How TiPolar works.

The idea that crystallographic orientation can be inferred from polarized light reflected off a polished metal surface is not new. It dates to the early history of metallography. In 1952, H. C. Vacher at the U.S. National Bureau of Standards demonstrated that polarized-light extinction positions correlated with crystal orientation in a 70 Ni–30 Cu alloy. Newton and Vacher extended the technique across additional metal systems in 1954. In 1966, Larson and Picklesimer published the first quantitative calibration of the method — a measured curve relating apparent polarization rotation to basal-pole tilt in zirconium, accurate to roughly ±3° on individual grains larger than ~5 µm.

Larson and Picklesimer also stated the scope of the technique plainly: the method extends, by the same principle, “to any other metal or alloy having an HCP or tetragonal crystal structure, such as Be, Ti, Hf, Zn, Mg, Re, Sn, and many of the rare-earth metals.”

The technique faded from practical use because it was manual, slow, and grain by grain — and because the automation of electron backscatter diffraction (EBSD) in the scanning electron microscope gave metallurgists a direct orientation measurement that applied universally to crystalline materials. The optical method's quantitative promise was there; the instrumentation to deliver it at industrial throughput was not.

What changed: digital imaging, automated polarization-state sampling, computational analysis, and a discipline of per-material calibration. TiPolar is the commercialization of that change for HCP metals, starting with alpha-titanium.

When linearly polarized light reflects from a polished, optically anisotropic crystalline surface, the reflectance for light polarized along one crystal direction is different from the reflectance for light polarized along an orthogonal direction. That differential reflectance — called diattenuation in modern Mueller-matrix polarimetry, or bireflectance in classical reflected-light metallography — is what TiPolar measures.

Diattenuation is not birefringence. Birefringence is a transmission-mode phenomenon for transparent anisotropic media: it's a phase difference between two polarization eigenmodes propagating through a material. A polished opaque metal cannot show birefringence, because the light never makes it through. What it shows instead is a difference in intensity of the reflected (or absorbed) light as a function of the polarization orientation relative to the underlying crystal — diattenuation.

For alpha-titanium (HCP), the magnitude and orientation of this diattenuation map directly to the c-axis of the crystal under the pixel. Light polarized parallel to the in-plane projection of the c-axis reflects differently than light polarized perpendicular to it. Sample the response across enough polarization states, against a calibrated model of the surface, and the c-axis orientation per pixel falls out.

The same principle applies in other HCP and tetragonal metals (zirconium, magnesium, beryllium, hafnium, zinc, rhenium, tin, many rare earths) and in many anisotropic minerals. Some materials whose bulk is not optically active produce a useful signal through a surface layer instead. Steels and brasses can be made measurable by appropriate etching that activates orientation-dependent contrast at the surface.

Diattenuation is not the only polarization signal physically on the table. Where the effect above lives in the reflection off an opaque metal, a thin, transparent surface film is a medium light actually passes through — down into the film, off the metal beneath, and back out — so an anisotropic film's dominant effect is a phase difference between the two polarization modes: birefringence, precisely the regime an opaque bulk cannot show. Some metals grow exactly such a film, and where that film is registered to the crystal beneath it, its birefringence could in principle carry information about the underlying orientation. Other systems can be etched to grow surface films that are optically active in the same way. The practical picture: for the well-studied, EBSD-validated materials we calibrate today, diattenuation dominates; for less-explored systems, birefringence through a surface film could in principle be the effect that carries the signal instead. None of these film-based routes is a calibrated or validated TiPolar capability today — they are physically available and, for the right material system, a research conversation we're glad to have.

The microscope offers two acquisition modes:

• Calibrated brightfield. A serious large-area metallographic mode in its own right — per-objective flat-field correction so vignetting doesn't survive into the output, calibrated tile positioning, and seamless montage assembly across the full stage envelope. Useful for inspection, region selection, and metallographic context, and a complete tool even when you never switch to polarimetric mode.
• TiPolar-modeTiPolar delivers light across a range of polarization states onto the sample and records how it responds to each. TiPolar-mode does exactly that — automatically — and turns the response into a measurement.. Our proprietary, patent-protected quantitative polarimetric acquisition. The optics operate near maximum extinction; the polarization state incident on the sample is rotated; the resulting per-pixel diattenuation signal is processed against a per-material calibration to produce a c-axis orientation map on an absolute scale.

How the polarization state is rotated, and the analysis pipeline that turns the diattenuation signal into orientation, are patent-protected and not disclosed here. The substance of the claim isn't the mechanism — it's the measurement's repeatability and physical fidelity, which the calibration discipline (next section) is built to demonstrate.

TiPolar measures, per pixel, the c-axis orientation of the HCP crystal under that pixel:

• Φ (declination) — the angle of the c-axis from the sample-normal direction, in the range [0°, 90°].
• φ₁ (azimuth) — the rotation of the c-axis projection in the sample plane, in the range [0°, 180°].

