Skip to content

Skip to content

You said:

Is there any material you can make in the cleanroom and nanopattern that is a piezo electric?

ChatGPT said:

Yes — there are several cleanroom-compatible, nanopatternable piezoelectric material stacks people routinely use. The right choice depends on whether you need strong piezo response, CMOS/process compatibility, thickness, and whether you can tolerate high-temperature steps.

Best-practice options for nanopatterned piezo in a cleanroom

1) Aluminum nitride (AlN) — the workhorse for MEMS/NEMS

  • Why it’s popular: sputter-deposited, robust, low loss, compatible with standard microfabrication, widely used for RF MEMS resonators and acousto-optics.

  • Patterning: typically pattern the metal electrodes (Ti/Pt, Mo, Al, etc.) by liftoff/etch; AlN itself can be patterned by dry etch (Cl-based ICP/RIE) if needed.

  • Notes: piezo strength is moderate (less than PZT), but it’s “easy” and very stable.

2) Scandium-doped AlN (ScAlN) — “AlN but much stronger”

  • Why: substantially enhanced piezo coefficients vs pure AlN.

  • Patterning/process: similar to AlN (sputter + dry etch), but film stress and composition control can be more finicky.

3) Zinc oxide (ZnO) — easy deposition, decent piezo, more fragile

  • Why: sputterable and piezoelectric.

  • Patterning: doable with liftoff (for electrodes) and etching routes exist, but ZnO can be chemically and electrically more temperamental than AlN, depending on environment.

4) Lead zirconate titanate (PZT) — the strongest common thin-film piezo

  • Why: very high piezo response; great if you need big actuation at low voltage.

  • Reality check in cleanrooms: often requires sol-gel or sputter plus high-temperature anneals (and lead handling rules). Many university cleanrooms restrict it.

  • Patterning: lift-off is hard for thick/annealed films; people often use hard masks + etch or pattern electrodes and define devices by etch/ion milling.

5) Lithium niobate (LiNbO₃, “LN”) thin films (LNOI) — excellent piezo + photonics

  • Why: strong piezoelectricity + fantastic optical properties (electro-optic too). For photonics labs, this is a huge one.

  • How it’s used: usually as commercial thin-film wafers (LNOI). You nanopattern LN waveguides/resonators with etching, and add electrodes to drive SAWs/BAWs, etc.

  • Cleanroom angle: you’re not “depositing LN” typically; you’re etching/patterning it.

6) Polymer piezoelectrics (PVDF, P(VDF-TrFE)) — low-temp, flexible, easy-ish

  • Why: spin-coatable, low-temperature, can be patterned (often by liftoff/etch), compatible with fragile substrates.

  • Catch: you usually need poling (strong electric field) to align domains; performance and stability can vary.

A practical “shortlist” depending on what you’re trying to do

  • Most universal / least drama: AlN (or ScAlN if you need more coupling).

  • Max actuation: PZT (if your facility allows it).

  • If you’re doing nanophotonics + piezo drive: LNOI (etch LN + electrodes).

  • Low-temp, fast prototyping: P(VDF-TrFE).

Quick sanity questions (answer in one line each and I’ll narrow it to 1–2 best stacks + pattern flow)

  1. Do you need actuation, sensing, or acoustic wave generation (SAW/BAW)?

  2. What substrate: Si, SiO₂, diamond, SiN, something else?

  3. Can you do >400–800 °C steps, or do you need low-temp only?

  4. Are lead-containing processes (PZT) allowed in your cleanroom?

If you tell me those constraints, I’ll suggest a concrete process stack (material + electrodes + etch/liftoff approach) that’s realistic in a university cleanroom.

You said:

Could I use microlithography to make a piezo and metal etc. that can tune 100 pm? How tall would it have to be

ChatGPT said:

Thought for 16s

Yes — microlithography + thin-film piezo + metal electrodes can absolutely give you 100 pm (0.1 nm) of tuning. The “how tall does it have to be?” part depends a lot on actuator geometry. There are two common regimes:

1) Out-of-plane (“stack”) actuation: film expands in thickness direction

If you have a piezo film of thickness tt sandwiched between two electrodes and it’s allowed to move out of plane, the ideal free displacement is

Δz≈d33 VΔz≈d33​V

Key surprise: in this idealized case, ΔzΔz does not depend on thickness tt. Thickness mostly determines:

  • how much voltage you can safely apply before dielectric breakdown (since E=V/tE=V/t)

  • how hard it is to fabricate / stress / residual curvature

What voltage gives 100 pm?

Use Δz=100 pmΔz=100 pm:

  • PZT (thin film/bulk-ish): d33∼200−500 pm/Vd33​∼200−500 pm/V
    ⇒V∼0.2−0.5 V⇒V∼0.2−0.5 V (in the free ideal case)

  • AlN: d33∼3−6 pm/Vd33​∼3−6 pm/V
    ⇒V∼17−33 V⇒V∼17−33 V

  • ScAlN: d33d33​ can be significantly higher than AlN (often ~2–5× depending on composition/film quality)
    ⇒⇒ maybe single-digit to ~10 V ballpark for 100 pm

Reality check (important)

On a wafer, thin films are often clamped by the substrate, so the motion you get at the surface can be much smaller than d33Vd33​V unless you:

  • pattern a released membrane/cantilever, or

  • use a flexural unimorph/bimorph design (see below)

So if your device is not released, you usually don’t design for pure “stack” motion.

2) In-plane strain (most common in MEMS): piezo strain causes bending or length change

If the piezo strain is in-plane (using d31d31​), the in-plane strain is

S≈d31 VtS≈d31​tV​

and a feature of length LL changes by

ΔL≈d31 Vt LΔL≈d31​tV​L

Here thickness does matter (thinner piezo = higher field for same V = more strain).

Quick example for scale

Take PZT with ∣d31∣∼100 pm/V∣d31​∣∼100 pm/V, L=100 μmL=100 μm, t=1 μmt=1 μm:

ΔL≈100×10−12⋅V10−6⋅10−4=10−9V m=1 nm per voltΔL≈100×10−12⋅10−6V​⋅10−4=10−9V m=1 nm per volt

So 100 pm is about 0.1 V worth of drive in this simplified picture.

Bending (unimorph) amplifies motion

If you put piezo on a passive layer (metal/SiN/SiO22​) and release it as a cantilever/bridge, the in-plane strain turns into much larger vertical tip deflection. That’s usually how people get easy picometer → nanometer motion with modest voltages.

So… “how tall would it have to be?”

If you truly mean a vertical stack under a moving element: height isn’t what sets 100 pm — d33d33​ and V do. But practically:

  • For AlN, people use t∼0.3−2 μmt∼0.3−2 μm films so that 10–50 V corresponds to reasonable fields.

  • For PZT, t∼0.2−2 μmt∼0.2−2 μm is common, but process constraints dominate (anneal, stress, facility rules).

If you mean “how tall must a released actuator beam be,” then thickness matters via stiffness and bending; typical released MEMS beams are a few hundred nm to a few µm thick total, depending on force/displacement targets.

One question that decides the right formula

Are you trying to tune something by:

  1. moving a surface up/down (out-of-plane gap tuning), or

  2. stretching/compressing a length (in-plane tuning), or

  3. bending a beam/membrane (most common for large motion)?

If you answer with your geometry (even a sketch in words: “cantilever”, “membrane over gap”, “ring resonator gap”, etc.) and what material you can deposit (AlN vs PZT), I’ll give you a realistic thickness + voltage estimate for getting 100 pm at the point you care about, including clamping/release effects.

You said:

I'm really interested in making a fabry-perot cavity and coupling to my quantum emitter. It's traditionally very difficult, but there are recent demonstrations where you can do it on a chip or in a monolithic platform. You can make the cavity with a quantum dot in the center and just stampt the concave mirror on top with no degrees of freedom. The concave mirror can be made easily by reflowing resist or by laser ablation. I want you to perform a literature review of all the platforms people have made of micro cavities in this geometry (Requirements include two mirrors, at least one of them concave, and at most one degree of freedom to change the cavity length). I am not interested in planar cavities like ring resonators and nanobeam cavities. I am not interested in these complicated experiments with control over all degrees of freedom for the mirrors.

ChatGPT said:

Thanks for the detailed background! To help narrow the search and focus on the most relevant work, could you please clarify the following:

  1. What types of quantum emitters are you most interested in? (e.g., quantum dots, color centers in diamond, atoms, etc.)

  2. Do you want recent literature only (e.g., past 5 years), or should I include foundational work as well?

  3. Should the review include both theoretical and experimental work, or only experimental demonstrations?

With that info, I’ll build a focused and comprehensive review of relevant microcavity platforms.

