NivaarExam PrepOfficial exam papers ↗

21-Mat-A2 Materials Transport Phenomena · December 2019

Question 5 of 5: Boltzmann–Matano Interdiffusion Coefficient

Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)

Notes on this paper

National Exams — December 2019 — 12-MTL-A2 Transport Phenomena in Materials Engineering. Three-hour, open-book exam (one textbook of the candidate's choice permitted, with margin notations, no loose notes); any non-communicating calculator permitted. Each of the five questions is worth 25 points, and any four constitute a complete paper — only the first four questions as they appear in the answer book are marked. All five are solved below for completeness. Candidates were told to state all assumptions clearly.

Reference texts: Bird, R. B., Stewart, W. E. & Lightfoot, E. N., Transport Phenomena — power-law non-Newtonian flow between parallel plates and annular fully-developed duct flow (Questions 1, 3), matching this paper's own Appendix A conservation-equation tables; Incropera, F. P. et al., Fundamentals of Heat and Mass Transfer — natural-convection correlations for a horizontal and a vertical flat plate (Question 2); Geiger, G. H. & Poirier, D. R., Transport Phenomena in Materials Processing — mould-resistance-controlled solidification of castings (Question 4); Shewmon, P. G., Diffusion in Solids — the Boltzmann–Matano graphical method for a concentration-dependent interdiffusion coefficient (Question 5).

Question 5: Boltzmann–Matano Interdiffusion Coefficient (25 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Given.

QuantitySymbolValue
Interdiffusion time$t$28 h $=1.008\times10^5$ s
Temperature—1200 °C (FCC alloy)
Target concentration$C_1$$0.02\ \text{mol/cm}^3$
Concentration profile$C_B(x)$read from the paper's own printed graph (0.10 → 0 mol/cm³, 0.01 cm grid scale)

Find. $D_{AB}$ at $C_B=0.02\ \text{mol/cm}^3$ by the Boltzmann–Matano graphical method.

[Figure not reproduced. See the official exam paper.]

Approach. The Boltzmann substitution $\eta=(x-x_M)/\sqrt{t}$ reduces Fick’s second law with a concentration-dependent diffusivity, $\partial C/\partial t=\partial/\partial x[D(C)\,\partial C/\partial x]$, to an ordinary differential equation in $\eta$ alone. Integrating it from the far field down to a chosen composition $C_1$ gives the Matano relation $$D(C_1)=-\frac{1}{2t}\left(\frac{dx}{dC}\right)_{C_1}\int_{0}^{C_1}\left(x-x_M\right)dC$$ in which distances must be measured from the Matano interface $x_M$ — the plane that balances the two areas between the profile and its terminal compositions, $\int_0^{C_L}(x-x_M)\,dC=0$. The printed graph’s own left-hand edge is not that plane: the drawn curve leaves its $C_B=0.10$ plateau only at $x\approx0.078$ cm, so $x_M$ has to be located from the curve before anything is integrated.

The profile was read from the printed figure and calibrated against the figure’s own scale bar: its two arrow tips touch adjacent vertical grid lines and the dimension is labelled 0.01 cm, so one grid square is 0.01 cm. On that scale the drawn curve runs from $x=0.078$ cm at $C_B\approx0.10$ to $x=0.236$ cm at $C_B\approx0.005$.