The Bunge-convention φ₂ — the rotation about the c-axis — is not measured. This is not a TiPolar limitation but a property of polarized-light reflection from a uniaxial crystal: the technique is insensitive to rotation about the unique axis itself. Larson and Picklesimer stated this in 1966; Jin and De Graef (Carnegie Mellon / AFRL, Materials Characterization 167, 110503, 2020) restated it in modern Mueller-matrix language. For HCP alpha-titanium, c-axis orientation alone is what governs the mechanical properties of interest — dwell-fatigue sensitivity, slip-system activation, MTR formation. The φ₂ that we don't measure is the part that doesn't matter for the use case TiPolar was built for.

In crystallographic terms, that pair (φ₁, Φ) is the first two Bunge Euler angles — the standard EBSD/texture convention — read as a Z–X–Z rotation from the sample frame to the crystal: φ₁ about the sample-normal Z, then Φ about the tilted X′. The third rotation, φ₂ about the c-axis itself, is the one polarized light can't see.

From the per-pixel (Φ, φ₁) field, the TiPolar software produces c-axis orientation maps, inverse pole figures using our own custom colormaps, pole figures, MTR identification, and texture metrics. Output is a single .tipolar file per scan — regardless of tile count — that opens as a seamless large-area montage with a complete per-frame audit trail.

TiPolar ships calibrated by us for HCP material systems. A calibration is produced by taking a representative sample of the target material, polishing it metallographically, mapping it with EBSD to obtain the ground-truth c-axis orientation per pixel, and using that mapped sample to calibrate the TiPolar response. Once calibrated, the optical measurement is on a non-relative scale — it produces orientation values, not relative contrast scores.

The validation discipline is the part that matters most. Because the calibration produces orientation on an absolute scale, the same calibration can be applied to independent samples in the same material system, and the c-axis maps produced from those independent samples can be checked against their own EBSD ground truth. When the calibration generalizes — when the c-axis map TiPolar produces on a sample it hasn't seen agrees with that sample's independently-measured EBSD c-axis map — two things are demonstrated at once: that the calibration is correct, and that TiPolar is sensitive in the same physical sense that EBSD is for c-axis measurement on that material system.

That's a stronger statement than “TiPolar agrees with EBSD.” It's the claim that TiPolar is measuring the same physics — a physical model of the material's optical response grounded in measured data, not a black-box fit to one sample's grains.

In independent peer-reviewed work, computational polarized-light microscopy placed at least 94% of alpha-titanium grains within 10° of their EBSD-measured c-axis (Jin & De Graef, Materials Characterization 167, 110503, 2020) — a research-level result, not a TiPolar specification, but direct third-party evidence that the optical method measures the same physics EBSD does.

Custom calibration of a new material system is the same process. If TiPolar doesn't yet ship a calibration for your material, that's a contract scope — see services.

TiPolar tolerates surface finishes that EBSD can't. A 1 µm diamond polish is our recommended target — best results come from a 1 µm finish — but the technique works across moderate metallographic finishes and degrades gracefully on coarser ones. There is no overnight colloidal-silica step required for first-pass results.

Preparation follows the proven three-step method Struers publishes for titanium: plane grinding on a rigid resin-bonded diamond disc, fine grinding with a 9 µm diamond suspension, then chemical-mechanical polishing with colloidal silica and hydrogen peroxide — carried on until the surface reads clean under the microscope, at which point the structure is already visible in polarized light without etching. For Ti-6Al-4V, our partners at Struers have tuned that method to bring a machined flat to a TiPolar-ready surface in roughly 20 minutes; other titanium alloys take comparable EBSD-quality prep time minus the overnight vibratory polish — typically 45–60 minutes. Independent academic work (Jin & De Graef 2020) reports the same property of the underlying technique: “a metallographic polish without vibratory polish is sufficient for CPLM data acquisition.”

For a production QC environment, the practical implication is significant. The single-largest cost in EBSD-based orientation work is often not the SEM time — it's the prep queue. TiPolar removes the bottleneck.

EBSD remains the gold standard for direct, universal crystallographic orientation measurement. It works on any crystalline material that can be properly prepared for the SEM. It measures full orientation (φ₁, Φ, φ₂). When the question is “what is the orientation in this small region, with full Euler answers, and we have time to do it,” EBSD wins.

TiPolar wins on the dimensions EBSD struggles with: area, speed, atmospheric operation, and sample-prep tolerance. When the question is “screen this forging cross-section for microtextured regions, characterize a part too large for the SEM chamber — up to a 120 × 100 mm stage that holds a whole forging cross-section or a tray of QC coupons at once — or fill the stage and inspect them all in an unattended run,” TiPolar is built for it.

The two instruments are complementary. EBSD validates TiPolar at the calibration stage and remains the appropriate tool when full orientation or universal applicability is required. TiPolar is the right tool when speed, scale, and shop-floor compatibility are what define the workflow.

Several orders of magnitude faster than EBSD on alpha-titanium, at equal pixel size and area, in air — the trade the two instruments actually make, side by side: