.709 573 0.941
400 0.666 800 0.707 673 0.924
500 0.666 1200 0.712 873 0.899
700 0.665 1600 0.715 1073 0.882

The Eucken relation predicts the data for argon to within 4% at 100 K, but within 0.5% or
better for T≥ 200 K. For nitrogen, the agreement is within 4% over the entire temperature
range. For steam, the disagreement ranges from 16% to 34%; coupling between rotational
and vibrational modes may be a factor.

Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV 382

 Problem 11.53 Three simplifications of the Maxwell-Stefan equations, (11.139), lead to
an effective, binary Fickian diffusion coefficient, 𝒟𝑖𝑚 . (a) Show that Blanc’s law, eqn. (11.109),
applies when species 𝑖 at low concentration diffuses into an otherwise uniform mixture. (b) Show
that if Ð𝑖𝑗 has the same value for each pair of species and the mixture is ideal, then 𝒟𝑖𝑚 = Ð𝑖𝑗 .
(c) Show that if species 𝑖 = 1, 2, 3, … , 𝑛 − 1 all have low concentrations, then 𝒟𝑖𝑚 = Ð𝑖,𝑛 , where
species 𝑛 is the dominant species (or solvent), as stated in eqn. (11.110).

Solution
a) Because species 𝑖 is dilute (𝑥1 ≪ 1), it may be considered ideal, with 𝛾𝑖 = 1. Then Γ𝑖𝑗 = 𝛿𝑖𝑗 ,
from eqn. (11.137b). The Maxwell-Stefan equations, (11.139) are
𝑛−1 𝑛
𝑥𝑖 𝐽𝑗 ⃗ − 𝑥𝑗 𝐽𝑖⃗
∑ Γ𝑖𝑗 ∇T,p 𝑥𝑗 = ∇T,p 𝑥𝑖 = ∑
𝑗=1 𝑗=1
𝑐Ð𝑖𝑗

and because species other than 𝑖 are approximately uniform, 𝐽𝑗 ⃗ ≅ 0 for 𝑗 ≠ 𝑖:

𝑛 𝑛
𝑥𝑗 𝐽𝑖⃗ 𝐽𝑖⃗ 𝑥
∇T,p 𝑥𝑖 = − ∑ =− ∑ 𝑖
𝑗=1
𝑐Ð𝑖𝑗 𝑐 𝑗=1 Ð𝑖𝑗
𝑗≠𝑖 𝑗≠𝑖
Putting this into the form of Fick’s law
𝑛
𝑥𝑖
𝐽𝑖⃗ = −𝑐∇T,p 𝑥𝑖 ∑= −𝑐𝒟𝑖𝑚 ∇T,p 𝑥𝑖
/𝑗=1 Ð𝑖𝑗
𝑗≠𝑖
where 𝑛
𝑥𝑖 Answer
−1
𝒟𝑖𝑚 =∑ ⟵−−−−−−−
Ð
𝑗=1 𝑖𝑗
𝑗≠𝑖

b) For an ideal mixture (such as an ideal gas mixture), Γ𝑖𝑗 = 𝛿𝑖𝑗 . The Maxwell-Stefan equations,
(11.139), reduce to eqns. (11.129b). Then
𝑛
𝑥𝑖 𝐽𝑗 ⃗ − 𝑥𝑗 𝐽𝑖⃗
∇T,p 𝑥𝑖 = ∑
𝑗=1
𝑐Ð𝑖𝑗
𝑛
1
= ∑ (𝑥 𝐽 ⃗ − 𝑥𝑗 𝐽𝑖⃗)
𝑐Ð𝑖𝑗 𝑗=1 𝑖 𝑗

⎛ 𝑛 𝑛 ⎞
1 𝐽⃗
= ⎜𝑥𝑖 ∑ 𝐽𝑗 ⃗ −𝐽𝑖⃗ ∑ 𝑥𝑗 ⎟ = − 𝑖
𝑐Ð𝑖𝑗 ⎜ 𝑗=1 𝑗=1 ⎟
𝑐Ð𝑖𝑗
⏟ ⏟
⎝ =0 =1 ⎠
so that
𝐽𝑖⃗ = −𝑐Ð𝑖𝑗 ∇T,p 𝑥𝑖 = −𝑐𝒟𝑖𝑚 ∇T,p 𝑥𝑖
with
Answer
𝒟𝑖𝑚 = Ð𝑖𝑗 ⟵−−−−−−−