You said:
  1. Any type of quantum emitter is fine, or even just the cavity with no emitter. 2. Both recent and foundational 3. Both. Remember, I don't care about in-plane cavities. I want free-space fabry perot with no degrees of freedom. Okay?
ChatGPT said:

Got it — I’ll compile a literature review of Fabry–Perot-type microcavities with two mirrors (at least one concave), where at most one degree of freedom is used to control cavity length. I’ll exclude in-plane resonators like ring or nanobeam cavities and omit platforms with full multi-axis control. I’ll include both foundational and recent experimental and theoretical work, even without embedded emitters.

I’ll let you know as soon as I have a thorough review ready. Feel free to continue chatting or add details in the meantime!

Research completed in 13m · 21 sources · 109 searches

You said:

Of those piezoelectric materials we talked about, are any of them highly optically transparent?

ChatGPT said:

Great follow-up. I’ll look into the optical transparency of commonly used piezoelectric materials — like AlN, ScAlN, PZT, ZnO, and PVDF — across visible and near-infrared wavelengths. I’ll report how transparent they are, their refractive indices, and whether they're suitable for optical applications like cavity integration.

I’ll let you know as soon as I have a detailed comparison ready. Feel free to keep asking questions in the meantime!

Research completed in 10m · 25 sources · 72 searches

ChatGPT said:

Literature Review of Minimal-Degree Fabry–Perot Microcavities

Fabry–Perot Microcavity Platforms with Concave Mirrors (Limited DOF Tuning)

Free-space Fabry–Perot microcavities consist of two facing mirrors forming a small optical resonator. To achieve a tightly confined mode (high Purcell factor) while maintaining stability and high finesse, at least one mirror is often concave. Below, we review such microcavity platforms that incorporate concave mirrors, with at most one active degree of freedom for tuning the cavity length. We exclude fully adjustable (multi-actuator) open cavities and planar in-plane resonators, focusing instead on monolithic or minimally tunable designs.

Fiber-Based Concave Microcavities

One widely used platform is the fiber-based Fabry–Perot microcavity, where a concave mirror is fabricated on the tip of an optical fiber. The concave fiber tip is typically made by laser ablation (e.g. CO₂ laser pulses melting the glass) or focused-ion-beam milling, producing a microscopic spherical depression. The fiber’s concave mirror is automatically aligned to the fiber core, ensuring efficient coupling of the cavity mode into the fiber (coupling efficiency >90% has been achieved). The other mirror is usually a macroscopic planar mirror or another concave fiber, forming a small air-gap cavity. After coating with high-reflectivity dielectric layers, such fiber-tip cavities can reach finesse $F$ on the order of $104$–$105$ and sub-picoliter mode volumes. For example, Hunger et al. (2010) reported a fiber cavity with finesse ∼37,000 at 780 nm and mode waist ~2.8 µm, while Muller et al. (2010) achieved an ultrahigh-finesse ~300,000 in a 6 µm length fiber microcavity. Fiber microcavities are tunable in length by moving one fiber along its axis (often via a single piezo actuator) – providing spectral tunability without requiring multi-axis alignment once initially set. Notably, Greuter et al. (2014) used CO₂-laser ablation to create a 10 µm radius concave fiber mirror, yielding a cubic-wavelength-scale mode volume (waist ~1 µm) and demonstrating sub-wavelength beam waists. Such small mode volumes, together with fiber integration, make these cavities attractive for quantum optics.

Integration with quantum emitters: Fiber-based cavities have been successfully coupled to solid-state quantum emitters. For instance, Albrecht et al. (2013) coupled a single NV center in diamond nanocrystal to a tunable fiber microcavity, observing enhanced emission. In that system, the fiber mirror was concave (CO₂-machined) and the cavity length ~3–4 µm, yielding a Purcell factor on the order of 10. Similarly, fiber cavities have been used to interface with semiconductor quantum dots (in membranes) and cold atoms. These demonstrations typically rely on an initial multi-axis alignment of the fiber to the emitter chip, but once aligned the cavity length is the only actively tuned parameter (e.g. by a piezo) during operation. Recent efforts aim to package fiber microcavities in a self-aligned manner so that only one degree of freedom (axial motion) is needed. For example, Qing et al. (2019) etched and laser-reflowed a concave fiber tip of 4.7 µm diameter (matched to the fiber core) that, after coating, formed a microcavity with $Q>1000$ and mode volume $V\approx13.6~\mu$m³. This device was wavelength-tunable and even demonstrated microlaser emission, all while maintaining high fiber-coupling efficiency due to the self-aligned mirror.

Microfabricated Concave-Mirror Cavities (Chip-Based)

Another class of platforms uses microfabricated concave mirrors on substrates (chips) to form stable microcavities. These typically involve fabricating an array of concave mirror templates on a flat substrate via lithographic or direct-write techniques, then combining that substrate with a second mirror (often a flat mirror or another concave mirror) to create many microcavities. A key goal is to achieve small mirror radii and smooth surfaces for high finesse, using monolithic fabrication so that alignment is fixed by design (with at most one tuning axis, usually the gap). Several fabrication methods have been developed:

  • Focused Ion Beam (FIB) milling: Dolan et al. (2010) pioneered FIB to mill concave pits (~5–20 µm radius) in a silica substrate, which were then dielectric-coated to serve as mirrors. They reported arrays of femtoliter-volume, open-access microcavities with mode volumes as low as 2.2 µm³ and finesse up to ~460. FIB allows flexible mirror geometries, and later work improved the finesse into the $104$–$105$ range. However, FIB is serial and can introduce surface roughness (limiting finesse). It remains useful for prototyping; for example, a FIB-fabricated concave mirror by Patel et al. achieved finesse $\sim104$–$105$, and arrays of FIB mirrors have been used for tunable cavity-QED experiments.

  • Laser micromachining: CO₂ laser ablation of glass can produce ultra-smooth concave profiles. This was used by groups like Warburton’s: Greuter et al. (2014) produced atomically smooth concave dimples (radius ~10–20 µm) in fused silica by CO₂ laser pulses. Pairing such a miniaturized curved mirror with a planar DBR mirror, they realized a high-finesse ($F\sim10^4$) microcavity with mode volume on the order of $(\lambda)^3$. The cavity was spatially and spectrally tunable, yet mechanically rigid enough to operate at cryogenic temperatures with sub-picometer stability. Laser-ablated concave mirrors on optical fibers or chips have achieved finesse up to $10^5$ in later works. Takahashi et al. (2014) and others employed ultrafast lasers to fabricate micron-scale concave features with good mode matching.

  • Isotropic etching and wafer fabrication: An early foundational work by Trupke et al. (2005) demonstrated a microfabricated high-finesse cavity using isotropic wet etching of silicon. In this approach, a concave pit is etched (e.g. with an isotropic HF or XeF₂ etch) and then coated with a DBR. Trupke’s device had mirror radius $R\sim150~\mu$m, cavity length ~$130~\mu$m, finesse on the order of a few thousand, and was intended for coupling to cold atoms. Subsequent improvements by the same team and others greatly enhanced the performance. Biedermann et al. (2010) achieved ultra-smooth concave silicon mirrors (surface roughness $\sim$2.2 Å) via a combination of plasma etching and thermal oxidation smoothing. Their cavities reached finesse $F\approx64{,}000$ at 780 nm. More recently, Wachter et al. (2019) reported silicon concave-mirror microcavity arrays with finesse up to $5×10^5$ (measured via linewidth and ring-down). They fabricated 100 concave mirrors (radius $R\sim100$–300 µm) by a CMOS-compatible two-step plasma etch plus oxidation smoothing, achieving mirror losses below 3 ppm. By bonding two such mirror chips with a precisely defined spacer, they formed rigid open-access cavities without any post-assembly alignment. Impressively, the best plano-concave cavities had $F\approx5.3×10^5$ (mode volume $V\sim330~\text{fL}\approx88,\lambda^3$) and $Q\sim1.1×10^7$. Even concave–concave symmetric cavities were demonstrated using lithographically aligned mirror pairs: alignment holes and micro-pillar spacers in the chips ensured <5 µm lateral accuracy. This yielded stable cavities with $F$ up to $(4.0\pm0.1)×10^5$ at $L\approx150~\mu$m ($R\approx200~\mu$m), corresponding to $Q\approx8×10^7$. Crucially, these wafer-bonded microcavities require no manual tuning of alignment – once the chips are bonded, only the cavity length (set by spacer thickness) determines resonance. Small length adjustments (one DOF) can be introduced by piezo actuation of one chip if needed, but the spatial overlap is fixed by the bonding process.

