Optical ring resonators are photonic cavities that circulate selected wavelengths. Their resonance linewidth reveals how long optical energy persists, but a linewidth alone cannot distinguish internal energy loss from escape into the bus waveguide. This distinction is central to ring resonator intrinsic vs loaded Q factor analysis. With a calibrated scan and a suitable coupling model, researchers can extract defensible quality factors instead of assigning every narrow dip to low propagation loss.
What a Resonator Quality Factor Measures
The quality factor, Q, compares energy stored in a resonant mode with energy dissipated per optical cycle. Its physical definition is Q = omega0 times stored energy divided by average power lost, where omega0 is the resonance angular frequency. A larger Q generally indicates a longer energy-storage time and a narrower resonance, provided the mode behaves as a linear, stationary cavity.
For an isolated, approximately Lorentzian resonance, the operational relation is Q = f0 / delta f, where f0 is center frequency and delta f is the full width at half maximum, or FWHM, expressed in frequency. For a sufficiently narrow optical resonance, the equivalent wavelength approximation is Q approximately equal to lambda0 / delta lambda. Here delta lambda must be the linewidth, not the spacing between adjacent resonances. The free spectral range describes mode spacing and is a different measurement.
High Q can improve selectivity and light-matter interaction, but increases sensitivity to temperature drift, fabrication variation, and laser wavelength instability. Match Q to system requirements.
Intrinsic, Coupling, and Loaded Q
Intrinsic Q, written Qi, quantifies loss mechanisms within the resonator when coupling to the measurement waveguide is excluded. Relevant losses include material absorption, sidewall scattering, radiation at bends, and other internal dissipative pathways. Coupling Q, written Qc, describes the deliberate energy leakage from the cavity into an external waveguide or port. Loaded Q, written Ql, describes the total decay observed when those loss channels act together.
For a single mode with independent linear decay channels, their inverse quality factors add: 1/Ql = 1/Qi + 1/Qc. This makes Ql smaller than either Qi or Qc. Equivalently, total photon decay rate equals intrinsic decay rate plus coupling decay rate. The measured linewidth is governed by that total rate, not by internal propagation loss alone.
Under-coupled means intrinsic loss dominates: Qi is smaller than Qc. Over-coupled means coupling leakage dominates: Qc is smaller than Qi. At critical coupling, Qi equals Qc, so Ql equals Qi/2. In an ideal single-bus traveling-wave ring with no parasitic interference, critical coupling produces a zero through-port transmission at the resonant wavelength. Real devices often show a nonzero minimum for additional reasons.
Prepare a Trustworthy Transmission Measurement
A typical experiment uses a wavelength-tunable laser, polarization control, stable input coupling, a through-port photodetector, and a calibrated wavelength or frequency reference. Optional drop-port data and phase measurements can constrain the coupling model. Establish a stable optical alignment before estimating any linewidth, and identify the polarization and spatial mode being measured.
- Begin at modest optical power so heating, nonlinear absorption, and intensity-dependent resonance shifts do not distort the line.
- Record an off-resonance baseline across a wavelength interval wide enough to distinguish the resonance from slow spectral drift.
- Acquire several fine scans across the same dip, in both sweep directions when possible. Use enough samples to resolve the linewidth rather than relying on a few points.
- Check the laser wavelength axis against a calibrated reference. Account for finite laser linewidth, detector bandwidth, sampling interval, and sweep-speed limitations.
- Normalize the through-port power to the local baseline and record input power, temperature, polarization, scan direction, and any filtering.
If the resonance shifts between forward and reverse scans, or its shape depends on optical power, a static Lorentzian fit may not be valid. Lower power, slow the scan, stabilize temperature, and document any residual drift before interpreting Q.
Extract Loaded Q from the Measured Linewidth
First find the resonance center and fit its normalized lineshape. For a clean, isolated dip, determine the FWHM relative to the local off-resonance baseline and the dip minimum, not simply the absolute transmission level of 50 percent. Convert the fitted wavelength width to frequency when a frequency-domain model is required. Fitting is usually more robust than locating crossings by eye.
As an illustration, a resonance centered at 1550 nm with a fitted width of 0.020 nm has Ql approximately equal to 1550 / 0.020 = 77,500. This is a loaded value. It cannot legitimately be reported as the intrinsic Q unless coupling loss has been independently shown negligible.
Do not fit a single Lorentzian across a visibly split doublet, asymmetric interference feature, or thermally skewed trace. Backscattering can couple clockwise and counterclockwise modes, producing two nearby resonances. Such data need a coupled-mode or other justified model, with uncertainty in every fitted parameter.
Estimate Intrinsic Q and Coupling Q
For an ideal single-bus, single-mode ring with a normalized through-port power minimum T0, temporal coupled-mode theory gives T0 = ((Qi – Qc)/(Qi + Qc)) squared. Define r as the square root of T0. Combining this relation with 1/Ql = 1/Qi + 1/Qc yields two mathematically possible solutions. The dip depth alone does not identify which coupling regime applies.
