Week 5 — Oct 12: Molecular Spectroscopy & Solid-Form Characterization

(Lecture 5.) Last week used electronic absorption to count molecules and to measure performance. This week uses vibrations — bonds stretching and bending at characteristic frequencies — to identify molecules and, with the solid-state techniques, to tell one crystal form from another. Both matter for the same reason: the solid form a patient swallows determines how it dissolves. These are also the techniques that make PAT possible.

The one idea

IR and Raman see the same molecular vibrations through opposite rules — a vibration shows in IR if it changes the dipole moment, in Raman if it changes the polarisability. They are two halves of one picture. The solid-state techniques add a second axis entirely: how the identical molecules are packed, which changes dissolution and bioavailability without changing a single bond.

Near-IR is the outlier: it sees only faint overtones and combination bands, broad and overlapping, carrying real information that no human can read off the plot — which is why NIR and chemometrics grew up together.

The vibrational and fluorescence techniques

The solid-state toolbox

The API is the same molecule in every form; what changes is the crystal lattice — and with it the solubility, dissolution rate, stability, and manufacturability.

The recurring teaching point: amorphous content. A small amorphous fraction in a mostly-crystalline API is more soluble, less stable, and often invisible to XRPD below ~5%; DVS, ssNMR, or microcalorimetry catch it. It is the classic hidden variable behind a batch that suddenly fails dissolution or grows a new degradant.

What all of this is used for

Making the result trustworthy

For an identity test: the reference spectrum or diffractogram, matched sampling (ATR vs transmission; XRPD sample-height and preferred-orientation effects), a documented acceptance criterion (a correlation threshold or specified band/peak positions, not “looks the same”), and awareness that a polymorph difference can fail an over-tight identity test for the right reason.

For a quantitative NIR or Raman method you are validating a model, not just an instrument:

• The calibration set must span every source of variation the method will meet — concentration, particle size, moisture, supplier, temperature — or real-sample predictions are extrapolation.
• A reference method (HPLC assay, Karl Fischer, LOD) supplies the truth values; the NIR method is no better than its reference.
• Preprocessing (SNV, MSC, derivatives — Week 10) is part of the method and locked with it.
• Validation per Q2 plus NIR-specific guidance, and a model lifecycle (Q14): monitoring, an update procedure, defined change control.

Compendial basis: USP ⟨854⟩ (MIR), ⟨858⟩ (Raman), ⟨1119⟩ (NIR), ⟨941⟩ (XRPD), ⟨891⟩ (thermal analysis); Ph. Eur. 2.2.24 / 2.2.48 / 2.2.40 / 2.9.33 / 2.2.34.

Worked case — measuring through the container

Incoming raw materials have traditionally been sampled: open a fraction of the drums, take material to the lab, run an identity test, quarantine the rest until the result comes back. Every opened container is a contamination risk and a delay, and the unopened ones are never tested at all.

Transmission Raman and spatially-offset Raman (SORS) measure the bulk contents of a sealed bottle, sachet, or blister through the packaging, in seconds, non-destructively — making 100% verification of incoming containers practical, none opened. It is a clean illustration of a theme from Week 1’s quality-control section: move the control to where the risk is, and test everything rather than infer from a sample. It also raises the data-integrity question the automation weeks return to — thousands of automated spectral pass/fail calls a day, each a GMP record.

Risk-assessment assignment

Take an in-line NIR blend-uniformity method and build the method FMEA against its analytical target profile (Week 2 approach). Weight the failure modes a spectroscopic model adds over a wet method: an out-of-calibration-range sample, probe-window fouling, feed-material drift, a stale model. Score their detectability — which would the running system actually catch?

Where the analyst sits

A spectrum takes seconds to acquire and can take a career to interpret responsibly. The instrument always returns a number or a “match”; the judgment is whether the match means what it appears to — a real form change or a sampling artefact, an NIR result inside the model’s world or outside it, an amorphous fraction that XRPD is quietly missing. That is the STEAM “A”, and for the process-facing methods it runs thousands of times an hour with nobody watching each call. The refrain: science → evidence → reduced uncertainty → control → regulatory confidence → patient trust.

For discussion

• A drug substance passes HPLC assay and impurities but fails its IR identity test. Give two innocent explanations and two that would stop the batch.
• XRPD shows a single crystalline phase. What could still be wrong with the solid form, and which technique would you reach for?
• Why did NIR spectroscopy and chemometrics develop together, when mid-IR and Raman are often used with simple band-position criteria?
• Your in-line Raman blend method was calibrated in winter. Blend results trend high in July. List the physical causes and how you would tell them apart.
• A batch of API passes every release test but tablets made from it fail dissolution. Walk through the solid-state investigation.
• When is a polymorph difference something an identity test should catch, and when is it something the test should be designed to tolerate?

Source note. Vibrational spectroscopy follows standard texts (Skoog; Smith, Modern Raman Spectroscopy); solid-state characterisation follows Brittain, Polymorphism in Pharmaceutical Solids. Compendial: USP ⟨854⟩/⟨858⟩/⟨1119⟩/⟨941⟩/⟨891⟩, Ph. Eur. equivalents, plus EMA/FDA NIR guidance. Through-container Raman follows the transmission-Raman / SORS literature (Matousek et al.). Model-lifecycle expectations connect to Q14 and the PAT week. (Instructor: confirm current ⟨858⟩/⟨941⟩ and NIR-guidance status; keep the method-FMEA assignment structurally identical to the Weeks 2 and 4 assignments so students see it three times.)

• ←4 · Oct 5
• 6 · Oct 19→