383
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 c) Because species 1, 2, 3, … , 𝑛 − 1 have low concentrations, they behave as ideal: Γ𝑖𝑗 = 𝛿𝑖𝑗 .
The Maxwell-Stefan equations again reduce to eqns. (11.129b). Further, with 𝑥𝑖 ≪ 𝑥𝑛 ≅ 1
for 𝑖 ≠ 𝑛, many terms are negligible:
𝑛
𝑥𝑖 𝐽𝑗 ⃗ − 𝑥𝑗 𝐽𝑖⃗
∇T,p 𝑥𝑖 = ∑
𝑗=1
𝑐Ð𝑖𝑗
𝑛 𝑛
𝐽𝑗 ⃗ 𝑥𝑗 𝐽𝑖⃗
= 𝑥𝑖 ∑ −∑
𝑗=1
𝑐Ð𝑖𝑗 𝑗=1 𝑐Ð𝑖𝑗
𝑛
𝐽𝑗 ⃗ 𝑥 𝐽⃗
≅ 𝑥𝑖 ∑ − 𝑛𝑖
𝑗=1
𝑐Ð𝑖,𝑗 𝑐Ð𝑖,𝑛
𝐽𝑖⃗
≅−
𝑐Ð𝑖,𝑛
and so
Answer
𝐽𝑖⃗ = −𝑐Ð𝑖,𝑛 ∇T,p 𝑥𝑖 𝑖 ≠ 𝑛 ⟵−−−−−−−
We see that 𝒟𝑖𝑚 = 𝒟𝑖,𝑛 = Ð𝑖,𝑛 .

384
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.54 A dilute solute 𝐴 in a liquid solution diffuses down the gradient of the
chemical potential, 𝜇𝐴 , because a gradient in potential energy causes a force per mole of 𝐴.
Compute this force and combine the result with the molar form of Fick’s law to derive the Einstein
relation, eqn. (11.111). Note that the Einstein relation considers the force per molecule and that
𝑅∘ = 𝑘𝐵 𝑁𝐴 , where 𝑁𝐴 is Avogadro’s number.

Solution The gradient in potential energy, or chemical potential, provides the force that
drives the diffusion of a solute 𝐴 against the opposing drag of the solvent molecules. From classical
mechanics, the force 𝐹 ⃗ resulting from a gradient in a potential energy 𝐸 is
𝑑𝐸
𝐹 ⃗ = −∇𝐸 or in 1-D 𝐹 = −
𝑑𝑧
In the case of a dilute solute, the potential energy is the chemical potential of species 𝐴 in J/mol,
eqn. (11.131):
𝜇𝐴 (𝑥) = 𝜇
𝐴 + 𝑅∘ 𝑇 ln(𝑥𝐴 )
The force driving species 𝐴 down the potential gradient, against the drag of solvent 𝐵, is then
𝑑𝜇𝐴 𝑅∘ 𝑇 𝑑𝑥𝐴
𝐹𝐴𝐵 = − =− (*)
𝑑𝑧 𝑥𝐴 𝑑𝑧
in units of N/mol, as given by eqn. (11.134). The drag force of the solvent on the solute is equal
and opposite the force caused by the potential gradient (i.e., acceleration is zero).
From here, we use the molar form of Fick’s law, eqn. (11.30), taken as an empirical premise:
𝑑𝑥𝐴
𝐽𝐴 = −𝑐𝒟𝐴𝐵
𝑑𝑧
Combining these equations,
𝑅∘ 𝑇 𝐽𝐴
𝐹𝐴𝐵 = (**)
𝑐𝐴 𝒟𝐴𝐵
Further, we recall that 𝐽𝐴 is defined by eqn. (11.23) as

𝐽𝐴 = 𝑐𝐴 (𝑣𝐴 − 𝑣∗ )

We put this into eqn. (**), and also notice that Einstein stated the drag per molecule, 𝐹𝐴 = 𝐹𝐴𝐵 /𝑁𝐴 :
𝑘𝐵 𝑇(𝑣𝐴 − 𝑣∗ )
𝐹𝐴 =
𝒟𝐴𝐵
To complete the calculation, we note that solute 𝐴 is dilute, so that eqn. (11.22) gives 𝑣∗ ≅ 𝑣𝐵 :
𝑘𝐵 𝑇(𝑣𝐴 − 𝑣𝐵 )
𝐹𝐴 =
𝒟𝐴𝐵
The velocity difference is just 𝑣𝐴̂ in eqn. (11.111), so we can rearrange to finish:
Answer
𝒟𝐴𝐵 = 𝑘B 𝑇 (𝑣𝐴̂ /𝐹𝐴 ) ⟵−−−−−−−

385
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.55 In Section 11.11, we evaluated the drag on one species as it diffuses past
another. Show that this force also leads to the Einstein relation, eqn. (11.111). Note that the Einstein
relation considers the force per molecule and that 𝑅∘ = 𝑘𝐵 𝑁𝐴 , where 𝑁𝐴 is Avogadro’s number.

