Steady Diffusion in Liquids, Solids and Curved Shells

Transport Phenomena · Diffusion

Dr. Ray’s Learning Center · covers: Steady Diffusion in Liquids, Solids and Curved Shells

  • 18 to 24 min
  • 4 figures
  • 3 worked examples
  • 4 practice sets

Before this lesson:

  • Steady Diffusion: Fick’s Law, Equimolar Counterdiffusion and Stagnant Films

A water layer carries ethanol out of a solvent, a rubber sheet slowly lets carbon dioxide through, and a rubber hose leaks hydrogen even with no hole in it. All three are steady diffusion, and all three use Fick’s law. Each adds one new detail: what the total concentration is in a liquid, what sets the concentration at the surface of a solid, and what area to use when the area changes along the path.

In this lesson you will work the liquid film, the solid membrane and the curved wall in turn (Figure 1). The equations come from Geankoplis, Transport Processes and Unit Operations, 3rd edition, Sections 6.3 and 6.5.

  1. Liquid filmcav, xBMc_{av},\ x_{BM}average c, log mean of B
  2. thensolids
  3. SolidcA=SpA22.414c_A = \dfrac{S\,p_A}{22.414}dissolve, then diffuse
  4. thencurved walls
  5. Curved wallWA=NAAW_A = N_A Amolar flow is constant
Figure 1. Three steady cases, one law: a liquid needs the right total concentration, a solid needs the solubility, and a curved wall needs the right mean area.

Learning Objectives

After this lesson, you should be able to:

  1. Calculate the steady flux of a solute through a liquid film, with the average total concentration and the log-mean mole fraction of the stagnant liquid.
  2. Find the dissolved concentration at the face of a solid from the solubility, and calculate the flux through a solid sheet.
  3. Explain why the molar flow, not the flux, is constant across a curved wall.
  4. Calculate the leak rate through a cylindrical or spherical wall with the right mean area.

Key Ideas

1. Diffusion in Liquids

A liquid obeys the same two flux forms as a gas, but its total concentration is not PRT\frac{P}{RT}. Density and mean molar mass change with composition, so Geankoplis uses the average of the two ends (Eq. 6.3-2):

cav=12(ρ1M1+ρ2M2)c_{av} = \dfrac{1}{2}\left(\dfrac{\rho_1}{M_1} + \dfrac{\rho_2}{M_2}\right)

For equimolar counterdiffusion (Eq. 6.3-1):

NA=DABcavz2−z1(xA1−xA2)N_A = \dfrac{D_{AB}\,c_{av}}{z_2-z_1}\,(x_{A1}-x_{A2})

The common liquid case is A diffusing through B that does not move, for example ethanol through a water film whose far side is an organic solvent that water does not enter. Then NB=0N_B = 0 and A rides a small bulk flow (Eqs. 6.3-3 and 6.3-4):

NA=DABcav(z2−z1)xBM(xA1−xA2)N_A = \dfrac{D_{AB}\,c_{av}}{(z_2-z_1)\,x_{BM}}\,(x_{A1}-x_{A2})
xBM=xB2−xB1ln⁡xB2xB1x_{BM} = \dfrac{x_{B2}-x_{B1}}{\ln\dfrac{x_{B2}}{x_{B1}}}

Here xBMx_{BM} is the log mean of the mole fractions of B at the two faces. It is less than 1, so dividing by it raises the flux. For a dilute solution xBM→1x_{BM} \to 1 and the flux reduces to NA=DAB(cA1−cA2)z2−z1N_A = \frac{D_{AB}(c_{A1}-c_{A2})}{z_2-z_1} (Eq. 6.3-5). Liquid diffusivities can change markedly with concentration, so use a value measured for the range you are working in.

Key idea: in a liquid, use cavc_{av} from the densities and molar masses at both faces, and divide by the log mean xBMx_{BM} when B is stagnant.

