lithium niobate / 04 Quirks / 4.10
4.10 Fabrication pitfalls: practitioner notes
L1 Phenomenon
Fabrication habits that work on silicon fail on lithium niobate. It is an insulator that charges under electron beams, a pyroelectric that charges on a hotplate, and a crystal whose two polar faces etch at different rates. Source: public practitioner notes (2018-2025); mechanism and primary source added where one exists.
Fabrication pitfalls: cause, mechanism, number, fix, statusnumbers from the 4.6 plate model
| Pitfall | Cause | Mechanism | Number | Fix as recorded | Status |
|---|---|---|---|---|---|
| Hotplate bake | Bakes fry devices or crack the whole substrate; worse on z-cut. | Pyroelectric charge: a 25 → 120 °C step on a 0.5 mm z-cut wafer, unscreened (4.6). | 7.5–17.8 kV | Conductive coating or aluminum-foil wrap as a discharge path. | practitionerderived |
| z-cut interferometers | 1980s z-cut channel-waveguide interferometers drift more than x-cut. | z-cut pyroelectric instability; unscreened field per kelvin on a z-cut plate: | 1.57–3.74 kV/cm per K | x-cut layout. | reportedpractitioner |
| 4 K cooldown | An LN chip's field damaged a device on another chip; released x-cut structures on Si collapsed after a change of wafer. | Pyroelectric charge on cooling; collapse attributed in the note to charge plus electrostatic pull-in (hypothesis). No p3 value below room temperature was found, so no number. | no number | Shielding electrodes; Al electrodes shorting LN parts to the Si substrate stopped the collapse. | practitionerU (specifics) |
| E-beam lithography | An insulating substrate charges under the electron beam. | Charging; a discharge layer and a clamp touching it carry the charge away. | 5-10 nm Al | Thin low-melting-point metal (Al), removed in an alkaline developer; a 5 nm Ti layer once ran hot and cross-linked the resist top. | practitioner |
| a-Si hard mask, Ar+ mill | An a-Si mask on oxide over y-cut TFLN cracked and etched non-uniformly; cracks transferred to the LN edges; the mask ran out first. | Film stress and fragile mask edges; films grown on LN may differ from the same films on Si (the note's hypothesis). | no number | Lower mill angle tried; conclusion: mask poor and fragile near edges. | practitioner |
| HF clean | Hole sidewalls of a y-cut nanobeam came out asymmetric after HF. | The −c end of domains etches much faster than +c, so sidewalls facing −c and +c recede differently. Which side faced +c is not recorded. | −c > +c | None stated; the chip had been annealed before fabrication. | P-comppractitioner |
p3, C/(m² K)
practitioner notes: public posts, cited by kind and datebake voltage: plate model of 4.6 derived−c etch rule P-comp
Six fabrication pitfalls
click a card for mechanism, number and fixPyroelectric coefficient, 10⁻⁵ C/(m² K)
practitioner notes, cited by kind and date practitionerbake voltage derived−c etch rule P-compcooldown specifics U
L2 Notes and mechanisms
| # | Pitfall (note date) | Mechanism and number | Fix in the note | Status |
|---|---|---|---|---|
| 1 | Hotplate bakes fry devices or crack the whole substrate; worse on z-cut (2025-12-29) | pyroelectric charge; a 95 K step gives 7.5-17.8 kV across a 0.5 mm z-cut wafer, unscreened (4.6) | conductive coating or aluminum-foil wrap as a discharge path | practitioner derived |
| 2 | 1980s z-cut channel-waveguide interferometers drift more than x-cut (2025-12-29) | z-cut pyroelectric instability | x-cut layout | Bulmer et al. 1986reported |
| 3 | During a 4 K cooldown an LN chip's field damaged a device on another chip; shielding electrodes protected the neighbor; interdigital transducers on the LN chip still failed (2025-12-29) | pyroelectric charge on cooling | shielding electrodes | practitioner U (specifics) |
| 4 | Released x-cut TFLN structures on Si collapsed during 4 K cooldowns after a change of wafer (2025-12-29) | hypothesis in the note: pyroelectric charge plus electrostatic pull-in; CTE mismatch ruled out because the onset was sudden | Al electrodes shorting LN parts to the Si substrate; collapse stopped | practitioner hypothesis |