Solution We evaluated the drag force on the diffusing solute 𝐴 by the solvent 𝐵, per unit
volume, in eqn. (11.126):
𝑥 𝑥 (𝑣 − 𝑣𝐵 )
𝑓𝐴𝐵 = (𝑐𝑅∘ 𝑇) 𝐴 𝐵 𝐴
Ð𝐴𝐵
Since species 𝐴 is dilute, 𝑥𝐵 ≅ 1; and 𝑐𝑖 = 𝑥𝑖 𝑐:
𝑐 𝑅∘ 𝑇(𝑣𝐴 − 𝑣𝐵 )
𝑓𝐴𝐵 = 𝑖
Ð𝐴𝐵
Einstein’s relation gives the drag force 𝐹𝐴 per molecule (not per mole). The number of molecules
of species 𝑖 per unit volume is 𝑐𝑖 𝑁𝐴 . Dividing
𝑅∘ 𝑇(𝑣𝐴 − 𝑣𝐵 ) 𝑘𝐵 𝑇(𝑣𝐴 − 𝑣𝐵 )
𝐹𝐴 = =
𝑁𝐴 Ð𝐴𝐵 Ð𝐴𝐵
With 𝑣𝐴̂ = 𝑣𝐴 − 𝑣𝐵 , we obtain the final result
Answer
𝒟𝐴𝐵 = 𝑘B 𝑇 (𝑣𝐴̂ /𝐹𝐴 ) ⟵−−−−−−−

386
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.56 Consider a solvent 𝑛 that contains 𝑛 − 1 dilute solutes with 𝑥𝑖 ≪ 𝑥𝑛 ≅ 1.
(a) Work Problem 11.53c. (b) Show that 𝑣∗⃗ ≅ 𝑣𝑛⃗ ≅ 𝑣 ⃗ if the solutes are dilute. (c) Suppose the
solutes are charged ions, such as Na+ and Cl – for table salt dissolved in water. An electric potential
gradient, −∇𝜙, applied to the solution will create a force per mole of ion, −𝑧𝑖 ℱ ∇𝜙, where 𝑧𝑖 is the
ion’s valence (±1 for Na+ and Cl – ) and ℱ = 9.6485 × 104 C/mol is Faraday’s constant. Starting
with the Maxwell-Stefan equations, derive the Nernst-Planck equation:
ℱ
𝑁𝑖⃗ = −𝑐𝒟𝑖,𝑛 ∇𝑥𝑖 + 𝑐𝑖 𝑣𝑛⃗ − 𝑐𝑖 𝒟𝑖,𝑛 𝑧𝑖 ∘ ∇𝜙
𝑅𝑇

Solution
b) With eqn. (11.22), if 𝑥𝑖 ≪ 𝑥𝑛 ≅ 1,
𝑐𝑣∗⃗ = ∑ 𝑐𝑖 𝑣𝑖⃗ = 𝑐 ∑ 𝑥𝑖 𝑣𝑖⃗ ≅ 𝑐𝑥𝑛 𝑣𝑛⃗ ≅ 𝑐𝑣𝑛⃗
𝑖 𝑖
so 𝑣 ⃗ ≅ 𝑣𝑖⃗ . With eqn. (11.16), if 𝑚𝑖 ≪ 𝑚𝑛 ≅ 1,
∗

𝜌𝑣 ⃗ = ∑ 𝜌𝑖 𝑣𝑖⃗ = 𝜌 ∑ 𝑚𝑖 𝑣𝑖⃗ ≅ 𝜌𝑚𝑛 𝑣𝑛⃗ ≅ 𝜌𝑣𝑛⃗

𝑖 𝑖

so 𝑣 ⃗ ≅ 𝑣𝑛⃗ ≅ 𝑣∗⃗ . Note that this condition implies that 𝐽𝑛⃗ ≡ 𝑐𝑛 (𝑣𝑛 − 𝑣∗ ) ≅ 0, by eqn. (11.23).

c) We add the electrostatic force to the left-hand side of eqn. (11.138):

𝑛
𝑥𝑖 𝐽𝑗 ⃗ − 𝑥𝑗 𝐽𝑖⃗
−𝑐𝑖 ∇𝜇𝑖 − 𝑐𝑖 𝑧𝑖 ℱ ∇𝜙 = −𝑐𝑅 𝑇 ∑ ∘
(*)
𝑗=1
𝑐Ð𝑖𝑗
In other words, we add the electrostatic force to the chemical potential driving force per unit
volume, −𝑐𝑖 ∇𝜇𝑖 , and equate the sum to the drag force from other species.
For dilute solutions, we have the same simplification as in Problem 11.53c, so that
𝑥𝑖 𝐽𝑗 ⃗ − 𝑥𝑗 𝐽𝑖⃗ 𝐽⃗
∑⃗ 𝑗=1
𝑛
≅− 𝑖
𝑐Ð𝑖𝑗 𝑐Ð𝑖,𝑛
and because dilute solutions are nearly ideal
𝑐𝑖 ∇𝜇𝑖 ≅ 𝑐𝑅∘ 𝑇∇𝑥𝑖
Putting these modifications into eqn. (*)
𝑅∘ 𝑇 ⃗
−𝑐𝑅∘ 𝑇∇𝑥𝑖 − 𝑐𝑖 𝑧𝑖 ℱ∇𝜙 = 𝐽
Ð𝑖,𝑛 𝑖
or
ℱ
𝐽𝑖⃗ = −𝑐Ð𝑖,𝑛 ∇𝑥𝑖 − 𝑐𝑖 Ð𝑖,𝑛 𝑧𝑖 ∇𝜙
𝑅∘ 𝑇
From part (b), 𝑣∗⃗ ≅ 𝑣𝑛⃗ ≅ 𝑣.⃗ Then, with 𝑁𝑖⃗ ≡ 𝐽𝑖⃗ + 𝑐𝑖 𝑣∗⃗ ≅ 𝐽𝑖⃗ + 𝑐𝑖 𝑣𝑛⃗ and 𝒟𝑖,𝑛 = Ð𝑖,𝑛 from
Problem 11.53c,
ℱ Answer
𝑁𝑖⃗ = −𝑐𝒟𝑖,𝑛 ∇𝑥𝑖 + 𝑐𝑖 𝑣𝑛⃗ − 𝑐𝑖 𝒟𝑖,𝑛 𝑧𝑖 ∘ ∇𝜙 ⟵−−−−−−−
𝑅𝑇
The Nernst-Planck equation is widely used in the study of electrochemical systems. The
terms on the right-hand side represent diffusion, convection, and electromigration.