Worked Example 1: Ethanol Through a Stagnant Water Film

Problem: Ethanol (A) diffuses at steady state through a water (B) film 1.5 mm thick at 293 K. On one side the solution is 16.8 wt% ethanol with density 972.8 kg m⁻³; on the other it is 6.8 wt% ethanol with density 988.1 kg m⁻³. Water does not diffuse (NB=0N_B = 0). Use DAB=0.740×10−9m2s−1D_{AB} = 0.740 \times 10^{-9}\,\text{m}^{2}\,\text{s}^{-1}, MA=46.05M_A = 46.05 and MB=18.02kgkmol−1M_B = 18.02\,\text{kg}\,\text{kmol}^{-1}. Find NAN_A. The data are from Geankoplis Example 6.3-1 (p. 399, diffusivity from Treybal, 1963), which uses a 2.0 mm film.

Step 1. Ethanol mole fractions. Take 100 kg of solution.

xA1=16.846.0516.846.05+83.218.02=0.07323xA2=6.846.056.846.05+93.218.02=0.02776\begin{aligned} x_{A1} &= \dfrac{\dfrac{16.8}{46.05}}{\dfrac{16.8}{46.05} + \dfrac{83.2}{18.02}} = 0.07323 \\ x_{A2} &= \dfrac{\dfrac{6.8}{46.05}}{\dfrac{6.8}{46.05} + \dfrac{93.2}{18.02}} = 0.02776 \end{aligned}

Step 2. Mean molar mass at each face.

M1=100kg(16.846.05+83.218.02)kmol=20.07kgkmol−1M2=100kg(6.846.05+93.218.02)kmol=18.80kgkmol−1\begin{aligned} M_1 &= \dfrac{100\,\text{kg}}{\left(\dfrac{16.8}{46.05} + \dfrac{83.2}{18.02}\right)\text{kmol}} = 20.07\,\text{kg}\,\text{kmol}^{-1} \\ M_2 &= \dfrac{100\,\text{kg}}{\left(\dfrac{6.8}{46.05} + \dfrac{93.2}{18.02}\right)\text{kmol}} = 18.80\,\text{kg}\,\text{kmol}^{-1} \end{aligned}

Step 3. Average total concentration (Eq. 6.3-2, p. 398).

cav=12(972.820.07+988.118.80)=48.4704+52.55852=50.514kmolm−3\begin{aligned} c_{av} = \dfrac{1}{2}\left(\dfrac{972.8}{20.07} + \dfrac{988.1}{18.80}\right) = \dfrac{48.4704 + 52.5585}{2} = 50.514\,\text{kmol}\,\text{m}^{-3} \end{aligned}

Step 4. Log mean of the water mole fractions.

xB1=1−0.07323=0.92677,xB2=1−0.02776=0.97224xBM=0.97224−0.92677ln⁡0.972240.92677=0.045470.04790=0.9493\begin{aligned} x_{B1} &= 1-0.07323 = 0.92677, \qquad x_{B2} = 1-0.02776 = 0.97224 \\ x_{BM} &= \dfrac{0.97224-0.92677}{\ln\dfrac{0.97224}{0.92677}} = \dfrac{0.04547}{0.04790} = 0.9493 \end{aligned}

Step 5. Substitute. Write the thickness in m.

NA=0.740×10−9×50.5141.5×10−3×0.9493(0.07323−0.02776)=1.194×10−6kmolm−2s−1≈1.19×10−6kmolm−2s−1\begin{aligned} N_A = \dfrac{0.740 \times 10^{-9} \times 50.514}{1.5 \times 10^{-3} \times 0.9493}\,(0.07323-0.02776) = 1.194 \times 10^{-6}\,\text{kmol}\,\text{m}^{-2}\,\text{s}^{-1} \approx 1.19 \times 10^{-6}\,\text{kmol}\,\text{m}^{-2}\,\text{s}^{-1} \end{aligned}

Final Answer: NA≈1.19×10−6kmolm−2s−1N_A \approx 1.19 \times 10^{-6}\,\text{kmol}\,\text{m}^{-2}\,\text{s}^{-1}

Check: leaving out xBMx_{BM} gives a flux 5.1% low, so bulk flow matters even at 7 mol% ethanol. Using pure water, 998.2318.02=55.40kmolm−3\frac{998.23}{18.02} = 55.40\,\text{kmol}\,\text{m}^{-3}, in place of cavc_{av} overstates the flux by 9.7%.

Note: this lesson uses the log mean xBMx_{BM} of Eq. 6.3-4. Geankoplis’s Example 6.3-1 uses the arithmetic average, here 0.94950.9495 against 0.94930.9493, a difference of about 0.02%. With the book’s own 2.0 mm film and the arithmetic average, carried without rounding to the end, the same method gives NA=8.9505×10−7kmolm−2s−1N_A = 8.9505 \times 10^{-7}\,\text{kmol}\,\text{m}^{-2}\,\text{s}^{-1}. Geankoplis reports 8.99×10−78.99 \times 10^{-7} because of small slips in its printed intermediate values (for example M2M_2 printed as 18.75 against 18.80 here).

Your turn: ethanol through a water film

Same steps as Worked Example 1: mole fractions, mean molar masses, cavc_{av}, the log mean xBMx_{BM}, then substitute with the thickness in m. The densities given are from Perry’s Handbook at 20 °C. Enter the flux as, for example, 1.23e-6.

Loading the practice problem…

2. Diffusion Through Solids

A gas such as H2\text{H}_2 or CO2\text{CO}_2 first dissolves in the face of a polymer or rubber, then diffuses through it. The dissolved gas is dilute, so the bulk-flow term is small and is dropped, and Fick’s law with constant DABD_{AB} gives the slab result (Eqs. 6.5-1 and 6.5-2):

NA=DAB(cA1−cA2)z2−z1N_A = \dfrac{D_{AB}\,(c_{A1}-c_{A2})}{z_2-z_1}

The concentration just inside each face is set by the solubility SS, in m³ of gas at STP (0 °C and 1 atm) per m³ of solid per atm of partial pressure (Eq. 6.5-5). Dividing by 22.414 m³(STP) kmol⁻¹ turns it into kmol m⁻³:

cA=SpA22.414c_A = \dfrac{S\,p_A}{22.414}

The concentration jumps at each face: the gas and the solid hold different amounts of A, and only their ratio is fixed (Figure 2). Combining the two equations gives the flux in terms of the permeability PMP_M (Eqs. 6.5-8 and 6.5-9):

NA=PM(pA1−pA2)22.414(z2−z1),PM=DABSN_A = \dfrac{P_M\,(p_{A1}-p_{A2})}{22.414\,(z_2-z_1)}, \qquad P_M = D_{AB}\,S
Dissolved carbon dioxide across a rubber sheetGas at 1.5 atm on the left, rubber 1.0 mm thick in the middle, no carbon dioxide on the right. Inside the rubber the dissolved concentration falls in a straight line from 0.06023 kmol per cubic metre, set by the solubility, to zero, and the flux N A runs left to right.0.000.020.040.060.08Position across the sheetDissolved CO2, kmol m⁻³gaspA1 = 1.5 atmgaspA2 = 0rubber, 1.0 mmcA1 = 0.06023cA2 = 0NA
Figure 2. Worked Example 2: the solubility sets the dissolved concentration just inside each face, and Fick’s law gives a straight line through the rubber.

For layers of different materials in series, the resistances LiPMi\frac{L_i}{P_{Mi}} add (Eq. 6.5-12):

NA=pA1−pA222.414⋅1L1PM1+L2PM2+⋯N_A = \dfrac{p_{A1}-p_{A2}}{22.414}\cdot\dfrac{1}{\dfrac{L_1}{P_{M1}} + \dfrac{L_2}{P_{M2}} + \cdots}

Table 6.5-1 (p. 411) lists DABD_{AB}, SS and PMP_M for several gases in rubber. These are separately measured data, so the tabulated PMP_M differs from the product DABSD_{AB}S by 0.6% to 3.4% for the hydrogen, oxygen, nitrogen and carbon dioxide rows. This lesson always uses DABSD_{AB}S. For gases such as H2\text{H}_2 and O2\text{O}_2 in metals such as Ni, Cd and Pt, the flux is approximately proportional to pA1−pA2\sqrt{p_{A1}}-\sqrt{p_{A2}}, so Eq. 6.5-8 does not hold.

Key idea: in a solid, find the face concentration from the solubility, cA=SpA22.414c_A = \frac{S\,p_A}{22.414}, then use Fick’s law for a slab. Permeability is PM=DABSP_M = D_{AB}S.

Worked Example 2: Carbon Dioxide Through a Rubber Sheet

Problem: A vulcanized rubber sheet 1.0 mm thick has CO2\text{CO}_2 at a partial pressure of 1.5 atm on one side and none on the other, at 298 K. From Geankoplis Table 6.5-1, DAB=0.11×10−9m2s−1D_{AB} = 0.11 \times 10^{-9}\,\text{m}^{2}\,\text{s}^{-1} and S=0.90m3(STP)m−3atm−1S = 0.90\,\text{m}^{3}(\text{STP})\,\text{m}^{-3}\,\text{atm}^{-1}. Find the steady flux NAN_A. Use PM=DABSP_M = D_{AB}S from these two values, not the tabulated PMP_M.

Step 1. Dissolved concentration at each face.

cA1=SpA122.414=0.90×1.522.414=0.06023kmolm−3,cA2=0\begin{aligned} c_{A1} = \dfrac{S\,p_{A1}}{22.414} = \dfrac{0.90 \times 1.5}{22.414} = 0.06023\,\text{kmol}\,\text{m}^{-3}, \qquad c_{A2} = 0 \end{aligned}

Step 2. Fick’s law across the slab.

NA=DAB(cA1−cA2)z2−z1=0.11×10−9×0.060231.0×10−3=6.625×10−9kmolm−2s−1≈6.63×10−9kmolm−2s−1\begin{aligned} N_A = \dfrac{D_{AB}\,(c_{A1}-c_{A2})}{z_2-z_1} = \dfrac{0.11 \times 10^{-9} \times 0.06023}{1.0 \times 10^{-3}} = 6.625 \times 10^{-9}\,\text{kmol}\,\text{m}^{-2}\,\text{s}^{-1} \approx 6.63 \times 10^{-9}\,\text{kmol}\,\text{m}^{-2}\,\text{s}^{-1} \end{aligned}

Step 3. Check with the permeability. Here PM=DABSP_M = D_{AB}S, not the tabulated value.

PM=0.11×10−9×0.90=9.90×10−11m3(STP)m−1s−1atm−1NA=9.90×10−11×1.522.414×1.0×10−3=6.625×10−9kmolm−2s−1\begin{aligned} P_M &= 0.11 \times 10^{-9} \times 0.90 = 9.90 \times 10^{-11}\,\text{m}^{3}(\text{STP})\,\text{m}^{-1}\,\text{s}^{-1}\,\text{atm}^{-1} \\ N_A &= \dfrac{9.90 \times 10^{-11} \times 1.5}{22.414 \times 1.0 \times 10^{-3}} = 6.625 \times 10^{-9}\,\text{kmol}\,\text{m}^{-2}\,\text{s}^{-1} \end{aligned}

Final Answer: NA≈6.63×10−9kmolm−2s−1N_A \approx 6.63 \times 10^{-9}\,\text{kmol}\,\text{m}^{-2}\,\text{s}^{-1}

Check: that is about 5.35×10−45.35 \times 10^{-4} m³(STP) of CO2\text{CO}_2 per m² per hour. The tabulated PM=1.01×10−10P_M = 1.01 \times 10^{-10} gives 6.759×10−9kmolm−2s−16.759 \times 10^{-9}\,\text{kmol}\,\text{m}^{-2}\,\text{s}^{-1}, 2.0% higher, only because the table’s DABD_{AB}, SS and PMP_M are separately measured; this lesson uses DABSD_{AB}S throughout.

Your turn: a gas through a rubber sheet

Same steps as Worked Example 2: find cAc_A at both faces from SS and the partial pressures, then divide the difference by the thickness in m.

Loading the practice problem…

3. Curved Shells: Cylinders and Spheres

In a tube wall or a spherical shell the area grows with radius: A=2πrLA = 2\pi r L for a cylinder and 4πr24\pi r^2 for a sphere. At steady state no A builds up anywhere, so the molar flow WA=NAAW_A = N_A A is the same at every radius, while the flux NAN_A falls as the area grows (Figure 3). Integrating Fick’s law with WAW_A constant gives, for a cylinder of length LL (Geankoplis Eqs. 6.5-3 and 6.5-4):

WA=2πLDAB(cA1−cA2)ln⁡r2r1=DABAlm(cA1−cA2)r2−r1,Alm=A2−A1ln⁡A2A1W_A = \dfrac{2\pi L\,D_{AB}\,(c_{A1}-c_{A2})}{\ln\dfrac{r_2}{r_1}} = \dfrac{D_{AB}\,A_{lm}\,(c_{A1}-c_{A2})}{r_2-r_1}, \qquad A_{lm} = \dfrac{A_2-A_1}{\ln\dfrac{A_2}{A_1}}
Flux and molar flow across a wall from r 5 mm to 8 mmThe molar flow is the same at every radius. The flux falls as the area grows: to 0.625 of its inner value at the outer face of a cylinder, and to 0.39 for a sphere.56780.00.20.40.60.81.01.2Radius r, mmValue divided by its value at r1molar flow WA: constantcylinder NA: 0.625sphere NA: 0.39
Figure 3. Across the wall of Worked Example 3 the molar flow WAW_A stays constant while the flux NAN_A falls as 1r\frac{1}{r} in a cylinder and as 1r2\frac{1}{r^2} in a sphere.

For a sphere, A=4πr2A = 4\pi r^2 and the same integration gives

WA=4πDAB(cA1−cA2)1r1−1r2=DAB(4πr1r2)(cA1−cA2)r2−r1W_A = \dfrac{4\pi D_{AB}\,(c_{A1}-c_{A2})}{\dfrac{1}{r_1}-\dfrac{1}{r_2}} = \dfrac{D_{AB}\,(4\pi r_1 r_2)\,(c_{A1}-c_{A2})}{r_2-r_1}

The second form comes from putting the two reciprocals over a common denominator:

1r1−1r2=r2−r1r1r24π1r1−1r2=4πr1r2r2−r1\begin{aligned} \dfrac{1}{r_1}-\dfrac{1}{r_2} &= \dfrac{r_2-r_1}{r_1 r_2} \\ \dfrac{4\pi}{\dfrac{1}{r_1}-\dfrac{1}{r_2}} &= \dfrac{4\pi r_1 r_2}{r_2-r_1} \end{aligned}

So the mean area of a spherical shell is 4πr1r24\pi r_1 r_2. A line of algebra shows that this is the square root of the inner area times the outer area:

4πr1r2=(4πr12)(4πr22)=A1A24\pi r_1 r_2 = \sqrt{(4\pi r_1^{2})(4\pi r_2^{2})} = \sqrt{A_1 A_2}

The same 4πr1r24\pi r_1 r_2 appears in heat conduction through a hollow sphere (Geankoplis Eq. 4.2-14, p. 222; Perry’s Eq. 5-8, p. 5-4). For example, the tube wall of Worked Example 3 (radii 5 and 8 mm) made into a sphere, with the same hydrogen and rubber, has

4πr1r2=4π×0.005×0.008=5.027×10−4m2WA=0.85×10−9×5.027×10−4×0.0035690.003=5.083×10−13kmols−1≈5.08×10−13kmols−1\begin{aligned} 4\pi r_1 r_2 &= 4\pi \times 0.005 \times 0.008 = 5.027 \times 10^{-4}\,\text{m}^{2} \\ W_A &= \dfrac{0.85 \times 10^{-9} \times 5.027 \times 10^{-4} \times 0.003569}{0.003} = 5.083 \times 10^{-13}\,\text{kmol}\,\text{s}^{-1} \approx 5.08 \times 10^{-13}\,\text{kmol}\,\text{s}^{-1} \end{aligned}

The choice of mean matters only for thick shells (Figure 4). At r2r1=1.1\frac{r_2}{r_1} = 1.1 the log-mean area is within 0.08% of the arithmetic mean and the sphere area within 0.5%; at r2r1=2\frac{r_2}{r_1} = 2 the gaps are 3.8% and 20%.

Log-mean and sphere mean areas compared with the arithmetic meanBoth ratios start at 1 for a thin wall. At a radius ratio of 2 the cylinder log-mean area is 3.8 percent below the arithmetic mean and the sphere area 20 percent below. The tube of Worked Example 3, at 1.6, is marked.1.01.52.02.50.60.70.80.91.0Radius ratio r2 to r1Ratio to the arithmetic meanWorked Example 3: 1.6cylinder, log meansphere, 4πr1r2
Figure 4. For a thin wall any mean area will do; for a thick one the cylinder needs AlmA_{lm} and the sphere 4πr1r24\pi r_1 r_2, which fall well below the arithmetic mean.

