Question 5 of 6: Fibre attenuation, modal dispersion and a symbolic dispersion derivation
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Notes on this paper
Paper format. National Examinations, December 2013, 98-Phys-A7 Optics. Three hours, closed book, one approved Casio or Sharp calculator. Question 1 (20 short parts) is mandatory; the rubric asks for any four of Questions 2–6 — all six are worked here, because the set is a study resource.
Reference texts. E. Hecht, Optics, 5th ed.; F. L. Pedrotti, L. M. Pedrotti and L. S. Pedrotti, Introduction to Optics, 3rd ed.
Question 5: Fibre attenuation, modal dispersion and a symbolic dispersion derivation (20 marks)
Rayleigh scaling. Loss coefficient $\alpha\propto1/\lambda^4$, so $$\alpha(1500)=\alpha(900)\left(\frac{900}{1500}\right)^4=(1.5\text{ dB/km})(0.6)^4=\boxed{0.194\text{ dB/km}}$$ — the classic argument for why telecom fibre moved from 850/900 nm to the 1300/1550 nm windows.
Part (b). Given. $n_1=1.46$ (core), $n_2=1.45$ (cladding), $L=1$ km.
Shortest path. The axial ray travels straight down the core: $$t_{short}=\frac{L\,n_1}{c}=\frac{(1000\text{ m})(1.46)}{2.998\times10^8}=\boxed{4.870\ \mu\text{s}}$$
Longest path. The ray guided at the critical angle grazes the core–cladding boundary; its angle from the fibre AXIS satisfies $\cos\theta_{max}=n_2/n_1=1.45/1.46$, giving $\theta_{max}=6.71^\circ$ and a zig-zag path length $L/\cos\theta_{max}=L\,n_1/n_2$: $$t_{long}=\frac{L\,n_1}{n_2}\cdot\frac{n_1}{c}=\frac{L\,n_1^2}{c\,n_2}=\frac{(1000)(1.46)^2}{(2.998\times10^8)(1.45)}=\boxed{4.904\ \mu\text{s}}$$ (path length 1006.9 m).
Modal spread. $$\Delta t=t_{long}-t_{short}=\boxed{33.6\text{ ns}}$$ over 1 km — the intermodal dispersion that limits step-index-fibre bandwidth.
Part (c). No numeric $S_o$, $\lambda_o$ are given — the part is answered symbolically.
(i) Units of $S_o$. $n$ is dimensionless and $\lambda^2-\lambda_o^4/\lambda^2$ has units of length$^2$, so $\boxed{[S_o]=\text{length}^{-2}}$ (e.g. nm$^{-2}$).
(ii) Zero dispersion and $D(\lambda)$. Material dispersion arises because $d^2n/d\lambda^2\ne0$ makes different wavelengths travel at different group velocities, spreading a pulse; the zero-dispersion wavelength $\lambda_o$ is where the dispersion parameter $D(\lambda)=-(\lambda/c)\,d^2n/d\lambda^2$ vanishes. Differentiating twice, $$\frac{dn}{d\lambda}=\frac{S_o}{4}\left(\lambda+\frac{\lambda_o^4}{\lambda^3}\right),\qquad \frac{d^2n}{d\lambda^2}=\frac{S_o}{4}\left(1-\frac{3\lambda_o^4}{\lambda^4}\right)$$ $$\boxed{D(\lambda)=-\frac{\lambda}{c}\frac{d^2n}{d\lambda^2}=\frac{S_o}{4c}\left(\frac{3\lambda_o^4}{\lambda^3}-\lambda\right)}$$
(iii) Group index. $n_g=n-\lambda\,dn/d\lambda$: substituting the two expressions above and simplifying, $$\boxed{n_g(\lambda)=-\frac{S_o}{8}\left(\lambda^2+\frac{3\lambda_o^4}{\lambda^2}\right)}$$
(iv) Transit-time difference. Group velocity $v_g=c/n_g$, so over a length $L$ the transit time is $t=Ln_g/c$; for two wavelengths $\lambda_1,\lambda_2$, $$\boxed{\Delta t=\frac{L}{c}\Big[n_g(\lambda_1)-n_g(\lambda_2)\Big]}=\frac{L S_o}{8c}\left[\left(\lambda_2^2-\lambda_1^2\right)+3\lambda_o^4\left(\frac{1}{\lambda_1^2}-\frac{1}{\lambda_2^2}\right)\right]$$
Check: substituting $\lambda=\lambda_o$ into the D(λ) expression derived directly from the GIVEN n(λ) gives D(λo)=Soλo/(2c), not zero — the printed formula does not algebraically zero its own dispersion at λo (the industry-standard Corning form is D(λ)=(So/4)[λ−λo4/λ3], which does vanish trivially at λo, and may be what the source intended). The calculus above is carried out faithfully on the formula exactly as printed; the concept explanation in (ii) is unaffected.