387
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.57 A liquid mixture of acetone and methanol sits at the bottom of a Stefan tube
that is 23.8 cm tall. The top is swept by a flow of pure air. Let species 1 be acetone, species 2
be methanol, and species 3 be air. The temperature is 328.5 K, and the pressure is 745.2 mmHg.
The mole fractions just above the liquid are 𝑥1,𝑠 = 0.319 and 𝑥2,𝑠 = 0.528. For these gases,
Ð12 = 0.0848 cm2 /s, Ð13 = 0.1372 cm2 /s, and Ð23 = 0.1991 cm2 /s [1].

a) Write the Maxwell-Stefan equations as a matrix o.d.e. for the vector X(𝑧) = {𝑥1 (𝑧), 𝑥2 (𝑧)}.
What is the mole flux of air, 𝑁3 ? What are the boundary conditions at the liquid surface
(𝑧 = 0) and at the top of the tube (𝑧 = 23.8 cm)?
b) Use the software of your choice to solve your equation from part (a) and plot the three
concentrations as a function of 𝑧. Note that the mole fluxes 𝑁1 and 𝑁2 are not known, so
that the equations must be solved iteratively to determine value correspond to the boundary
condition at the top of the tube. [Ans: 𝑁1 = 1.782 × 10−7 mol/cm2 s, 𝑁2 = 3.126 ×
10−7 mol/cm2 s.]
c) If part (b) seems too complex, instead use the answers given for part (b) to solve the equations
as if the fluxes were known. Does your solution meet the boundary conditions at the top of
the tube? (It should.)
d) Does 𝐽air = 0 at 𝑧 = 0? Explain.

Solution

a) Only two of the three MS equations are independent because ∑ 𝑥𝑖 = 1. Further, in a quasi-
steady system, the mole fluxes are constant along the length of the tube. And, because air
does not pass through the liquid surface, 𝑁3 = 0.
Equations (11.129a) can be written for 𝑥1 and 𝑥2 as

0
3
𝑑 𝑥1 𝑁𝑗 − 𝑥𝑗 𝑁1 𝑥 𝑁 − 𝑥2 𝑁1 𝑥1 𝑁
3 − 𝑥3 𝑁1
𝑐 𝑥1 = ∑ = 1 2 +
𝑑𝑧 𝑗=1
Ð1𝑗 Ð12 Ð13
0
3
𝑑 𝑥2 𝑁𝑗 − 𝑥𝑗 𝑁2 𝑥 𝑁 − 𝑥1 𝑁2 𝑥2 𝑁
3 − 𝑥3 𝑁2
𝑐 𝑥2 = ∑ = 2 1 +
𝑑𝑧 𝑗=1
Ð2𝑗 Ð21 Ð23

But 𝑥3 = 1 − 𝑥1 − 𝑥2 and Ð𝑖𝑗 = Ð𝑗𝑖 , so

𝑑 𝑥 𝑁 − 𝑥2 𝑁1 (1 − 𝑥1 − 𝑥2 )𝑁1
𝑐 𝑥1 = 1 2 −
𝑑𝑧 Ð12 Ð13
𝑁2 𝑁1 𝑁 𝑁 𝑁1
= 𝑥1 ( + ) + 𝑥2 ( 1 − 1 ) −
Ð12 Ð13 Ð13 Ð12 Ð13
𝑑 𝑥 𝑁 − 𝑥1 𝑁2 (1 − 𝑥1 − 𝑥2 )𝑁2
𝑐 𝑥2 = 2 1 −
𝑑𝑧 Ð12 Ð23
𝑁 𝑁 𝑁 𝑁 𝑁2
= 𝑥 1 ( 2 − 2 ) + 𝑥2 ( 1 + 2 ) −
Ð12 Ð13 Ð12 Ð23 Ð23