  • Photoresist reflow & etching: A very recent advance is the use of reflowed polymer microlenses as mirror templates. In 2022, Westly et al. (NIST) reported fabrication of concave mirrors via photoresist thermal reflow and transfer etching, achieving unprecedented surface quality. In this method, photoresist droplets are melted to form near-perfect spherical caps, then etched into a substrate and coated. The result was an array of microcavities with measured finesse up to 1.3 million and scatter losses <1 ppm. Radii of curvature could be user-defined from ~50 µm up to many mm in the same wafer. The average coating-limited $F\approx1.05×10^6$ across 43 devices (at 1560 nm), demonstrating scalable, ultrahigh-finesse mirror fabrication. This approach marries semiconductor processing with traditional mirror quality, opening the door to integrated microcavities for quantum and precision metrology on a chip.

Tuning and emitter integration: Chip-based concave microcavities are often assembled in a fixed geometry (or with a single tunable axis). For example, Dolan’s 2010 FIB-mirror array was mounted with a macroscopic mirror on a translation stage to tune all cavities’ length uniformly. In sensing applications, some bonded-cavity systems incorporate a PZT or thermal element to slightly adjust the gap for calibration – but typically only one actuator is used (no multi-axis controllers) because the mirror alignment is lithographically pre-set. These microcavities have been used to enhance and study emitters as well. Najer et al. (2019) embedded a gated GaAs quantum dot in a microfabricated open cavity with $Q\sim10^6$, achieving the strong-coupling regime of cavity QED. (This used a fiber-aligned open cavity, which does allow multi-DOF alignment, but it underscores the potential of high-$Q/V$ cavities for emitters.) On the chip-integrated side, Fait et al. (2021) demonstrated high-finesse silicon microcavities operating at 1280 nm (telecom O-band) with Erbium dopants, aiming for Purcell-enhanced emission for quantum memory applications. The general trend is that as microfabricated cavity finesse and quality have improved, researchers are incorporating emitters (quantum dots, rare-earth ions, etc.) either by bonding material into the cavity or by positioning the cavity around the emitter – all while maintaining minimal degrees of freedom in operation.

Monolithic “Buckled-Dome” Microcavities

A distinct approach to concave microcavities is the monolithic air-gap Fabry–Perot formed by a buckled mirror. In these devices, both mirrors are fabricated in place on a single substrate stack, and a controlled stress release causes one mirror to deform into a concave “dome” shape. The group of DeCorby and Davis introduced this concept: in 2011, Allen et al. used a multilayer $\text{a-Si/SiO}_2$ film under residual stress, which upon release buckled into a curved mirror (like a tiny dome). Two such curved DBR mirrors facing each other can form a microcavity. More commonly, a half-symmetric design is used: a planar DBR mirror on a substrate and a second DBR that buckles outward to create a concave top mirror. The cavity is an “open-access” air gap, but it is monolithically self-aligned – the dome remains concentric with the planar mirror below it, by virtue of the fabrication process.

Fabrication and performance: Potts et al. (2016) reported on-chip dome microcavities fabricated by this guided self-assembly method. They deposited a dielectric DBR (e.g. 10–15 alternate layers of $\text{SiO}2$ and $\text{Ta}_2\textO_5$) on a substrate, then a thin sacrificial release layer, and finally another DBR. Heating the stack caused the top DBR to delaminate in circular regions (defined by photolithography) and _buckle upward due to compressive stress. The result is a concave mirror of diameter 100–300 µm and sag height ~2.5–10 µm (controllable by diameter). Warburton’s group characterized similar buckled cavities, noting interference fringe patterns confirming high symmetry (cylindrical curvature). A typical 100 µm-diameter dome cavity had an optical length $L\approx3.3~µ$m and effective radius $R_c\approx210~µ$m. With high-reflectivity DBRs (designed for 780 nm), these devices achieved finesse $F\approx3500$ (consistent with the 99.91% mirror reflectivity). The measured fundamental mode linewidth was ~27 pm (at $\sim778$ nm), matching $F\approx3560$. The mode waist was $w_0\approx2.5~µ$m, giving a mode volume $V \approx 35λ^3$ (for $\lambda\sim780$ nm). These values indicate a reasonably small mode volume and moderate finesse – suitable for cavity QED in the weak to intermediate coupling regime.

Because the mirrors are fixed on-chip, tuning is limited to one degree (global cavity length changes, e.g. via temperature or strain). Bitarafan et al. (2015) showed that thermal expansion can tune the resonance of buckled-dome cavities over several free spectral ranges. In one example, heating from 20°C to 100°C red-shifted the cavity mode by >0.5 nm, enough to align to atomic transitions. The thermal stability is high; Potts et al. noted that resonance could be stabilized to within $\pm 5$ pm in a cryostat. In principle, one could integrate microheaters or MEMS actuators to adjust the dome height (one axis); however, no multi-axis alignment is needed since the mirrors are permanently concentric. Recent studies even observed optical bistability in these air-gap cavities due to intracavity heating, a sign that the cavity-emitter or cavity-optomechanical coupling is significant.

Integration with emitters: The buckled-cavity platform is appealing for interfacing with atomic systems, since it can be made to resonate at atomic lines (e.g. Rb D₂ at 780 nm) and potentially integrated with vapor cells or cold atom setups. Indeed, the 2016 devices were designed to be tunable to the Rb D₂ line. In a solid-state context, one could imagine depositing a thin emitter layer before mirror deposition so that emitters reside inside the cavity. To date, most demonstrations have been passive; however, Merkel et al. (2020) inserted erbium ions (at 1530 nm) into a cryogenic high-Q microcavity, achieving Purcell-enhanced coherent emission. That system utilized a macroscopic reference cavity, but it foreshadows using on-chip DBR microcavities for emitter coupling. The buckled-dome approach provides an integrated, alignment-free microcavity with one tunable parameter (e.g. temperature). The trade-off is somewhat larger mode volume compared to fiber or etched microcavities, but ongoing work aims to reduce the diameter and increase the curvature of buckled mirrors for tighter confinement.

Summary and Outlook

In summary, a variety of Fabry–Perot microcavity platforms with concave mirrors have been developed, balancing high optical performance with limited degrees of freedom for tuning:

  • Fiber-tip microcavities (concave fiber + mirror) achieve small volumes and high finesse with built-in fiber coupling, requiring only one axis of motion (length) during use. They have enabled strong light-matter interaction with NV centers and quantum dots, and even form the basis of microcavity lasers.

  • Microfabricated mirror cavities on chips (concave pits made by FIB, laser, or etching) offer scalable arrays and extremely high finesse. Foundational work showed finesses ~$103$–$104$, and recent monolithic assemblies now reach $F>105$–$106$ with sub-λ³ volumes. These systems are often bonded or fixed such that only the cavity length can be tuned (e.g. via a piezo or strain), while mirror alignment is lithographically defined. This stability makes them promising for quantum networks and precision sensors.

  • Buckled-dome (self-assembled) cavities integrate the entire FP cavity on one chip, with a curved DBR mirror emerging through stress engineering. They require no post-fabrication alignment and can be thermally or mechanically tuned in length. Achieved finesses ($\sim10^3$) are lower than the best open cavities, but adequate for many strong-coupling experiments, and improvements in mirror coatings could push this higher.

Each platform involves different mirror fabrication techniques – from laser-reflowed glass (giving atomically smooth concave surfaces), to FIB milled pits (flexible patterning), to thermal shaping of polymers, to strained film buckling. Crucially, all use at least one concave mirror to support a stable, near-Gaussian mode despite the microscopic cavity length, which mitigates diffraction loss and allows finesses comparable to macroscopic supermirrors. By restricting the design to at most one active tuning axis, these platforms forego the full adjustability of macroscopic cavities in exchange for robustness and scalability (no constant re-alignment). This has enabled deploying microcavities in challenging settings: cryogenic setups, vibration-prone environments, or large arrays for sensing.

Looking forward, such concave-mirror microcavities are poised to play a key role in quantum photonics and sensing. Ongoing research is integrating solid-state emitters (quantum dots, color centers, 2D materials) into these tiny cavities to realize efficient single-photon sources and strong emitter–cavity coupling on chip. The combination of high $Q$ and low $V$ (for large Purcell factors) with minimal alignment overhead makes these platforms attractive for scalable quantum devices. Advances like the photoresist-reflow mirrors reaching $F>10^6$ suggest that performance no longer needs to be sacrificed for compactness. In short, the field has progressed from early tunable microcavities with modest finesse to today’s integrated microcavities rivaling macroscopic resonators in $Q$ – all while maintaining one simple degree of freedom for tuning. This marriage of precision optics and microfabrication will likely underpin next-generation quantum interfaces and sensors built around Fabry–Perot microcavity platforms.

Sources:

  • Hunger et al., NJP 12, 065038 (2010) – High-finesse fiber Fabry–Perot design

  • Muller et al., Opt. Lett. 35, 2293 (2010) – Ultrahigh-$F$ low-$V$ microcavity

  • Dolan et al., Opt. Lett. 35, 3556 (2010) – FIB-fabricated femtoliter cavity arrays

  • Greuter et al., APL 105, 121105 (2014) – Small-mode-volume CO₂-laser microcavity

  • Bitarafan et al., JOSA B 32, 1214 (2015) – Buckled dome microcavities (thermo-optic tuning)

  • Potts et al., APL 108, 041103 (2016) – On-chip open-access dome cavity for cQED

  • Wachter et al., Nat. Commun. 10, 5020 (2019) – Si microcavity array, $F>5×10^5$

  • Najer et al., Nature 575, 622 (2019) – Strongly coupled QD–microcavity ($Q∼10^6$)

  • Westly et al., arXiv:2203.15931 (2022) – Reflow-etched mirrors with $F>10^6$

  • Albrecht et al., PRL 110, 243602 (2013) – NV center in fiber microcavity

  • (Additional citations provided in-line.)

ar5iv.labs.arxiv.org&sz=32

Sources

ChatGPT said:

Optical Transparency of Common Piezoelectric Thin Films

Optical Transparency of Common Piezoelectric Materials (400–1600 nm)

Material Transparency (400–1600 nm) Refractive Index Absorption / Optical Loss Transparent-Device Suitability Piezo vs. Optical Trade-offs
Aluminum Nitride (AlN) Highly transparent across visible & NIR (wide bandgap ≈6.1 eV, absorption edge ~200–210 nm). Thin AlN films show ~75–85% transmittance for λ > 400 nm (limited mainly by surface reflections). ~2.0 at visible wavelengths (typical n ≈1.98–2.15 for polycrystalline AlN at 632.8 nm; up to ~2.2 for epitaxial films). Mild normal dispersion (higher n in UV, ~2.2 at 400 nm, dropping to ~1.9–2.0 at 1500 nm). Negligible absorption in 400–1600 nm. AlN is an insulator with essentially zero extinction coefficient in the visible/NIR (e.g. k ≈0.001 by 1200 nm). Optical losses are extremely low (AlN waveguides can achieve sub-dB/cm loss if high-quality). Yes. AlN’s wide transparency window and high index make it ideal for piezo-optical devices (e.g. optomechanical resonators). It has been used in UV-optoelectronics and integrated photonics. Its piezoelectric response (though moderate) can tune optical cavities without introducing absorption. AlN offers superb optical clarity but relatively modest piezoelectric coefficients. No special trade-off is needed to maintain transparency – it is intrinsically transparent. However, attempts to boost piezoelectricity (e.g. alloying with Sc) can reduce the bandgap and introduce slight optical loss or scattering (see ScAlN). In short, AlN favors optical purity over piezo strength.

| Scandium-doped AlN (ScAlN) | Remains highly transparent in vis–NIR for moderate Sc content. Alloying Sc into AlN (wurtzite structure stable up to ~50% Sc) red-shifts the bandgap but it stays in the UV range for ≤30–40% Sc. Thus, ScₓAl₁₋ₓN with typical x~0.2–0.3 absorbs only at UV wavelengths (visible and 1.55 µm light passes through). | Slightly higher index than pure AlN (Sc lowers the bandgap, increasing polarizability). For example, Sc₀.2Al₀.8N films report n ~2.1–2.2 at 633 nm (vs ~2.0 for AlN). Generally n remains in the ~2.0–2.3 range in the visible, depending on Sc fraction. | Low intrinsic loss in NIR – suitable for waveguides. Sc₀.3Al₀.7N racetrack resonators have shown propagation loss ~1.6 dB/cm at 1550 nm (dominated by material absorption). High-quality single-crystal ScAlN can achieve optical losses on par with undoped AlN, but sputtered polycrystalline films may suffer higher scattering loss. | Yes. ScAlN is emerging as an excellent platform for transparent photonic devices. Sc doping boosts AlN’s piezoelectric and Pockels coefficients dramatically while retaining a broad transparency window. It is being explored for high-speed modulators and acousto-optic devices where its enhanced electromechanical coupling is useful. | The piezoelectric enhancement from Sc comes at some cost to optical clarity if not carefully managed. Higher Sc content or poor crystallinity can introduce optical absorption (due to the lower bandgap and defects) and increased scattering. There is a trade-off between maximizing d₃₃/nonlinearity (at high Sc%) and maintaining low optical loss. Optimizing film quality (epitaxy, low-defect growth) is crucial to benefit from high piezo performance without spoiling transparency. |

| Lead Zirconate Titanate (PZT)
(Pb(Zr,Ti)O₃) | Opaque as bulk ceramic, but thin films can be transparent above ~400 nm. PZT is a ferroelectric with bandgap ~3.5–4.0 eV (composition-dependent), so it absorbs UV light <350 nm but not visible photons. Sol–gel PZT films on glass achieve 60–80% transparency in the visible, and PLZT ceramics (La-doped PZT) can exceed 60% transmittance with proper processing. | Relatively high index. PZT’s n is in the 2.0–2.6 range across visible–NIR. For a 53/47 PZT film, n peaks ~2.98 near 320 nm and gradually falls with increasing wavelength (dropping to ~2.2–2.4 in the red/NIR). This high refractive index aids optical confinement. | Very low absorption above the band edge. In the transparent range (>400 nm) PZT behaves as a lossless dielectric (extinction k ≈0). Any residual optical losses stem from scattering (grain boundaries, porosity) or electrode absorption rather than bulk material absorption. For example, PZT films are noted for “low material absorption loss” in the context of integrated optics. With optimized processing (dense, crack-free films), optical loss can be kept small. | Conditionally. While bulk PZT is not optically clear, thin PZT films or transparent PLZT ceramics are used in electro-optic devices. The combination of high electro-optic coefficients and adequate transparency enables PZT-based waveguides and modulators. Its high index allows sub-wavelength light confinement in photonic structures. PZT has been demonstrated in transparent actuators, electro-optic modulators, and “transparent ultrasound” transducers (with >80% optical transmission) that marry acoustic and optical functions. | There is an inherent trade-off between piezoelectric performance and optical clarity in ferroelectric oxides. The domain structures and grain boundaries that maximize PZT’s piezo response tend to scatter light. Thus, making PZT transparent often involves compromises: e.g. doping and microstructural control (La-doping, ultrafine grains) to reduce light scattering, which can slightly reduce ferroelectric strength. In general, pure PZT offers superior piezoelectricity but is opaque, whereas modified or thinned PZT sacrifices some performance to gain transparency. Achieving both high _d_₃₃ and high transparency simultaneously remains challenging. |

| Zinc Oxide (ZnO) | Transparent in the visible and NIR. ZnO is a wide-bandgap semiconductor (E_g ≈3.3 eV) that absorbs only UV (< ~380 nm) and is essentially transparent for 400–1600 nm. Even polycrystalline ZnO films have very low visible/NIR absorption (ZnO is well known as a transparent conductive oxide). | Moderate index (~1.9–2.0 in visible). For ZnO, n ≈2.35 at 300 nm and falls to ~1.85 at 800 nm. In the green (500–550 nm) n ~2.0; in the telecom band (1550 nm) n ~1.7–1.8. The refractive index is lower than AlN or PZT but still sufficient for optical waveguiding on lower-index substrates. | Minimal absorption in its transparent range. Undoped ZnO has virtually no absorption from 0.4–1.6 µm. Optical losses are mainly due to reflection (refractive index contrast) or any free-carrier absorption if the film is doped. (Heavily Al-doped ZnO, used as a TCO, remains ~85–90% transparent in visible but can show increased NIR reflectance/absorption due to Drude carriers.) In general, piezoelectric-grade ZnO (undoped, insulating) has extremely low optical loss in vis–NIR. | Yes. ZnO’s transparency and piezoelectricity allow its use in optical devices – for example, as a transparent piezo layer in acoustic-optical modulators or as the active layer in UV or visible microlasers. It is less common in integrated photonics than AlN, but it is used in transparent electronics (touch panels, displays) where its piezoelectric or acoustic wave properties can be harnessed without blocking light. ZnO can thus serve in tunable optical elements (e.g. transparent SAW devices on optical substrates) as long as electrodes or doping are designed to be optically benign. | ZnO offers a good balance of piezoelectric coupling and optical clarity, but there are trade-offs when tuning its properties. For instance, adding dopants to improve conductivity (for electrode functionality) introduces free carriers that degrade infrared transparency. Likewise, increasing film thickness or crystal size can increase optical scattering. Compared to AlN, ZnO’s piezoelectric coefficient is somewhat lower and it is softer, so achieving the same mechanical tuning might require a thicker layer (slightly reducing transparency). Overall, ZnO’s trade-off is between electrical/piezo optimization (doping, thickness) and maintaining its inherent transparency. |

