Crystal Ion Slicing (CIS) in Photonics

• Crystal Ion Slicing (CIS) is a fabrication method that employs controlled He⁺ implantation to create a defined damage layer enabling the exfoliation of thin BaTiO₃ films.
• The process balances reliable flake release and implantation-induced lattice damage, necessitating precise post-slicing annealing to restore crystallinity and ferroelectric order.
• Scalable and CMOS-compatible, CIS offers a viable route for integrated photonic applications by recovering bulk-like optical responses and efficient nonlinear behavior.

Crystal Ion Slicing (CIS) is a thin-film fabrication route in which ion implantation creates a damaged layer at a controlled depth so that a crystalline flake can later be exfoliated and transferred. In BaTiO$_3$, a material described as compelling for integrated photonics because of its strong electro-optic and second-order nonlinear properties, CIS is presented as a scalable and CMOS-compatible route for fabricating thin films, but one whose ion implantation step introduces lattice damage that can degrade structural and optical performance. For CIS-processed BaTiO$_3$, post-slicing thermal annealing has been shown to restore structural integrity, recover optical quality, and re-establish functional ferroelectric and nonlinear-optical behavior, thereby qualifying CIS combined with annealing as a viable manufacturing strategy for BaTiO$_3$-on-insulator (BTOI) photonic platforms (Esfandiar et al., 7 Aug 2025).

1. Definition and material context

In the reported BaTiO$_3$ implementation, CIS is a sequence centered on He$^+$ implantation into double-side polished, (001)-oriented single crystal BaTiO$_3$ substrates, followed by thermally assisted exfoliation and transfer of a thin flake. The implantation step is used to create a damaged layer, or defect plane, at a depth corresponding to the desired flake thickness, reported as approximately $1.1$–$1.3\,\mu$m. This depth selection is central to the method because it defines the eventual thickness of the released crystalline layer (Esfandiar et al., 7 Aug 2025).

Within this framework, CIS is not described simply as a mechanical thinning method. Its defining feature is the deliberate use of ion-induced subsurface damage to localize fracture or delamination. For BaTiO$_3$, this is technically important because the same material attributes that motivate photonic integration—strong electro-optic response and second-order nonlinearity—are also sensitive to implantation-induced disorder. The paper therefore treats CIS as a coupled structural and functional processing problem: fabrication of a transferable thin crystal must be balanced against preservation, or later restoration, of crystallinity, ferroelectric order, and the associated nonlinear susceptibility.

2. Process sequence and operating parameters

The BaTiO$_3$ CIS process reported in the source is defined by specific implantation, pre-annealing, exfoliation, and post-annealing conditions.

Stage	Parameters	Reported outcome
Ion implantation	480 keV He$_3$0; $_3$1 or $_3$2 ions/cm$_3$3; 7° off-axis tilt	Damaged layer at $_3$4–$_3$5m; disorder; domain reorientation
Exfoliation	Hot plate 270 °C for 1 h; mechanical exfoliation	Helium bubble formation, blistering, interface weakening; flakes of $_3$6m
Thermal annealing	700 °C for 21 h in ambient air	Recovery of crystallinity, domain reorientation, enhanced $_3$7

The implantation stage uses 480 keV He$_3$8 at two fluences: a low fluence of $_3$9 ions/cm$_3$0 and a high fluence of $_3$1 ions/cm$_3$2. To avoid ion channeling, the samples are tilted 7° off-axis during implantation. After implantation, a hot plate pre-anneal at 270 °C for 1 hr induces helium bubble formation, blistering, and interface weakening. Mechanical exfoliation then yields flakes of approximately $_3$3m thickness, consistent with the target defect-plane depth and with SRIM-based expectations reported in the paper. Only high-fluence samples provide reliable flake release (Esfandiar et al., 7 Aug 2025).

This operating window reveals an intrinsic trade-off. The higher fluence is the regime that enables dependable slicing, yet the same regime also produces the strongest structural damage signatures. A plausible implication is that CIS process optimization in BaTiO$_3$4 cannot be reduced to implantation alone; it depends on a coupled implantation-and-recovery workflow in which defect generation is followed by sufficiently strong post-slicing annealing.

3. Implantation-induced structural disorder

The as-implanted state is characterized in the paper through Raman spectroscopy, X-ray diffraction (XRD), and Rutherford backscattering (RBS), and the reported signatures are consistent with implantation-induced strain, disorder, and loss of tetragonality. In pristine BaTiO$_3$5, Raman spectra exhibit clear tetragonal peaks, especially the B$_3$6 mode near $_3$7 cm$_3$8 and split A$_3$9(TO), E, and LO modes. After implantation, the B$_3$0 feature is suppressed, which is identified as indicating tetragonality and macro polar order loss, while the relative increase of A$_3$1(TO) modes is associated with domain reorientation. At high fluence, the Raman response undergoes overall reduction and broadening, approaching an amorphous or disordered state (Esfandiar et al., 7 Aug 2025).

The XRD response shows broadening and low-angle shift of the (002) reflection, whereas the (200) reflection remains visible but broadened. These features are identified as indicative of c-axis expansion, strain, and disorder. At high fluence, the loss of clear peak splitting is reported as evidence of decreased tetragonality. RBS independently shows enhanced surface backscattering yield, especially at high fluence, thereby quantifying the degree and depth of disorder.

Taken together, these probes indicate that CIS-relevant implantation damage is not restricted to a narrow sacrificial layer in a purely mechanical sense. It modifies lattice order, tetragonal symmetry, and polar-domain configuration in ways that are directly relevant to electro-optic and nonlinear-optical performance. This is the central limitation that motivates the subsequent annealing step.

4. Post-slicing annealing and structural restoration

Defect healing is performed after slicing by annealing the transferred flakes at 700 °C for 21 hours in ambient air in a conventional furnace. The stated purpose is to promote recrystallization, restore structural order, and remove or anneal out defects. Raman mapping before and after annealing, including k-means clustered maps, shows pronounced recovery of A$_3$2(TO) modes after annealing. An increased B$_3$3/A$_3$4(TO) ratio is reported as evidence of restored tetragonality and order, while partial recovery of the A$_3$5(LO) mode at 720 cm$_3$6 suggests partial return of c-domain character (Esfandiar et al., 7 Aug 2025).

The recovery is spatially nonuniform in a systematic way: it is reported to start from flake edges. The paper attributes this to healing via increased surface area and defect mobility. Flakes are also reported to display low surface roughness of approximately $_3$7–$_3$8 nm, and the transferred exfoliated layers are described as flat and clean.

A technical feature of this annealing protocol is that it operates above the BaTiO$_3$9 Curie temperature, reported as $^+$0C. This enables domains to reconfigure upon cooling. In that sense, annealing is not merely a defect-erasure step. It is simultaneously a crystallographic recovery step and a ferroelectric-state reset, coupling lattice recrystallization to domain reorganization.

5. Ferroelectric domains, SHG, and recovery of the nonlinear susceptibility

Polarization-resolved second-harmonic generation (SHG) microscopy is used to probe ferroelectric domain orientation through the nonlinear susceptibility tensor $^+$1. The reported setup uses 900 nm excitation and maps SHG intensity together with in-plane domain direction on a per-pixel basis. Before annealing, several domain orientations are present, and SHG intensity persists even in some regions where Raman signatures are weak. The paper interprets this as evidence that long-range ferroelectric order can survive despite partial lattice disruption (Esfandiar et al., 7 Aug 2025).

After annealing, domains realign and the angular distribution narrows, often along a preferred in-plane axis, although with some flake-to-flake variation. The magnitude and orientation of $^+$2 partially recover, and polarization-dependent SHG is fitted to a tensor model. The reported expressions are

$^+$3

$^+$4

$^+$5

with $^+$6 pm/V, $^+$7 pm/V, and $^+$8 pm/V for BaTiO$^+$9.

A common over-simplification would be to treat weak Raman response as equivalent to complete loss of ferroelectric functionality. The reported coexistence of weak Raman signatures with persistent SHG directly argues against that equivalence. In the paper’s terms, lattice order and long-range polar order can decouple under moderate disorder. This decoupling is one of the conceptually important outcomes of the study because it refines how implantation damage should be interpreted in ferroic photonic materials.

6. Linear optical response and implications for BTOI platforms

The linear optical properties of annealed CIS flakes are evaluated by reflectometry over 500–950 nm. Fabry-Pérot oscillations are observed, indicating well-defined film thickness and refractive index. Transfer matrix modeling using bulk BaTiO$_3$0 dispersion data reproduces the experimental spectra well, and the paper states that the linear dispersion of annealed CIS flakes closely matches that of bulk BaTiO$_3$1 (Esfandiar et al., 7 Aug 2025).

This optical result complements the Raman and SHG data. Raman recovery indicates substantial restoration of crystallinity; SHG demonstrates persistence and partial recovery of nonlinear functionality; reflectometry shows bulk-like linear optical behavior. The combined interpretation offered by the paper is that CIS, when paired with proper thermal annealing, yields high-quality thin BaTiO$_3$2 films with restored structural order, bulk-like linear optical response, and strong second-order nonlinear optical properties.

The manufacturing significance is stated explicitly in terms of BTOI integration. CIS is described as scalable and CMOS-compatible, suitable for integration on substrates such as Si or SiO$_3$3/Si, and enabling waveguide fabrication, high-$_3$4 photonic structures, integrated modulators, and frequency converters. The abstract further identifies modulation, frequency conversion, and quantum optics as target application areas. The additional suggestion that domain engineering may be possible through annealing protocols for programmable $_3$5 is presented in the source as a prospective direction rather than as an already established device capability.