If a gas A diffuses outward from a sphere through a stagnant gas B, the log term of the stagnant film carries over unchanged (Geankoplis Eq. 6.2-31, p. 391, written with c=PRTc = \frac{P}{RT} and yA=pAPy_A = \frac{p_A}{P}):

WA=4πcDAB1r1−1r2ln⁡1−yA21−yA1,c=PRTW_A = \dfrac{4\pi\,c\,D_{AB}}{\dfrac{1}{r_1}-\dfrac{1}{r_2}}\,\ln\dfrac{1-y_{A2}}{1-y_{A1}}, \qquad c = \dfrac{P}{RT}

Key idea: in a curved wall the molar flow WAW_A is constant, not the flux. Use the slab equation with the log-mean area AlmA_{lm} for a cylinder and 4πr1r24\pi r_1 r_2 for a sphere.

Worked Example 3: Hydrogen Leaking Through a Rubber Tube

Problem: Hydrogen at 2.0 atm flows through a vulcanized rubber tube 2.0 m long with inner radius 5 mm and outer radius 8 mm, at 298 K. The partial pressure of hydrogen outside is zero. From Geankoplis Table 6.5-1, DAB=0.85×10−9m2s−1D_{AB} = 0.85 \times 10^{-9}\,\text{m}^{2}\,\text{s}^{-1} and S=0.040m3(STP)m−3atm−1S = 0.040\,\text{m}^{3}(\text{STP})\,\text{m}^{-3}\,\text{atm}^{-1}. Find the leak rate WAW_A.

Step 1. Dissolved concentration at the inner face.

cA1=0.040×2.022.414=0.003569kmolm−3,cA2=0\begin{aligned} c_{A1} = \dfrac{0.040 \times 2.0}{22.414} = 0.003569\,\text{kmol}\,\text{m}^{-3}, \qquad c_{A2} = 0 \end{aligned}

Step 2. Inner, outer and log-mean areas.

A1=2πr1L=2π×0.005×2.0=0.06283m2,A2=2π×0.008×2.0=0.10053m2Alm=A2−A1ln⁡A2A1=2πL(r2−r1)ln⁡r2r1=2π×2.0×0.003ln⁡0.0080.005=0.037700.4700=0.08021m2\begin{aligned} A_1 &= 2\pi r_1 L = 2\pi \times 0.005 \times 2.0 = 0.06283\,\text{m}^{2}, \qquad A_2 = 2\pi \times 0.008 \times 2.0 = 0.10053\,\text{m}^{2} \\ A_{lm} &= \dfrac{A_2-A_1}{\ln\dfrac{A_2}{A_1}} = \dfrac{2\pi L\,(r_2-r_1)}{\ln\dfrac{r_2}{r_1}} = \dfrac{2\pi \times 2.0 \times 0.003}{\ln\dfrac{0.008}{0.005}} = \dfrac{0.03770}{0.4700} = 0.08021\,\text{m}^{2} \end{aligned}

Step 3. Leak rate.

WA=DABAlm(cA1−cA2)r2−r1=0.85×10−9×0.08021×0.0035690.003=8.111×10−11kmols−1≈8.11×10−11kmols−1\begin{aligned} W_A = \dfrac{D_{AB}\,A_{lm}\,(c_{A1}-c_{A2})}{r_2-r_1} = \dfrac{0.85 \times 10^{-9} \times 0.08021 \times 0.003569}{0.003} = 8.111 \times 10^{-11}\,\text{kmol}\,\text{s}^{-1} \approx 8.11 \times 10^{-11}\,\text{kmol}\,\text{s}^{-1} \end{aligned}

Step 4. Compare with a flat wall. The leak scales with the area used, so divide each other choice by AlmA_{lm}:

A1Alm=0.062830.08021=0.7833,A2Alm=0.100530.08021=1.2533A1+A22Alm=0.081680.08021=1.0183\begin{aligned} \dfrac{A_1}{A_{lm}} &= \dfrac{0.06283}{0.08021} = 0.7833, \qquad \dfrac{A_2}{A_{lm}} = \dfrac{0.10053}{0.08021} = 1.2533 \\ \dfrac{A_1 + A_2}{2A_{lm}} &= \dfrac{0.08168}{0.08021} = 1.0183 \end{aligned}

So the inner area alone gives a leak 21.7% low, the outer area 25.3% high, and the arithmetic mean of the two areas 1.83% high.