388
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 And in matrix form, these equations are
𝑁2 𝑁1 𝑁1 𝑁 𝑁1
⎛Ð + Ð − 1⎞ ⎛
𝑑 𝑥1 12 13 Ð13 Ð12 𝑥1 Ð ⎞
𝑐 ( )= ⎜ ⎟ ( ) − ⎜ 13 ⎟
𝑑𝑧 ⏟𝑥2 ⎜ 𝑁2 − 𝑁2 𝑁1 + 𝑁2 ⎟ 𝑥2 ⎜ 𝑁2 ⎟
=X ⎝ Ð12 Ð13 Ð12 Ð23 ⎠
⏟⎵⎵⎵⎵⎵⎵⎵⏟⎵⎵⎵⎵⎵⎵⎵⏟ ⏟⎝⎵Ð
⏟23
⎵⏟⎠
=A =B
or just
𝑑 Answer
𝑐 X = 𝐀𝐗 − 𝐁 ⟵−−−−−−−
𝑑𝑧
The boundary conditions are
0.319 0 Answer
X=( ) at 𝑧 = 0, X = ( ) at 𝑧 = 23.8 cm ⟵−−−−−−−
0.528 0

b) An example of appropriate code follows, using Matlab.

1 % Solve ODEs iteratively to find fluxes that match bc : x1 = x2 = 0

at top of tube
2 % species 1 = acetone ; species 2 = methanol ; species 3 = air
3 % Carty & Schrodt had : N (1) = 1.779 e -7 , N (2) = 3.121 e -7 [ mol / cm ^2 - s ]
4 %
5 % xi = mole fraction ;
6 % D_ij = MS diffusivity [ cm ^2/ s ]
7 % N( i ) = mole flux [ mol / cm ^2 - s ]
8
9 global zmax
10 zmax = 23.8; % length of Stefan tube [ cm ]
11
12 % iterative root finding : get fluxes that make top - end mole
fractions zero .
13 fun = @root2d ;
14 N0 = [0 ,0]; % initial guess for fluxes
15 options = optimset ( ' Display ' ,' off '); % don ' t show fsolve message
16 N = fsolve ( fun ,N0 , options ) % iteratively solve ODE system in fun
17
18 [x1 ,x2 , x3 ] = Conc (N ) % display final solutions
19
20 % Plot results and save as a pdf file
21 clf
22 fplot (x2 , ' LineWidth ' ,2) ;
23 ax = gca ;
24 ax . FontSize = 16;
25 xlabel ( ' Distance from surface , z [ cm ] ')
26 ylabel ( ' Mole fraction ')
27 axis ([0 zmax 0 1])
28 grid on
29 hold on
30 fplot (x1 , ' LineWidth ' ,2)
31 fplot (x3 , ' LineWidth ' ,2)

389
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 32 legend ( ' Methanol ' , ' Acetone ' , ' Air ' ,' Location ' , ' northwest ')
33 print ( ' Ternary_Stefan_tube_concentrations ' , ' - dpdf ' , ' - r600 '); % 600
dpi , pdf
34
35 % % function that evaluates mole fractions at zmax as function of
flux vector N
36 function F = root2d (N);
37 [x1 ,x2 , x3 ] = Conc (N );
38 global zmax ;
39 F (1) = double ( subs ( x1 , zmax )) ; % convert symbols to numbers for
fsolve
40 F (2) = double ( subs ( x2 , zmax )) ;
41 end
42
43 % % function that solves M - S odes for given values of flux vector N
44 function [x1 ,x2 , x3 ] = Conc (N );
45 % data reported by Carty & Schrodt (1975)
46 D12 = 0.0848; % cm ^2/ s
47 D13 = 0.1372;
48 D23 = 0.1991;
49 cden = 3.6375 e -5; % molar density [ mol / cm3 ] calculated for their
conditions
50
51 % initial conditions ( bottom boundary condition )
52 x10 = 0.319; % solution is very sensitive to these ICs .
53 x20 = 0.528; % ICs are mole fractions above liquid surface .
54
55 % Maxwell - Stefan equations in matrix form , after eliminating eqn
for x3
56 B1 = -N (1) / D13 ;
57 B2 = -N (2) / D23 ;
58 A11 = N (1) / D13 +N (2) / D12 ;
59 A12 = N (1) *(1/ D13 -1/ D12 );
60 A21 = N (2) *(1/ D23 -1/ D12 );
61 A22 = N (1) / D12 +N (2) / D23 ;
62
63 % https :// www . mathworks . com / help / symbolic / solve -a - system - of -
differential - equations . html
64 % X ' = AX + B
65 sympref ( ' FloatingPointOutput ' , true );
66 syms x1 (z) x2 ( z) x3 (z)
67 X = [ x1 ; x2 ];
68 A = [ A11 A12 ; A21 A22 ]/ cden ;
69 B = [ B1 ; B2 ]/ cden ;
70 C = X (0) == [ x10 ; x20 ]; % initial conditions
71
72 odes = diff (X ) == A*X + B;
73 [ x1Sol ( z) , x2Sol ( z)] = dsolve ( odes ,C) ;
74
75 % solutions for yi ( z )

390
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 76 x1 (z) = simplify ( x1Sol (z)) ; % acetone
77 x2 (z) = simplify ( x2Sol (z)) ; % methanol
78 x3 (z) = 1 - x1 (z ) - x2 (z ) ; % air
79 end

The computed results are shown in Figure 1. The code returns 𝑁1 = 1.782×10−7 mol/cm2 s
and 𝑁2 = 3.126 × 10−7 mol/cm2 s.