- Under-coupled solution: Qi = 2 Ql / (1 + r), and Qc = 2 Ql / (1 – r).
- Over-coupled solution: Qi = 2 Ql / (1 – r), and Qc = 2 Ql / (1 + r).
- At critical coupling, T0 approaches zero, r approaches zero, and both Qi and Qc approach 2 Ql.
For example, if Ql is 80,000 and T0 is 0.36, then r is 0.6. The under-coupled interpretation gives Qi = 100,000 and Qc = 400,000. The over-coupled interpretation gives Qi = 400,000 and Qc = 100,000. Both predict the same measured linewidth and transmission minimum. Claiming one intrinsic value without additional evidence would conceal a fundamental ambiguity.
Resolve the regime using independent information, such as a controlled series of bus-to-ring coupling gaps, calibrated phase response, a second measurement port, or a validated device model. These formulas also assume that the measured baseline and extinction accurately describe the cavity rather than polarization leakage, reflections, or another optical pathway.
Check the Model Before Converting Q to Loss
Resonators with multiple bus waveguides have more than one external coupling channel. Their inverse Q contributions must all be included in the loaded decay rate. A single-bus formula applied to a through-and-drop structure can produce a misleading Qi. Likewise, mode splitting, coherent interference, and strong nonlinear behavior can invalidate the simple dip-depth relationship.
Intrinsic Q can sometimes be related to propagation loss using the modal group index, resonance wavelength, and round-trip attenuation model. That conversion requires a defensible group index and a clear definition of whether loss is expressed as power attenuation per unit length. Q alone does not uniquely separate scattering from absorption, and a measured loaded Q cannot be converted directly into intrinsic waveguide loss.
Common Errors and Practical Diagnostics
- Coarse wavelength sampling: Too few data points across the dip bias linewidth estimates; refine the scan and test fit convergence.
- Laser frequency uncertainty: Uncalibrated sweep nonlinearity and source linewidth can broaden or distort narrow features; verify the wavelength axis.
- Thermal pulling: Heating makes scans asymmetric or direction-dependent; reduce power and compare forward and reverse traces.
- Baseline errors: Fiber fringes, coupling drift, and detector offsets change apparent extinction; normalize locally and inspect background structure.
- Polarization mixing: Multiple resonances may overlap when polarization is unstable; control and document the launched mode.
- Unresolved mode splitting: A doublet fit as one dip produces a spurious linewidth; inspect residuals and use an appropriate mode-coupling model.
- Unrecognized loading: Additional ports or scattering pathways change the relation between Ql and Qi; enumerate loss channels before fitting.
Report Reproducible Quality-Factor Results
Report device geometry, resonance wavelength, polarization, optical power, temperature, sweep direction, fit model, linewidth, Ql, and measurement uncertainty. State whether the trace is normalized through-port power or another measured quantity. If Qi and Qc are inferred, disclose the assumed coupling regime and the independent observation supporting it.
Include raw or minimally processed traces, fit residuals, repeat measurements, and the treatment of distorted or split resonances. When data establish only Ql, avoid unsupported intrinsic-Q estimates.
Frequently Asked Questions
Can loaded Q ever equal intrinsic Q?
It can approach intrinsic Q when coupling to the external waveguide is extremely weak, meaning Qc is much larger than Qi. Exact equality is the limiting case of zero external coupling, which also makes optical interrogation difficult. Finite coupling generally lowers Ql.
Does a deeper resonance dip always mean a higher Q?
No. Dip depth depends on the balance between internal and external decay, while linewidth depends on their sum. A deep dip can occur at critical coupling even when Ql is substantially lower than Qi. Depth and width must be interpreted together.
Why are there two possible intrinsic Q values from one dip?
Through-port intensity measures the square of a field amplitude, which loses the sign distinguishing under-coupling from over-coupling. Both regimes can produce the same normalized minimum. Phase, additional ports, or a controlled coupling series can break that ambiguity.
Should linewidth be measured in wavelength or frequency?
Either is acceptable for a narrow resonance when conversion is handled consistently. Frequency-domain decay models are more fundamental; the approximation Q = lambda0 / delta lambda is accurate when the width is small compared with the center wavelength. Never substitute free spectral range for linewidth.
What if the resonance appears asymmetric or split?
Do not force a conventional Lorentzian through it. Test for backscattering, thermal effects, nearby modes, and interference, then select a physical model that explains the trace. Report the model limitation if the available data cannot distinguish these mechanisms.
Safety and Installation Boundaries
Optical laboratories can involve invisible laser radiation, energized instruments, hot components, and fragile fiber ends. Follow the site’s laser safety procedures, use appropriate beam containment and protective measures, and never inspect an active optical fiber directly. Optical alignment, enclosure changes, and instrument installation should be supervised by qualified professionals; electrical supply modifications must be performed by licensed professionals where required. These measurement principles are educational guidance, not a substitute for a site-specific safety assessment or approved installation procedure.
For company information and updates, visit the PhotoBattery Inc. homepage.
