Title: EE 5336/7336:
1EE 5336/7336
Integrated Photonics Chapter 4 Step-Index
Circular Waveguides
Marc P. Christensen Assistant Professor Electrical
Engineering Department Southern Methodist
University Dallas, TX 75275 mpc_at_engr.smu.edu (214)
768-1407
2Chapter 4
- Step-index Circular Waveguides
- a.k.a. step-index fiber
- Most prevalent geometry.
- Harder to intuit the modes.
- (at least I think so)
- Same approach
- Pick a co-ordinate system that makes sense
- Write down the wave equation in each region
- Use boundary equations to relate the regions
- Numerically solve any equations which cannot be
solved analytically
3Fiber is everywhere!
- Why?
- No inherent performance advantage
- Cost!
- 100s of 1000s of kms each year
- When you think long haul telecom,
- Think fiber.
- When you think photonic integrated circuitry
(i.e., signal processing and routing) - Think rectangular or ridge waveguides.
4Geometry of a step index fiber
5Solutions for Modes
- Cylindrical Co-ordinate System
- Solve wave equation in this geometry
- 2-D
- Leads to 2 different mode numbers
- Coupling between differing field components.
6The Wave Eqn in Cylindrical Co-ordinates
7Field Coupling
r, f
z
8z-component Wave Eqn.
9Solving the Ez Wave Eqn
10Solving the Ez Wave Eqn
11Solving the Ez Wave Eqn
12Solving the Ez Wave Eqn
13Bessel Functions
- Just a function, no different than Sin or Cos
- Orthogonal to each other.
- Defined everywhere (all r)
- (check out Appendix B for some useful properties,
well introduce them as needed here)
14Bessel Functions of 1st Kind - J
15Modified Bessel Functions of 2nd Kind - K
16Bessel Function Approximations
- At large values of the arguments
17Guiding Condition
18Solution for Ez and Hz
19Take a step back?
20Getting to r and f
21Inside the core (rlta)
22Outside the core (rgta)
23End of Session
24Boundary Conditions
25Boundary Conditions
26Characteristic Equation
27Relating the Coefficients
28Simplest Case is when n 0
29Bessel Identity Simplifies Characteristic Eqn
30Bessel Identity Simplifies Characteristic Eqn
31Mathematica Example
32Hybrid Modes (n !0)
- A0 TE
- B0 TM
- AgtB Ez is larger than Hz so most
- transverse field is H,
- so we call it HE
- AgtB Hz is larger than Ex so most
- transverse field is E,
- so we call it EH
33Meridonal vs. Skew
r, f
z
34Too Ugly to Solve!
35Weakly Guiding Approximation
36Using that same old Bessel Identity
37Now we have
38Degeneracy
- Same Beta (same propagation speed)
- TE0m is degenerate with TM0m
- HEn1,m is degenerate with EHn-1,m
39Hypothesize it and see if it is True
40Is it nearly Transverse E?
41Is it nearly Transverse E?
42Is it nearly Transverse E?
43Is it nearly Transverse E?
44Linearly Polarized Modes
- LP1,m sum of TE0,m or TM0,m and HE2,m
- LPn,m sum of HEn1,m and EHn-1,m
- LP0,m HE1,m
45Linearly Polarized Modes
- Useful, but not real!
- Useful because they tell how a polarized source
will launch into a fiber. - Not real because the weakly guiding approximation
is just that (an approximation), Actual betas
are nearly equal which means that over long
distances this breaks down
46End of Session
47Normalized Frequency and Cutoff
- Determine if a given wavelength will have a given
mode in the fiber. - Recall We are looking for the intersections of
curves. In 1-D we said that every time Tan(kh)
went to Infinity we picked up another mode. - This time the equation contains Jn1/Jn
- So now we pick up a new mode at every root of
the denominator (Jn)
481st Graph of Solutions
49List of Cutoffs
- TE0,m kagtmth root of J0(ka)
- HE1,m kagtmth root of J1(ka)
- EHn,m kagtmth root of Jn(ka)
- HEn,m
50Approximate Graph
51V and lc
- Vkmaxa
- All modes with ka less than V can exist within a
fiber! - We call the cutoff condition Vlt2.405 because
TE0,1, TM0,1, and HE2,1 cease to exist there.
Only HE1,1 will propagate. - Alternatively we can define the wavelength lc
above which V2.405 and the fiber is single mode.
52Example 1
- l1.3 mm, ncore1.5, nclad1.48
- What is the maximum radius of the fiber for which
the fiber will be single mode? - Va (2 p/1.3 mm)Sqrt1.52-1.4822.405
- a2.038 mm
53Example 2
- l1.3 mm, ncore1.5, nclad1.48
- If we increase the radius to a10 mm,
- What modes exist in the waveguide?
- V10 mm(2 p/1.3 mm)Sqrt1.52-1.482
- V11.799
- So now what?
54J0
55J1
56J2
57J3
58J4
59J5
60J6
61J7
62J8
632J1ka J2
642J2ka/2 J3
652J3ka/3 J4
662J4ka/4 J5
672J5ka/5 J6
682J6ka/6 J7
692J7ka/7 J8
702J8ka/8 J9
712J9ka/9 J10
72The Fundamental Mode
- HE11 mode is the lowest mode (cutoff of zero!).
- It looks Gaussian!!!????
73Mode Field Diameter
74So Do I Always Need to Count Every Mode???
- Did I have to before?
- How did I avoid it?
- Can we do the same here?
- Well need to use the approximations for the
Bessel Functions Then as good as the
approximations are, our estimates of modes are
75Characteristic Eqn for LP modes
76Characteristic Eqn for LP modes
77Number of Modes
78Allowed Modes Triangle