| Polyvinylidene Fluoride (PVDF) & P(VDF–TrFE) | Transparent, flexible polymer films. PVDF and its trifluoroethylene copolymer are generally clear in the visible and IR, with strong absorption only in the UV (<300 nm). In practice, multilayer PVDF-TrFE thin films show 50–85% optical transmittance in the visible (higher for thinner films). The transparency extends through 1.55 µm (no significant absorption in the telecom band). | Low index (~1.4). As an organic polymer, PVDF-TrFE has a refractive index around 1.5 in the visible. (Measured values ~1.443–1.598 in the 600–650 nm range.) This relatively low n means less optical confinement and higher Fresnel transmission (low reflection losses at interfaces). | Very low intrinsic absorption in 400–1600 nm. PVDF is basically an optically clear plastic in that range – any losses are from reflections or slight haze. The films show high absorption in the UV but can be considered loss-free in visible/NIR. (No significant extinction coefficient in the 0.4–1.6 µm range; e.g. a 10 µm PVDF film can transmit most visible light.) Thus PVDF is well-suited when optical loss must be minimal. | Yes. PVDF-based piezoelectric films are used in flexible transparent devices – for example, as transparent actuators, tunable lenses/mirrors, or wearable sensors that optically couple to displays. Thanks to their optical clarity and nonlinear optical properties, P(VDF-TrFE) films have been investigated for frequency conversion and organic photonics. While their refractive index is low, they can be integrated as optically passive layers or sandwiched in optical cavities without significantly perturbing light. The key advantage is that they add electromechanical functionality with almost no optical absorption and mechanical flexibility. | The main compromise is performance vs. strength: PVDF’s piezoelectric coefficients (d₃₃ ~10–30 pC/N) are much lower than ceramics like PZT, but it achieves transparency effortlessly. Improving the piezoelectric phase (β-phase crystallites) can slightly reduce optical clarity due to increased scattering (more crystalline domains). Similarly, increasing film thickness or stacking layers raises the piezoelectric output but linearly reduces transparency (e.g. transmittance drops from ~85% to ~50% as multilayer film thickness increases). In sum, PVDF excels in optical clarity and flexibility at the cost of piezoelectric strength – it trades brute piezoelectric performance for transparency and mechanical compliance. |

Summary

Overview: AlN and ZnO are wide-bandgap piezoelectric ceramics that are intrinsically transparent from the visible into the infrared. They have moderate refractive indices (∼2.0) and exhibit essentially no absorption in the 400–1600 nm range, making them well-suited for optomechanical devices like tunable optical cavities. Scandium-doped AlN (ScAlN) represents a midpoint – it substantially boosts AlN’s piezoelectric and nonlinear optical coefficients while remaining largely transparent (the band edge shifts into the near-UV as Sc content increases). By 20–30% Sc, ScAlN still transmits visible/IR light with low loss, enabling high-performance transparent actuators and modulators.

Lead Zirconate Titanate (PZT), a classic high-piezoelectric perovskite, is opaque in bulk form but can be made optically transparent in thin-film or ceramic form via microstructural engineering. It has a high refractive index (~2.2–2.5 in VIS) and negligible band-edge absorption above ~400 nm. Transparent PZT-based materials (e.g. PLZT) leverage PZT’s strong electromechanical coupling in optical systems – for example, transparent PLZT ceramics and films have been used as electro-optic modulators and acoustic-optic elements, albeit with some performance trade-offs. The need to minimize light scattering (by reducing ferroelectric domains, porosity, etc.) often slightly compromises PZT’s maximal piezoelectric performance. Nevertheless, advances in processing now yield PZT ceramics with >60–80% transparency and reasonably high d-constants, bridging the gap between optics and ultrasonics.

PVDF and P(VDF-TrFE) polymers offer the opposite balance: exceptional optical clarity and flexibility, but comparatively low piezoelectric response. These fluoropolymer films are transparent across the visible and IR (typical n ~1.5, no absorption except in the UV) and can be easily integrated into optical setups as transparent piezoelectric layers. Their piezoelectric coefficients are an order of magnitude lower than PZT’s, meaning they induce smaller mechanical deformations or electric signals. To enhance PVDF’s piezo activity (through increasing β-phase crystallinity or adding ceramic fillers), one often incurs increased light scattering or opacity. Thus, PVDF is ideal where moderate electromechanical coupling is acceptable in exchange for plastic-like optical transparency and elasticity (for instance, in stretchable or bio-integrated optical devices).

Trade-off Highlights: In summary, achieving both high piezoelectric performance and high optical transparency is challenging – materials like PZT excel in piezoelectricity but require special tuning to be transparent, whereas materials like AlN/PVDF excel in transparency but have lower piezo activity. ScAlN and modern transparent-ceramic techniques represent promising solutions that push the trade-off curve, offering improved piezoelectric coefficients without sacrificing too much optical clarity. Researchers continue to explore doped composites and novel processing (e.g. grain refinement, domain engineering) to maximize electromechanical coupling while minimizing optical loss. The appropriate choice thus depends on application priorities: for high optical finesse (e.g. ultra-low-loss photonics), AlN or ZnO are preferred; for maximum strain or dielectric tuning (e.g. actuators), PZT/PLZT may be necessary, with careful optical design; for flexible or hybrid systems, PVDF-based films provide an optically invisible piezoelectric layer. Each material involves a compromise, but ongoing advances are steadily mitigating these trade-offs.

Sources: The comparison table and summary above are based on experimental literature values and reports, with citations provided for key data (refractive indices, transmission spectra, absorption coefficients, etc.). The sources include spectroscopic measurements of thin-film optical constants, device demonstrations of optical loss in doped AlN and PZT films, and reviews on transparent piezoelectric ceramics. These references support the quantitative claims and illustrate the state-of-the-art for integrating piezoelectric materials into transparent devices.

Sources

ChatGPT can make mistakes. Check important info.

Fabry–Perot Microcavity Platforms with Concave Mirrors (Limited DOF Tuning)

Free-space Fabry–Perot microcavities consist of two facing mirrors forming a small optical resonator. To achieve a tightly confined mode (high Purcell factor) while maintaining stability and high finesse, at least one mirror is often concave. Below, we review such microcavity platforms that incorporate concave mirrors, with at most one active degree of freedom for tuning the cavity length. We exclude fully adjustable (multi-actuator) open cavities and planar in-plane resonators, focusing instead on monolithic or minimally tunable designs.

Fiber-Based Concave Microcavities

One widely used platform is the fiber-based Fabry–Perot microcavity, where a concave mirror is fabricated on the tip of an optical fiber. The concave fiber tip is typically made by laser ablation (e.g. CO₂ laser pulses melting the glass) or focused-ion-beam milling, producing a microscopic spherical depression. The fiber’s concave mirror is automatically aligned to the fiber core, ensuring efficient coupling of the cavity mode into the fiber (coupling efficiency >90% has been achieved). The other mirror is usually a macroscopic planar mirror or another concave fiber, forming a small air-gap cavity. After coating with high-reflectivity dielectric layers, such fiber-tip cavities can reach finesse $F$ on the order of $104$–$105$ and sub-picoliter mode volumes. For example, Hunger et al. (2010) reported a fiber cavity with finesse ∼37,000 at 780 nm and mode waist ~2.8 µm, while Muller et al. (2010) achieved an ultrahigh-finesse ~300,000 in a 6 µm length fiber microcavity. Fiber microcavities are tunable in length by moving one fiber along its axis (often via a single piezo actuator) – providing spectral tunability without requiring multi-axis alignment once initially set. Notably, Greuter et al. (2014) used CO₂-laser ablation to create a 10 µm radius concave fiber mirror, yielding a cubic-wavelength-scale mode volume (waist ~1 µm) and demonstrating sub-wavelength beam waists. Such small mode volumes, together with fiber integration, make these cavities attractive for quantum optics.

Integration with quantum emitters: Fiber-based cavities have been successfully coupled to solid-state quantum emitters. For instance, Albrecht et al. (2013) coupled a single NV center in diamond nanocrystal to a tunable fiber microcavity, observing enhanced emission. In that system, the fiber mirror was concave (CO₂-machined) and the cavity length ~3–4 µm, yielding a Purcell factor on the order of 10. Similarly, fiber cavities have been used to interface with semiconductor quantum dots (in membranes) and cold atoms. These demonstrations typically rely on an initial multi-axis alignment of the fiber to the emitter chip, but once aligned the cavity length is the only actively tuned parameter (e.g. by a piezo) during operation. Recent efforts aim to package fiber microcavities in a self-aligned manner so that only one degree of freedom (axial motion) is needed. For example, Qing et al. (2019) etched and laser-reflowed a concave fiber tip of 4.7 µm diameter (matched to the fiber core) that, after coating, formed a microcavity with $Q>1000$ and mode volume $V\approx13.6~\mu$m³. This device was wavelength-tunable and even demonstrated microlaser emission, all while maintaining high fiber-coupling efficiency due to the self-aligned mirror.