Final Answer: WA≈8.11×10−11kmols−1W_A \approx 8.11 \times 10^{-11}\,\text{kmol}\,\text{s}^{-1}, about 2.92×10−42.92 \times 10^{-4} mol h⁻¹

Check: the first form of the cylinder equation in Key Idea 3, with 2πL2\pi L and the logarithm of the radius ratio, gives the same 8.111. With r2r1=1.6\frac{r_2}{r_1} = 1.6 the wall is thick, so the area you use matters far more than which mean you take.

Your turn: a gas leaking from a hollow sphere

Same idea as Worked Example 3, for a sphere: find cA1c_{A1} from SS, use the mean area 4πr1r24\pi r_1 r_2 with the radii in m, then multiply by DAB(cA1−cA2)r2−r1\frac{D_{AB}(c_{A1}-c_{A2})}{r_2-r_1}.

Loading the practice problem…

Common Mistakes

Common mistake: Using the dilute form for a concentrated film. With xBMx_{BM} left out, Worked Example 1 comes out 5.1% low. Divide by the log mean of the B mole fractions whenever B is stagnant.

Common mistake: Using pure water for the total concentration. 998.2318.02=55.40kmolm−3\frac{998.23}{18.02} = 55.40\,\text{kmol}\,\text{m}^{-3} is pure water. The solution has a different density and molar mass at each face: use cavc_{av}.

Common mistake: Treating the partial pressure as a concentration. A gas must dissolve before it diffuses: cA=SpA22.414c_A = \frac{S\,p_A}{22.414}, not pAp_A.

Common mistake: Forgetting the 22.414. SpAS\,p_A is in m³(STP) per m³ of solid. Divide by 22.414 m³(STP) kmol⁻¹ to get kmol m⁻³.

Common mistake: Leaving lengths in mm. Write every thickness and radius in m: 1 mm is 10−310^{-3} m. Otherwise the flux is 1000 times too small.

Common mistake: Using the inner or outer area as if the wall were flat. In Worked Example 3 that is 21.7% low or 25.3% high. Use AlmA_{lm} for a cylinder and 4πr1r24\pi r_1 r_2 for a sphere.

Common mistake: Mixing up the cylinder and sphere means. The log-mean area belongs to the cylinder, 4πr1r24\pi r_1 r_2 to the sphere. For a thick sphere with r2r1=2\frac{r_2}{r_1} = 2 the arithmetic mean area is 25% too large.

Common mistake: Rounding too early. Keep 4 significant figures in the mole fractions, cavc_{av}, xBMx_{BM} and the areas; round only the final answer to 3.

Key Takeaway

Liquids, solids and curved walls all use Fick’s law. In a liquid, use cavc_{av} and divide by xBMx_{BM} when B is stagnant. In a solid, the face concentration is SpA22.414\frac{S\,p_A}{22.414} and the permeability is DABSD_{AB}S. In a curved wall the molar flow is constant, so use the slab equation with the right mean area: AlmA_{lm} for a cylinder, 4πr1r24\pi r_1 r_2 for a sphere.

Practice: randomized problems

Each problem below uses fresh numbers every time and mixes the three kinds of problem in this lesson: work it out, check your answer, then open the worked solution to compare your steps.

Loading the practice problem…

Your results

This summary collects every practice problem you try on this page and suggests what to review next.

Loading your results…

Figures: original work by Dr. Ray’s Learning Center, licensed CC BY 4.0. Methods and data: C. J. Geankoplis, Transport Processes and Unit Operations, 3rd edition (1993): Eqs. 6.3-1 to 6.3-5 (p. 398) for liquids; Example 6.3-1 (p. 399, diffusivity from Treybal, 1963) for the ethanol and water film; Eqs. 6.5-1 to 6.5-12 (pp. 408 to 410) and Table 6.5-1 (p. 411) for solids; Eq. 6.2-31 (p. 391) for a sphere in a stagnant gas; Eq. 4.2-14 (p. 222) for the sphere mean area. Perry’s Chemical Engineers’ Handbook, 9th edition: Eq. 5-8 (p. 5-4) for the sphere mean area and Table 2-58 (p. 2-108) for ethanol and water densities. The practice problems use illustrative conditions and are audited; see the lesson audit log.