  1. Locate the Matano interface. Writing $x=x_M+s\,\mathrm{erfc}^{-1}(2C_B/C_L)$ with the left-hand terminal composition $C_L=0.10$, the 44 digitized points fall on a straight line ($R^2=0.987$) — that is, this couple is an ordinary error-function profile. The least-squares line gives $$x_M=0.159\ \text{cm},\qquad s\equiv2\sqrt{Dt}=0.0581\ \text{cm}$$ The equal-area condition $\int_0^{0.10}(x-x_M)\,dC=0$, integrated numerically over the digitized curve with the fitted tails, returns the same plane ($x_M=0.1588$ cm), and the half-height point $C_B=C_L/2=0.05$ sits at $x=0.155$ cm, as it must for a symmetric couple.
  2. Read $x$ at $C_1=0.02$. From the digitized curve, $$x(C_B=0.02)=0.191\ \text{cm}\qquad\Rightarrow\qquad x-x_M=0.032\ \text{cm}$$
  3. Tangent slope at $C_1=0.02$. Differentiating the fitted profile at that composition — equivalently, the tangent drawn on Fig. 4 — gives $$\left(\frac{dx}{dC}\right)_{0.02}=-1.47\ \text{cm}\,/\,(\text{mol/cm}^3)$$ A local straight-line fit through the printed-curve points bracketing $C_B=0.02$ returns $-1.95$ instead; the gap between the two is the honest precision of a tangent read from a printed graph, and it is carried through below.
  4. Matano integral (area). Trapezoidal integration of $(x-x_M)$ over $0\le C_B\le0.02$, with the $C_B\to0$ tail supplied by the fitted profile: $$\int_0^{0.02}\left(x-x_M\right)dC=1.20\times10^{-3}\ \text{cm}\cdot(\text{mol/cm}^3)$$
  5. Assemble $D_{AB}$. With $t=28\times3600=1.008\times10^5\ \text{s}$: $$D_{AB}=\frac{1}{2t}\left|\frac{dx}{dC}\right|\int_0^{0.02}(x-x_M)\,dC=\frac{(1.47)(1.20\times10^{-3})}{2(1.008\times10^{5})}$$ $$\boxed{D_{AB}(C_B=0.02)\approx8.7\times10^{-9}\ \text{cm}^2/\text{s}=8.7\times10^{-13}\ \text{m}^2/\text{s}}$$ Using the steeper raw tangent instead would give $1.2\times10^{-8}\ \text{cm}^2/\text{s}$, so to the precision a graphical read can support the answer is $D_{AB}\sim1\times10^{-8}\ \text{cm}^2/\text{s}=1\times10^{-12}\ \text{m}^2/\text{s}$.
  6. Independent check. Because the profile is an error function, its fitted width alone fixes the diffusivity with no tangent and no area at all: $$D=\frac{(s/2)^2}{t}=\frac{(0.0291)^2}{1.008\times10^{5}}=8.4\times10^{-9}\ \text{cm}^2/\text{s}$$ within 4% of the Matano result. Both sit in the published band for substitutional interdiffusion in FCC metals at 1200 °C (e.g. Cu–Ni couples report $\tilde D\sim10^{-13}$–$10^{-12}\ \text{m}^2/\text{s}$ near this temperature).
QuantityValue
Distance calibration (printed scale bar)1 grid square = 0.01 cm
Matano interface, $x_M$0.159 cm
$x$ at $C_B=0.02$0.191 cm  ($x-x_M=0.032$ cm)
Tangent slope, $(dx/dC)_{0.02}$−1.47 cm/(mol/cm³)
Matano area, $\int_0^{0.02}(x-x_M)\,dC$$1.20\times10^{-3}$ cm·(mol/cm³)
$D_{AB}(C_B=0.02)$$8.7\times10^{-9}\ \text{cm}^2/\text{s}=8.7\times10^{-13}\ \text{m}^2/\text{s}$
Cross-check from the fitted erf width$8.4\times10^{-9}\ \text{cm}^2/\text{s}$ (4% apart)
Check — the scale and the origin, which is what this answer turns on

Scale. The distance axis carries no numbers. The only calibration printed on the page is the pair of dimension arrows near the foot of the plot, whose tips touch two adjacent vertical grid lines and which are labelled 0.01 cm — so one grid square is 0.01 cm and the whole drawn profile is about 0.16 cm wide.

Origin. The Matano relation measures distance from the Matano interface, not from the plotted axis. Integrating from the graph’s own $x=0$ instead would raise the integral from $1.20\times10^{-3}$ to $4.35\times10^{-3}$ and the answer from $8.7\times10^{-9}$ to $3.2\times10^{-8}$ cm²/s — nearly a factor of four — so the choice is stated here rather than left implicit.

Reading precision. The tangent is the weakest step ($-1.47$ from the fitted profile against $-1.95$ from a local fit to the raw points), which puts the answer between $8.7\times10^{-9}$ and $1.2\times10^{-8}$ cm²/s. Quoted to the one significant figure the method actually supports, $D_{AB}\approx1\times10^{-8}$ cm²/s $=1\times10^{-12}$ m²/s.

Back to the paper →