Microfabricated Concave-Mirror Cavities (Chip-Based)

Another class of platforms uses microfabricated concave mirrors on substrates (chips) to form stable microcavities. These typically involve fabricating an array of concave mirror templates on a flat substrate via lithographic or direct-write techniques, then combining that substrate with a second mirror (often a flat mirror or another concave mirror) to create many microcavities. A key goal is to achieve small mirror radii and smooth surfaces for high finesse, using monolithic fabrication so that alignment is fixed by design (with at most one tuning axis, usually the gap). Several fabrication methods have been developed:

  • Focused Ion Beam (FIB) milling: Dolan et al. (2010) pioneered FIB to mill concave pits (~5–20 µm radius) in a silica substrate, which were then dielectric-coated to serve as mirrors. They reported arrays of femtoliter-volume, open-access microcavities with mode volumes as low as 2.2 µm³ and finesse up to ~460. FIB allows flexible mirror geometries, and later work improved the finesse into the $104$–$105$ range. However, FIB is serial and can introduce surface roughness (limiting finesse). It remains useful for prototyping; for example, a FIB-fabricated concave mirror by Patel et al. achieved finesse $\sim104$–$105$, and arrays of FIB mirrors have been used for tunable cavity-QED experiments.

  • Laser micromachining: CO₂ laser ablation of glass can produce ultra-smooth concave profiles. This was used by groups like Warburton’s: Greuter et al. (2014) produced atomically smooth concave dimples (radius ~10–20 µm) in fused silica by CO₂ laser pulses. Pairing such a miniaturized curved mirror with a planar DBR mirror, they realized a high-finesse ($F\sim10^4$) microcavity with mode volume on the order of $(\lambda)^3$. The cavity was spatially and spectrally tunable, yet mechanically rigid enough to operate at cryogenic temperatures with sub-picometer stability. Laser-ablated concave mirrors on optical fibers or chips have achieved finesse up to $10^5$ in later works. Takahashi et al. (2014) and others employed ultrafast lasers to fabricate micron-scale concave features with good mode matching.

  • Isotropic etching and wafer fabrication: An early foundational work by Trupke et al. (2005) demonstrated a microfabricated high-finesse cavity using isotropic wet etching of silicon. In this approach, a concave pit is etched (e.g. with an isotropic HF or XeF₂ etch) and then coated with a DBR. Trupke’s device had mirror radius $R\sim150~\mu$m, cavity length ~$130~\mu$m, finesse on the order of a few thousand, and was intended for coupling to cold atoms. Subsequent improvements by the same team and others greatly enhanced the performance. Biedermann et al. (2010) achieved ultra-smooth concave silicon mirrors (surface roughness $\sim$2.2 Å) via a combination of plasma etching and thermal oxidation smoothing. Their cavities reached finesse $F\approx64{,}000$ at 780 nm. More recently, Wachter et al. (2019) reported silicon concave-mirror microcavity arrays with finesse up to $5×10^5$ (measured via linewidth and ring-down). They fabricated 100 concave mirrors (radius $R\sim100$–300 µm) by a CMOS-compatible two-step plasma etch plus oxidation smoothing, achieving mirror losses below 3 ppm. By bonding two such mirror chips with a precisely defined spacer, they formed rigid open-access cavities without any post-assembly alignment. Impressively, the best plano-concave cavities had $F\approx5.3×10^5$ (mode volume $V\sim330~\text{fL}\approx88,\lambda^3$) and $Q\sim1.1×10^7$. Even concave–concave symmetric cavities were demonstrated using lithographically aligned mirror pairs: alignment holes and micro-pillar spacers in the chips ensured <5 µm lateral accuracy. This yielded stable cavities with $F$ up to $(4.0\pm0.1)×10^5$ at $L\approx150~\mu$m ($R\approx200~\mu$m), corresponding to $Q\approx8×10^7$. Crucially, these wafer-bonded microcavities require no manual tuning of alignment – once the chips are bonded, only the cavity length (set by spacer thickness) determines resonance. Small length adjustments (one DOF) can be introduced by piezo actuation of one chip if needed, but the spatial overlap is fixed by the bonding process.

  • Photoresist reflow & etching: A very recent advance is the use of reflowed polymer microlenses as mirror templates. In 2022, Westly et al. (NIST) reported fabrication of concave mirrors via photoresist thermal reflow and transfer etching, achieving unprecedented surface quality. In this method, photoresist droplets are melted to form near-perfect spherical caps, then etched into a substrate and coated. The result was an array of microcavities with measured finesse up to 1.3 million and scatter losses <1 ppm. Radii of curvature could be user-defined from ~50 µm up to many mm in the same wafer. The average coating-limited $F\approx1.05×10^6$ across 43 devices (at 1560 nm), demonstrating scalable, ultrahigh-finesse mirror fabrication. This approach marries semiconductor processing with traditional mirror quality, opening the door to integrated microcavities for quantum and precision metrology on a chip.

Tuning and emitter integration: Chip-based concave microcavities are often assembled in a fixed geometry (or with a single tunable axis). For example, Dolan’s 2010 FIB-mirror array was mounted with a macroscopic mirror on a translation stage to tune all cavities’ length uniformly. In sensing applications, some bonded-cavity systems incorporate a PZT or thermal element to slightly adjust the gap for calibration – but typically only one actuator is used (no multi-axis controllers) because the mirror alignment is lithographically pre-set. These microcavities have been used to enhance and study emitters as well. Najer et al. (2019) embedded a gated GaAs quantum dot in a microfabricated open cavity with $Q\sim10^6$, achieving the strong-coupling regime of cavity QED. (This used a fiber-aligned open cavity, which does allow multi-DOF alignment, but it underscores the potential of high-$Q/V$ cavities for emitters.) On the chip-integrated side, Fait et al. (2021) demonstrated high-finesse silicon microcavities operating at 1280 nm (telecom O-band) with Erbium dopants, aiming for Purcell-enhanced emission for quantum memory applications. The general trend is that as microfabricated cavity finesse and quality have improved, researchers are incorporating emitters (quantum dots, rare-earth ions, etc.) either by bonding material into the cavity or by positioning the cavity around the emitter – all while maintaining minimal degrees of freedom in operation.

Monolithic “Buckled-Dome” Microcavities

A distinct approach to concave microcavities is the monolithic air-gap Fabry–Perot formed by a buckled mirror. In these devices, both mirrors are fabricated in place on a single substrate stack, and a controlled stress release causes one mirror to deform into a concave “dome” shape. The group of DeCorby and Davis introduced this concept: in 2011, Allen et al. used a multilayer $\text{a-Si/SiO}_2$ film under residual stress, which upon release buckled into a curved mirror (like a tiny dome). Two such curved DBR mirrors facing each other can form a microcavity. More commonly, a half-symmetric design is used: a planar DBR mirror on a substrate and a second DBR that buckles outward to create a concave top mirror. The cavity is an “open-access” air gap, but it is monolithically self-aligned – the dome remains concentric with the planar mirror below it, by virtue of the fabrication process.

Fabrication and performance: Potts et al. (2016) reported on-chip dome microcavities fabricated by this guided self-assembly method. They deposited a dielectric DBR (e.g. 10–15 alternate layers of $\text{SiO}2$ and $\text{Ta}_2\textO_5$) on a substrate, then a thin sacrificial release layer, and finally another DBR. Heating the stack caused the top DBR to delaminate in circular regions (defined by photolithography) and _buckle upward due to compressive stress. The result is a concave mirror of diameter 100–300 µm and sag height ~2.5–10 µm (controllable by diameter). Warburton’s group characterized similar buckled cavities, noting interference fringe patterns confirming high symmetry (cylindrical curvature). A typical 100 µm-diameter dome cavity had an optical length $L\approx3.3~µ$m and effective radius $R_c\approx210~µ$m. With high-reflectivity DBRs (designed for 780 nm), these devices achieved finesse $F\approx3500$ (consistent with the 99.91% mirror reflectivity). The measured fundamental mode linewidth was ~27 pm (at $\sim778$ nm), matching $F\approx3560$. The mode waist was $w_0\approx2.5~µ$m, giving a mode volume $V \approx 35λ^3$ (for $\lambda\sim780$ nm). These values indicate a reasonably small mode volume and moderate finesse – suitable for cavity QED in the weak to intermediate coupling regime.