| 5 | Heating z-cut LN to 120 °C could expose resist (2018) | pyroelectric electron emission from the polar face (Bourim et al. 2006reported (title)) | U | |
| 6 | Electron-beam lithography on LN needs a discharge layer; 5-10 nm Al preferred; a 5 nm Ti layer once cross-linked the top of the resist because it ran too hot; the clamp must contact the discharge layer (2024-10-13) | charging of an insulating substrate | thin low-melting-point metal (Al), removed in an alkaline resist developer | practitioner |
| 7 | a-Si hard mask on oxide over y-cut TFLN cracked and etched non-uniformly under Ar+ milling; cracks transferred to LN edges; the mask ran out before LN was etched through (2018-08-20) | film stress and fragile mask edges; films grown on LN may differ from the same films on Si (the note's hypothesis) | lower mill angle tried; conclusion: mask poor and fragile near edges | practitioner |
| 8 | HF clean: post-HF SEM of a y-cut nanobeam shows asymmetric hole sidewalls (2018) | the −c end of domains etches much faster than +c (Volk & Wöhlecke 2009, p. 168P-comp; Nassau et al. 1966reported (title), Barry et al. 1998reported (title), Sones et al. 2001reported (title)); in x- and y-cut devices sidewalls facing −c and +c recede differently | none stated; the chip was annealed before fabrication | practitioner |
| 9 | Ion milling depends strongly on residual gas partial pressure; O2 oxidizes metal masks or ashes resist (2018) | practitioner | ||
| 10 | Li does not form volatiles with Br or Cl (2018) | U |
L3 Limits of the notes
- The orientation of c in the post-HF SEM is not recorded in the note; the asymmetry and the −c rule are stated, not which side is which.
- Bake voltages are ranges spanning the pyroelectric coefficient spread (4.6); cooldowns carry no number: no p3 value below room temperature was found.
- Charging under electron beams and on hotplates, face-dependent etching and mask stress connect to chapter 05 (milling, lithography, poling).
Sources
- practitioner note 2025-12-29Practitioner note (public post), 2025-12-29.
- practitioner note 2018-04-28Practitioner note (public post), 2018-04-28.
- practitioner note 2024-10-13Practitioner note (public post), 2024-10-13.
- practitioner note 2018-08-20Practitioner note (public post), 2018-08-20.
- Bulmer et al. 1986C. H. Bulmer, W. K. Burns, S. C. Hiser, Pyroelectric effects in LiNbO3 channel-waveguide devices. Applied Physics Letters 48 1036-1038 (1986), doi:10.1063/1.96640.Read: abstract
- Bourim et al. 2006El Mostafa Bourim, Chang-Wook Moon, Seung-Woon Lee, In Kyeong Yoo, Investigation of pyroelectric electron emission from monodomain lithium niobate single crystals. Physica B: Condensed Matter 383 171-182 (2006), doi:10.1016/j.physb.2006.02.034.Read: title
- Nassau et al. 1966K. Nassau, H. J. Levinstein, G. M. Loiacono, Ferroelectric lithium niobate. 1. Growth, domain structure, dislocations and etching. Journal of Physics and Chemistry of Solids 27 983-988 (1966), doi:10.1016/0022-3697(66)90070-9.Read: title
- Barry et al. 1998I. E. Barry, G. W. Ross, P. G. R. Smith, R. W. Eason, G. Cook, Microstructuring of lithium niobate using differential etch-rate between inverted and non-inverted ferroelectric domains. Materials Letters 37 246-254 (1998), doi:10.1016/S0167-577X(98)00100-1.Read: title
- Sones et al. 2001C. L. Sones, S. Mailis, W. S. Brocklesby, R. W. Eason, J. R. Owen, Differential etch rates in z-cut LiNbO3 for variable HF/HNO3 concentrations. Journal of Materials Chemistry 12 295-298 (2001), doi:10.1039/b106279b.Read: title
- Volk & Wöhlecke 2009Tatyana Volk, Manfred Wöhlecke, Lithium Niobate: Defects, Photorefraction and Ferroelectric Switching. Springer Series in Materials Science 115 (2009), doi:10.1007/978-3-540-70766-0.Read: selected chapters: Ch. 2.2 (pp. 14-19), Ch. 3 (pp. 51-61), Ch. 4 (pp. 75-77, 82-85, 94-97), Sec. 7.1.3 (p. 168)