1
Methanol
0.9 Acetone
Air
0.8

0.7
Mole fraction

0.6

0.5

0.4

0.3

0.2

0.1

0
0 5 10 15 20
Distance from surface, z [cm]

Figure 1. Concentration distribution of ternary gas mixture in Stefan tube, com-

puted from Maxwell-Stefan equations. Liquid mixture of acetone and methanol (at
𝑧 = 0) evaporates into tube swept by pure air at the top (𝑧 = 23.8 cm).

c) The simpler code follows. The results are almost exactly the same, and the top boundary
condition is well met.

1 % Values reported by Carty & Schrodt (1975) in cgs units

2 % 1 = acetone ; 2 = methanol ; 3 = air
3 % y = mole fraction ; N_i = mole flux ; D_ij = MS diffusivity ;
4 N1 = 1.779 e -7; % mol / cm ^2 - s
5 N2 = 3.121 e -7;
6 D12 = 0.0848; % cm ^2/ s
7 D13 = 0.1372;
8 D23 = 0.1991;
9

391
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 10 cden = 3.6375 e -5; % molar density [ mol / cm3 ] calculated for their
conditions
11
12 y10 = 0.319; % the solution is very sensitive to these ICs .
13 y20 = 0.528;
14 zmax = 23.8; % length of Stefan tube [ cm ]
15
16 % Maxwell - Stefan equations in matrix form , after eliminating x3
17 B1 = - N1 / D13 ;
18 B2 = - N2 / D23 ;
19 A11 = N1 / D13 + N2 / D12 ;
20 A12 = N1 *(1/ D13 -1/ D12 );
21 A21 = N2 *(1/ D23 -1/ D12 );
22 A22 = N1 / D12 + N2 / D23 ;
23
24 % https :// www . mathworks . com / help / symbolic / solve -a - system - of -
differential - equations . html
25 % Y ' = AY + B
26 sympref ( ' FloatingPointOutput ' , true );
27 syms y1 (z) y2 ( z) y3 (z)
28 Y = [ y1 ; y2 ];
29 A = [ A11 A12 ; A21 A22 ]/ cden ;
30 B = [ B1 ; B2 ]/ cden ;
31 C = Y (0) == [ y10 ; y20 ];
32
33 odes = diff (Y ) == A*Y + B
34 [ y1Sol ( z) , y2Sol ( z)] = dsolve ( odes ,C) ;
35
36 y1Sol (z) = simplify ( y1Sol (z))
37 y2Sol (z) = simplify ( y2Sol (z))
38
39 % Air mole fraction can be obtained from y1 + y2 + y3 = 1
40 y3 (z) = 1 - y1Sol (z) - y2Sol (z);
41
42 clf
43 fplot ( y2Sol , ' LineWidth ' ,2) ;
44 ax = gca ;
45 ax . FontSize = 16;
46 xlabel ( ' Distance from surface , z [ cm ] ')
47 ylabel ( ' Mole fraction ')
48 axis ([0 zmax 0 1])
49 grid on
50 hold on
51 fplot ( y1Sol , ' LineWidth ' ,2)
52 fplot (y3 , ' LineWidth ' ,2)
53 legend ( ' Methanol ' , ' Acetone ' , ' Air ' ,' Location ' , ' northwest ')
54 print ( ' Ternary_Stefan_tube_concentrations ' , ' - dpdf ' , ' - r600 '); % 600
dpi , pdf

392
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 d) We now know that 𝑁 = 𝑁1 + 𝑁2 = 1.782 × 10−7 + 3.126 × 10−7 = 4.908 × 10−7 mol/cm2 s.
We also know that 𝑁3 = 𝑁air = 0 and that 𝑥3 = 𝑥air = 1 − 0.319 − 0.528 = 0.153 at 𝑧 = 0.
Thus
0
𝐽air = 𝐽3 = 𝑁
3 − 𝑥3 𝑁 = −(0.153)(4.908 × 10−7 ) = −0.751 × 10−7 mol/cm2 s
The diffusional flux of air is not zero because the net mass transfer away from the surface
produces convection. Because the air mole flux at the surface is zero, the concentration
gradient will have steepened until air diffuses toward the surface at just the rate it is being
convected away from the flux.

Comment 1: The linear differential equations can of course be solved analytically (as an initial
value problem from 𝑧 = 0), but the boundary conditions at the top must still be matched by a
correct choice of the mole fluxes. Matlab returns the following equations for the concentration
profiles with the correct fluxes, for 𝑧 in cm:
𝑥1 = 0.3631 − 0.0020 𝑒0.1591 𝑧 − 0.0421 𝑒0.0789 𝑧
𝑥2 = 0.0020 𝑒0.1591 𝑧 − 0.1109 𝑒0.0789 𝑧 + 0.6369
𝑥3 = 0.1530 𝑒0.0789 𝑧

Comment 2: The curves are plotted in Fig. 2 with one of the data sets from Carty’s dissertation
(Table 1).
Table 1. Data from Carty’s Ternary Run T-5 [2]

Position [cm] 𝑥acetone 𝑥methanol 𝑥air

4.472 0.2684 0.4568 0.2748
6.954 0.2515 0.4574 0.2911
9.437 0.2332 0.4218 0.3450
11.920 0.2205 0.3799 0.3996
14.402 0.1856 0.3047 0.5096
16.884 0.1551 0.2457 0.5992
19.367 0.1137 0.1668 0.7195

References:
[1] R. Carty and J. Thomas Schrodt, “Concentration Profiles in Ternary Gaseous Diffusion,” Indus-
trial & Engineering Chemistry Fundamentals, 14(3):276–278, 1975. doi:10.1021/i160055a025.

[2] Ronald Haden Carty, Concentration Profiles in a Ternary Gaseous Diffusion System, PhD
Dissertation, Department of Chemical Engineering, University of Kentucky, 1973.

393
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 1
Methanol
0.9 Acetone
Air
0.8 Methanol
Acetone
0.7 Air
Mole fraction

0.6

0.5

0.4

0.3

0.2

0.1

0
0 5 10 15 20
Distance from surface, z [cm]

Figure 2. Concentration distribution of ternary gas mixture in Stefan tube, com-

puted from Maxwell-Stefan equations, with measurements from [2].

394
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 Problem 11.58 A simple model of an ablating heat shield assumes that ablated material is
rapidly removed without effecting the flow and that the aerodynamic heat flux, 𝑞aero , is constant.
a) If the shield is at 𝑇0 when heating starts at 𝑡 = 0, how long until the surface reaches the
ablation temperature, 𝑇𝑎 , for either melting or sublimation?
b) Once ablation starts, assume the surface recedes at a constant speed, 𝑉𝑎 . Find the temperature
distribution in the material below the surface for 𝑡 > 𝑡𝑎 . Hint: Change to a coordinate
𝜂 = 𝑥 − 𝑉𝑎 (𝑡 − 𝑡𝑎 ) attached to the receding surface, and ignore the initial distribution of 𝑇.
c) Use an energy balance to determine 𝑉𝑎 in terms of material properties and 𝑞aero .
d) What combination of material properties best reduces the heat conducted into the heat shield?

Heat Shield
surface,
t < ta
surface, t > ta

qaero

Va

0 η

0 x

11.58.auxlock

Solution
a) Using eqn. (5.56), the time to reach the ablation temperature, 𝑡𝑎 , is
𝑞aero 𝛼𝑡𝑎
𝑇𝑎 = 𝑇0 + 2
𝑘 √ 𝜋
or
2 2
𝜋 𝑘(𝑇𝑎 − 𝑇0 ) 𝜋 (𝑇 − 𝑇0 ) Answer
𝑡𝑎 = [ ] = (𝑘𝜌𝑐𝑝 ) [ 𝑎 ] ⟵−−−−−−−
𝛼 2𝑞aero 4 𝑞aero

b) The governing equation is simply the 1D unsteady heat conduction equation

𝜕𝑇 𝜕2 𝑇
=𝛼 2
𝜕𝑡 𝜕𝑥
where: 𝑥 is measured from the initial position of the surface; 𝑇 → 𝑇0 as 𝑥 → ∞; and
𝑇 = 𝑇𝑎 for 𝑥 = 𝑉𝑎 𝑡. The last boundary condition is at the ablating surface for 𝑡 > 𝑡𝑎 . Define