Because the mirrors are fixed on-chip, tuning is limited to one degree (global cavity length changes, e.g. via temperature or strain). Bitarafan et al. (2015) showed that thermal expansion can tune the resonance of buckled-dome cavities over several free spectral ranges. In one example, heating from 20°C to 100°C red-shifted the cavity mode by >0.5 nm, enough to align to atomic transitions. The thermal stability is high; Potts et al. noted that resonance could be stabilized to within $\pm 5$ pm in a cryostat. In principle, one could integrate microheaters or MEMS actuators to adjust the dome height (one axis); however, no multi-axis alignment is needed since the mirrors are permanently concentric. Recent studies even observed optical bistability in these air-gap cavities due to intracavity heating, a sign that the cavity-emitter or cavity-optomechanical coupling is significant.

Integration with emitters: The buckled-cavity platform is appealing for interfacing with atomic systems, since it can be made to resonate at atomic lines (e.g. Rb D₂ at 780 nm) and potentially integrated with vapor cells or cold atom setups. Indeed, the 2016 devices were designed to be tunable to the Rb D₂ line. In a solid-state context, one could imagine depositing a thin emitter layer before mirror deposition so that emitters reside inside the cavity. To date, most demonstrations have been passive; however, Merkel et al. (2020) inserted erbium ions (at 1530 nm) into a cryogenic high-Q microcavity, achieving Purcell-enhanced coherent emission. That system utilized a macroscopic reference cavity, but it foreshadows using on-chip DBR microcavities for emitter coupling. The buckled-dome approach provides an integrated, alignment-free microcavity with one tunable parameter (e.g. temperature). The trade-off is somewhat larger mode volume compared to fiber or etched microcavities, but ongoing work aims to reduce the diameter and increase the curvature of buckled mirrors for tighter confinement.

Summary and Outlook

In summary, a variety of Fabry–Perot microcavity platforms with concave mirrors have been developed, balancing high optical performance with limited degrees of freedom for tuning:

  • Fiber-tip microcavities (concave fiber + mirror) achieve small volumes and high finesse with built-in fiber coupling, requiring only one axis of motion (length) during use. They have enabled strong light-matter interaction with NV centers and quantum dots, and even form the basis of microcavity lasers.

  • Microfabricated mirror cavities on chips (concave pits made by FIB, laser, or etching) offer scalable arrays and extremely high finesse. Foundational work showed finesses ~$103$–$104$, and recent monolithic assemblies now reach $F>105$–$106$ with sub-λ³ volumes. These systems are often bonded or fixed such that only the cavity length can be tuned (e.g. via a piezo or strain), while mirror alignment is lithographically defined. This stability makes them promising for quantum networks and precision sensors.

  • Buckled-dome (self-assembled) cavities integrate the entire FP cavity on one chip, with a curved DBR mirror emerging through stress engineering. They require no post-fabrication alignment and can be thermally or mechanically tuned in length. Achieved finesses ($\sim10^3$) are lower than the best open cavities, but adequate for many strong-coupling experiments, and improvements in mirror coatings could push this higher.

Each platform involves different mirror fabrication techniques – from laser-reflowed glass (giving atomically smooth concave surfaces), to FIB milled pits (flexible patterning), to thermal shaping of polymers, to strained film buckling. Crucially, all use at least one concave mirror to support a stable, near-Gaussian mode despite the microscopic cavity length, which mitigates diffraction loss and allows finesses comparable to macroscopic supermirrors. By restricting the design to at most one active tuning axis, these platforms forego the full adjustability of macroscopic cavities in exchange for robustness and scalability (no constant re-alignment). This has enabled deploying microcavities in challenging settings: cryogenic setups, vibration-prone environments, or large arrays for sensing.

Looking forward, such concave-mirror microcavities are poised to play a key role in quantum photonics and sensing. Ongoing research is integrating solid-state emitters (quantum dots, color centers, 2D materials) into these tiny cavities to realize efficient single-photon sources and strong emitter–cavity coupling on chip. The combination of high $Q$ and low $V$ (for large Purcell factors) with minimal alignment overhead makes these platforms attractive for scalable quantum devices. Advances like the photoresist-reflow mirrors reaching $F>10^6$ suggest that performance no longer needs to be sacrificed for compactness. In short, the field has progressed from early tunable microcavities with modest finesse to today’s integrated microcavities rivaling macroscopic resonators in $Q$ – all while maintaining one simple degree of freedom for tuning. This marriage of precision optics and microfabrication will likely underpin next-generation quantum interfaces and sensors built around Fabry–Perot microcavity platforms.

Sources:

  • Hunger et al., NJP 12, 065038 (2010) – High-finesse fiber Fabry–Perot design

  • Muller et al., Opt. Lett. 35, 2293 (2010) – Ultrahigh-$F$ low-$V$ microcavity

  • Dolan et al., Opt. Lett. 35, 3556 (2010) – FIB-fabricated femtoliter cavity arrays

  • Greuter et al., APL 105, 121105 (2014) – Small-mode-volume CO₂-laser microcavity

  • Bitarafan et al., JOSA B 32, 1214 (2015) – Buckled dome microcavities (thermo-optic tuning)

  • Potts et al., APL 108, 041103 (2016) – On-chip open-access dome cavity for cQED

  • Wachter et al., Nat. Commun. 10, 5020 (2019) – Si microcavity array, $F>5×10^5$

  • Najer et al., Nature 575, 622 (2019) – Strongly coupled QD–microcavity ($Q∼10^6$)

  • Westly et al., arXiv:2203.15931 (2022) – Reflow-etched mirrors with $F>10^6$

  • Albrecht et al., PRL 110, 243602 (2013) – NV center in fiber microcavity

  • (Additional citations provided in-line.)

Citations

[

On-Chip High-Finesse Fabry-Perot Microcavities for Optical Sensing and Quantum Information

https://www.mdpi.com/1424-8220/17/8/1748

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=Many%20of%20the%20shortcomings%20of,were%20characterized%20by%20relatively%20large)[

On-Chip High-Finesse Fabry-Perot Microcavities for Optical Sensing and Quantum Information

https://www.mdpi.com/1424-8220/17/8/1748

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=analytes%20or%20atomic%20emitters%2C%20etc,M%7D%20remains%20at%20a)[

A simple approach to fiber-based tunable microcavity with high coupling efficiency | Request PDF

https://www.researchgate.net/publication/330463548_A_simple_approach_to_fiber-based_tunable_microcavity_with_high_coupling_efficiency

](https://www.researchgate.net/publication/330463548_A_simple_approach_to_fiber-based_tunable_microcavity_with_high_coupling_efficiency#:~:text=a%20concave%20mirror%20on%20the,6%20%CE%BCm%C2%B3%2C%20and)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=Hunger%C2%A0et%C2%A0al,m)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=Biedermann%202010%20Biedermann%C2%A0et%C2%A0al,780%20nm%20%20Projected%3A%20130k)[

[PDF] Ultrahigh-finesse, low-mode-volume Fabry–Perot microcavity

https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=905212

](https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=905212#:~:text=Ashmore%2C%20and%20E,13%20%2F%20OPTICS%20LETTERS)[

[1408.1357] A small mode volume tunable microcavity: development and characterization

https://arxiv.org/abs/1408.1357

](https://arxiv.org/abs/1408.1357#:~:text=,With%20these%20small)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=,2013)[

On-Chip High-Finesse Fabry-Perot Microcavities for Optical Sensing and Quantum Information

https://www.mdpi.com/1424-8220/17/8/1748

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=mirrors%20were%20originally%20reported%20by,finesse%20as%20high%20as%2040%2C000)[

Femtoliter tunable optical cavity arrays - ORA - Oxford University Research Archive

https://ora.ox.ac.uk/objects/uuid:1c373433-7b05-48dc-add8-73a0d76e5050

](https://ora.ox.ac.uk/objects/uuid:1c373433-7b05-48dc-add8-73a0d76e5050#:~:text=Abstract%3A%20Large%20arrays%20of%20uniform%2C,2010%20Optical%20Society%20of%20America)[

On-Chip High-Finesse Fabry-Perot Microcavities for Optical Sensing and Quantum Information

https://www.mdpi.com/1424-8220/17/8/1748

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=match%20at%20L1182%20mirrors%20were,finesse%20as%20high%20as%2040%2C000)[

[1408.1357] A small mode volume tunable microcavity: development and characterization

https://arxiv.org/abs/1408.1357

](https://arxiv.org/abs/1408.1357#:~:text=silica.%20We%20demonstrate%20excellent%20mode,fluctuations%20are%20below%205%20pm)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=radius%20of%20curvature%20mirrors%20,Jacobs%C2%A0et%C2%A0al.%20%282012)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=%282010%20%29%3B%20Muller%C2%A0et%C2%A0al,they%20are%20limited%20to%20the)[

Silicon microcavity arrays with open access and a finesse of half a million - PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=,Strong%20atom%E2%80%93field%20coupling%20for%20Bose%E2%80%93Einstein)[