395
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 a similarity coordinate, 𝑥 − 𝑉𝑎 (𝑡 − 𝑡𝑎 ), which is attached to the receding surface. Then
𝜕𝑇 𝜕𝑇 𝜕𝜂 𝑑𝑇
= = −𝑉𝑎
𝜕𝑡 𝜕𝜂 𝜕𝑡 𝑑𝜂
𝜕𝑇 𝜕𝑇 𝜕𝜂 𝑑𝑇
= =
𝜕𝑥 𝜕𝜂 𝜕𝑥 𝑑𝜂
𝜕 𝑇 𝜕𝜂 𝜕 𝜕𝑇 𝑑 2 𝑇
2
= =
𝜕𝑥2 𝜕𝑥 𝜕𝜂 𝜕𝜂 𝑑𝜂2
so that
𝑑𝑇 𝑑 2𝑇
−𝑉𝑎 =𝛼 2
𝜕𝜂 𝑑𝜂
Integrating once (use the trick 𝑦 = 𝑑𝑇/𝑑𝜂 to do this)
𝑑𝑇 𝑉𝜂
= 𝐶 exp(− 𝑎 )
𝑑𝜂 𝛼
where 𝐶 is a constant. A second integration gives us (collecting constant terms)
𝑇(𝜂) = 𝐶1 + 𝐶2 exp(−𝑉𝑎 𝜂/𝛼)
The surface is at 𝜂 = 0 where 𝑇 = 𝑇𝑎 , and for 𝜂 → ∞, 𝑇 → 𝑇0 . Applying these conditions
Answer
𝑇(𝜂) = 𝑇0 + (𝑇𝑎 − 𝑇0 ) exp(−𝑉𝑎 𝜂/𝛼) ⟵−−−−−−−
c) At the surface, 𝑞aero is divided between conduction into the material and the latent heat of
melting or subliming (call this 𝐿 in J/kg). Material is removed from the surface at a rate of
𝜌𝑉𝑎 in kg/m2 s. Then
𝜕𝑇 |
𝑞aero = −𝑘 | + 𝜌𝐿𝑉𝑎
𝜕𝜂 |𝜂=0
and the first right-hand side term is
𝜕𝑇 |
𝑞cond = −𝑘 | = −𝑘(𝑇𝑎 − 𝑇0 )(−𝑉𝑎 /𝛼) = +𝜌𝑐𝑝 (𝑇𝑎 − 𝑇0 )𝑉𝑎
𝜕𝜂 |𝜂=0
so that
𝑞aero Answer
𝑉𝑎 = ⟵−−−−−−−
𝜌[𝑐𝑝 (𝑇𝑎 − 𝑇0 ) + 𝐿]
d) The heat shield should prevent heat from being conducted into the spacecraft behind the
shield, so the following ratio should be low
𝑞cond 𝜌𝑉𝑎 𝑐𝑝 (𝑇𝑎 − 𝑇0 ) 1 Answer
= = ⟵−−−−−−−
𝑞aero 𝜌𝑉𝑎 [𝑐𝑝 (𝑇𝑎 − 𝑇0 ) + 𝐿] 1 + 𝐿/𝑐𝑝 (𝑇𝑎 − 𝑇0 )
Thus, to ensure that 𝑞cond ≪ 𝑞aero , the shield material should have 𝐿 ≫ 𝑐𝑝 (𝑇𝑎 − 𝑇0 ). In
addition, 𝑉𝑎 should minimized, which favors high values of both 𝜌𝐿 and 𝜌𝑐𝑝 (𝑇𝑎 − 𝑇0 ).

Comment: The initial condition for the moving surface problem should be the temperature dis-
tribution of the solution of part (a) at 𝑡𝑎 . We have not provided an equation for that temperature
profile in the text, but it is [2]
2𝑞aero √𝛼𝑡𝑎 𝑥
𝑇(𝑥, 𝑡𝑎 ) = 𝑇0 + ierfc ( )
𝑘 2√𝛼𝑡𝑎

396
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV
 where ierfc is an integral of the error function, also tabulated as a special function. This profile is
similar to a decaying exponential function in 𝑥. Since we know 𝑡𝑎 from part (a), we can do some
algebra to show
𝑇(𝑥, 𝑡𝑎 ) − 𝑇0 𝑥
= √𝜋 ierfc ( ) (*)
(𝑇𝑎 − 𝑇0 ) 2√𝛼𝑡𝑎
where we note that ierfc (0) = 1/√𝜋.
On the other hand, evaluating the result of part (b) at 𝜂 = 𝑥 − 𝑉𝑎 (𝑡𝑎 − 𝑡𝑎 ) gives a different
decaying function of 𝑥
𝑇(𝑥, 𝑡𝑎 ) = 𝑇0 + (𝑇𝑎 − 𝑇0 ) exp(−𝑉𝑎 𝑥/𝛼)
or
𝑇(𝑥, 𝑡𝑎 ) − 𝑇0
= exp(−𝑉𝑎 𝑥/𝛼) (**)
(𝑇𝑎 − 𝑇0 )
Equations (*) and (**) both go from 1 at 𝑥 = 0 to 0 as 𝑥 → ∞, but they are otherwise different
functions. We cannot do a detailed comparison unless we choose a specific material. Given that
we have already made the vast simplification of a constant heat flux, the smaller difference in the
shape of the initial condition is not very significant.
References:
[1] P. M. Sforza, “Thermal protection systems.” In P. M. Sforza, ed., Manned Spacecraft Design
Principles, Chapter 9, pp. 391-452. Butterworth-Heinemann, Boston, 2016. doi:10.1016/B978-0-
12-804425-4.00009-X
[2] H. S. Carslaw and J. C. Jaeger, Conduction of Heat in Solids, 2nd ed., Oxford University
Press, 1959, §2.9.

397
Copyright 2024, John H. Lienhard, V and John H. Lienhard, IV