Silicon microcavity arrays with open access and a finesse of half a million - PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=curvature%2C%20we%20use%20a%20complementary,quality%20approaching%20the%20atomic%20limit)[

Silicon microcavity arrays with open access and a finesse of half a million - PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=a%20silicon%20wafer,to%20create%20a%20range%20of)[

Silicon microcavity arrays with open access and a finesse of half a million - PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=measurement%20yields%20a%20lifetime%20of,Cumulative%20distribution%20of%20the%20finesse)[

Silicon microcavity arrays with open access and a finesse of half a million - PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=determined%20as%20in%20,transmission%20of%20the%20mirror%20coatings)[

Silicon microcavity arrays with open access and a finesse of half a million - PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=curvature%2C%20we%20use%20a%20complementary,quality%20approaching%20the%20atomic%20limit)[

Silicon microcavity arrays with open access and a finesse of half a million - PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=modulation%20of%20350%20MHz,green)[

Silicon microcavity arrays with open access and a finesse of half a million - PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=1%2C%20column%201%20to%20row,photons%20from%20integrated%20quantum%20dots16)[

Silicon microcavity arrays with open access and a finesse of half a million - PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=Furthermore%2C%20we%20created%20rigidly%20assembled,such%20high%20finesse%20could%20be)[

Silicon microcavity arrays with open access and a finesse of half a million - PMC

https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=accurate%20and%20robust%20alignment%20of,that%20the%20micromachined%20cavity%20pre)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=of%20curvature%20spanning%20four%20orders,finesse)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=generation%20compact%20atomic%20clocks,small%20and%20large%20radii%20of)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=Image%3A%20Refer%20to%20caption%20Figure,modify%20the%20impact%20of%20surface)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=finesse%20mirrors%20relies%20on%20centuries,This%20substrate%20is%20coated)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=resonators%20with%20finesse%20values%20as,optics%20and%20frequency%20metrology%20technologies)[

A gated quantum dot far in the strong-coupling regime of cavity-QED ...

https://arxiv.org/abs/1812.08662

](https://arxiv.org/abs/1812.08662#:~:text=A%20gated%20quantum%20dot%20far,tunable%20quantum%20dot)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=,%282020%29%20Benjamin%C2%A0Merkel%2C%20Alexander%C2%A0Ulanowski)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=,2020)[

On-Chip High-Finesse Fabry-Perot Microcavities for Optical Sensing and Quantum Information

https://www.mdpi.com/1424-8220/17/8/1748

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=20,Lab%20Chip%202015%2C%2015)[

On-Chip High-Finesse Fabry-Perot Microcavities for Optical Sensing and Quantum Information

https://www.mdpi.com/1424-8220/17/8/1748

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=buckled%20dome%20Fabry%E2%80%93Perot%20microcavities,211%20Google)[

https://arxiv.org/pdf/1601.03344

](https://arxiv.org/pdf/1601.03344#:~:text=Following%20the%20deposition%20of%20the,de%02tail%20the%20100%20%C2%B5m%20diameter)[

https://arxiv.org/pdf/1601.03344

](https://arxiv.org/pdf/1601.03344#:~:text=,and%20a%20simulated%20optical%20mode)[

https://arxiv.org/pdf/1601.03344

](https://arxiv.org/pdf/1601.03344#:~:text=where%20Rc%20is%20the%20radius,implying%20that%20Vm%20%E2%89%88%2035%CE%BB)[

https://arxiv.org/pdf/1601.03344

](https://arxiv.org/pdf/1601.03344#:~:text=%CE%BA%20%3D%20%CF%80c%2F2LF%2C%20an%20estimate,agreement%20with%20the%20predicted%20value)[

https://arxiv.org/pdf/1601.03344

](https://arxiv.org/pdf/1601.03344#:~:text=in%20greater%20detail%2C%20with%20a,nH%20%E2%88%92%20nL%29%20%E2%88%921)[

https://arxiv.org/pdf/1601.03344

](https://arxiv.org/pdf/1601.03344#:~:text=The%20100%20%C2%B5m%20buckled,Similar%20values)[

Thermomechanical characterization of on-chip buckled dome Fabry ...

https://ui.adsabs.harvard.edu/abs/2015JOSAB..32.1214B/abstract

](https://ui.adsabs.harvard.edu/abs/2015JOSAB..32.1214B/abstract#:~:text=,guided%20assembly%20of%20circular)[

[1408.1357] A small mode volume tunable microcavity: development and characterization

https://arxiv.org/abs/1408.1357

](https://arxiv.org/abs/1408.1357#:~:text=the%20effective%20optical%20radius%20matches,fluctuations%20are%20below%205%20pm)[

Bistability in buckled dome microcavities - PubMed

https://pubmed.ncbi.nlm.nih.gov/26565878/

](https://pubmed.ncbi.nlm.nih.gov/26565878/#:~:text=Bistability%20in%20buckled%20dome%20microcavities,fabricated%20by%20controlled%20formation)[

https://arxiv.org/pdf/1601.03344

](https://arxiv.org/pdf/1601.03344#:~:text=ing%20an%20ellipsometer,on%20sin%02gle%20films%2C%20a%20peak)[

[2203.15931] Micro-fabricated mirrors with finesse exceeding one million

https://ar5iv.labs.arxiv.org/html/2203.15931

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=match%20at%20L465%20,2020)[

[PDF] Integrated Devices for On-Chip Quantum Optics - SciSpace

https://scispace.com/pdf/integrated-devices-for-on-chip-quantum-optics-1nft2o2r47.pdf

](https://scispace.com/pdf/integrated-devices-for-on-chip-quantum-optics-1nft2o2r47.pdf#:~:text=SciSpace%20scispace,optics%20applications)[

https://arxiv.org/pdf/1601.03344

](https://arxiv.org/pdf/1601.03344#:~:text=cQED%20applications%2C%20it%20should%20allow,The)[

On-Chip High-Finesse Fabry-Perot Microcavities for Optical Sensing and Quantum Information

https://www.mdpi.com/1424-8220/17/8/1748

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=12,mode%20volume%20tunable%20microcavity%3A%20Development)[

On-Chip High-Finesse Fabry-Perot Microcavities for Optical Sensing and Quantum Information

https://www.mdpi.com/1424-8220/17/8/1748

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=Reichel%2C%20J,CrossRef)[

On-Chip High-Finesse Fabry-Perot Microcavities for Optical Sensing and Quantum Information

https://www.mdpi.com/1424-8220/17/8/1748

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=15,expressions%20for%20the%20reflection%20delay)[

On-Chip High-Finesse Fabry-Perot Microcavities for Optical Sensing and Quantum Information

https://www.mdpi.com/1424-8220/17/8/1748

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=16,CrossRef)

All Sources

[

mdpi

](https://www.mdpi.com/1424-8220/17/8/1748#:~:text=Many%20of%20the%20shortcomings%20of,were%20characterized%20by%20relatively%20large)[

researchgate

](https://www.researchgate.net/publication/330463548_A_simple_approach_to_fiber-based_tunable_microcavity_with_high_coupling_efficiency#:~:text=a%20concave%20mirror%20on%20the,6%20%CE%BCm%C2%B3%2C%20and)[

ar5iv.labs.arxiv

](https://ar5iv.labs.arxiv.org/html/2203.15931#:~:text=Hunger%C2%A0et%C2%A0al,m)[

tsapps.nist

](https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=905212#:~:text=Ashmore%2C%20and%20E,13%20%2F%20OPTICS%20LETTERS)[

arxiv

](https://arxiv.org/abs/1408.1357#:~:text=,With%20these%20small)[

ora.ox.ac

](https://ora.ox.ac.uk/objects/uuid:1c373433-7b05-48dc-add8-73a0d76e5050#:~:text=Abstract%3A%20Large%20arrays%20of%20uniform%2C,2010%20Optical%20Society%20of%20America)[

pmc.ncbi.nlm.nih

](https://pmc.ncbi.nlm.nih.gov/articles/PMC6456601/#:~:text=,Strong%20atom%E2%80%93field%20coupling%20for%20Bose%E2%80%93Einstein)[

ui.adsabs.harvard

](https://ui.adsabs.harvard.edu/abs/2015JOSAB..32.1214B/abstract#:~:text=,guided%20assembly%20of%20circular)[

pubmed.ncbi.nlm.nih

](https://pubmed.ncbi.nlm.nih.gov/26565878/#:~:text=Bistability%20in%20buckled%20dome%20microcavities,fabricated%20by%20controlled%20formation)[

scispace

](https://scispace.com/pdf/integrated-devices-for-on-chip-quantum-optics-1nft2o2r47.pdf#:~:text=SciSpace%20scispace,optics%20applications)