Infrared, Raman, and Photoluminescence Analysis

FTIR, Raman, and PL are often mentioned together because they are spectroscopic methods. But they do not pose the same question or measure the same physical phenomenon. Their greatest value emerges precisely when their functions are not conflated.

Three methods—three principal questions

FTIR: which bulk impurity and defect groups are present, and to which diamond type does the sample belong?

Raman: which crystalline or molecular phase is present at the measurement location?

PL: which optically active defect centers emit light after excitation, and what is their spectroscopic context?

[VISUAL 70.1: FTIR vs. Raman vs. PL—bulk type / local phase / defect emission]

FTIR: diamond type and the bulk impurity picture

Diamond’s infrared absorption contains both intrinsic lattice features and impurity-related features. Different nitrogen configurations are particularly important in the one-phonon region, while boron produces its own characteristic absorptions.

On this basis, diamonds are classified as:

• type Ia — aggregated nitrogen;
• type Ib — predominantly isolated substitutional nitrogen;
• type IIa — no nitrogen measurable by the routine IR criterion for type I;
• type IIb — boron as an important impurity and possible electrical conductivity.

FTIR may also reveal hydrogen-related, platelet, amber, and other features, depending on the sample.

Diamond type is a filter, not an origin verdict

Diamond type is extremely useful because it narrows the possibilities statistically and physically. The great majority of natural colorless diamonds belong to the type Ia population, while many commercial colorless laboratory-grown diamonds are type II.

But natural type IIa and IIb diamonds exist, natural type Ib is rare, and laboratory-grown material may exhibit different types and mixed configurations. Even unusual CVD examples with aggregated nitrogen have been published.

Therefore:

type IIa ≠ CVD

type Ib ≠ HPHT

type Ia ≠ automatically natural

FTIR is a powerful filter and part of the body of evidence, not a standalone verdict on origin.

Nitrogen aggregation as a record of time and temperature

Given sufficient time and appropriate temperatures, isolated nitrogen can aggregate into more complex configurations. The relationship among C, A, and B forms of nitrogen may therefore contain information about thermal history.

But this information does not provide a simple “clock” without an additional model. Concentration, time, temperature, multiphase growth, and subsequent treatment can alter the interpretation.

Raman: phase identity, not diamond origin

Raman spectroscopy measures the inelastic scattering of light, and the result is typically displayed as Raman shift in cm⁻¹.

Crystalline diamond has a highly characteristic first-order Raman feature at approximately 1332 cm⁻¹. It is extremely useful for confirming the diamond phase and as a reference feature in many measurements.

But 1332 cm⁻¹ does not mean “natural diamond.” Natural, HPHT-grown, and CVD-grown diamonds all consist of the diamond phase.

[VISUAL 70.2: Raman—the 1332 cm⁻¹ host-diamond feature and a local spectrum of a mineral inclusion]

Local identification of inclusions and surface phases

One of Raman’s greatest gemological advantages is the ability to focus on a small area. This can distinguish among:

• the host diamond;
• a mineral inclusion;
• graphite or another carbon material;
• a surface deposit;
• a particular component within a fracture or on the surface, if the geometry and signal are sufficiently favorable.

Raman phase identification does not automatically resolve its geological relationship with the diamond. If Raman confirms that an inclusion is garnet, chemistry, texture, and context are still required for a petrogenetic conclusion.

PL: sensitivity to defect centers

Photoluminescence occurs when optical excitation brings electrons to an excited state and their return produces light emission. In diamond, PL can be extremely sensitive to trace defect centers below the detection limit of routine FTIR absorption.

Examples include:

• NV⁰ at approximately 575 nm;
• NV⁻ at approximately 637 nm;
• SiV⁻ at approximately 737 nm;
• numerous nickel-related and other centers.

These values are useful as reference points, but a peak must never be interpreted separately from the excitation wavelength, temperature, charge state, phonon sideband, instrument, and remainder of the spectrum.

Why PL is not a “barcode origin detector”

SiV is often important in CVD identification because silicon may be incorporated during growth from the reactor environment. More recent GIA literature, however, emphasizes two limitations: SiV is not present in every CVD diamond and may occasionally occur in natural diamonds.

Current evidence is especially important here. In GIA’s ten-year CVD dataset analyzed in 2024, the average normalized SiV⁻ PL signal decreased by approximately two orders of magnitude, while the number of CVD diamonds without a detectable SiV⁻ peak also increased. This is not a universal rate for global production, but it demonstrates how a marker that was once extremely practical can gradually lose standalone diagnostic power as growth technology changes.

Similarly, NV centers occur in different natural, laboratory-grown, and treatment contexts.

Therefore:

SiV ≠ CVD, always

NV ≠ treatment, always

PL is powerful precisely because it is interpreted as a multipeak, condition-dependent fingerprint.

Excitation, detector range, and temperature

A PL spectrum exists only in relation to measurement conditions.

The excitation wavelength determines which centers are efficiently excited. The detector range determines what the instrument can detect at all. Temperature changes line widths, the relationship among emission pathways, and the visibility of weak features.

Low-temperature PL often greatly increases diagnostic resolution, but it is not automatically necessary for every sample. The professional record must contain the conditions under which the spectrum was obtained.

Peak intensity is not concentration without a model

PL peak intensity depends on far more than the number of defects:

• laser power and wavelength;
• absorption of the excitation;
• focus and sampling volume;
• collection geometry;
• temperature;
• quenching;
• charge state;
• other competing emission centers;
• detector and instrumental response.

The ratio of two peaks can therefore be used as a diagnostic criterion only when it has been validated for a defined population and conditions.

Raman and PL can share the same optical system

A laser-excited spectrum may contain both Raman and luminescence features. This is useful, but it also creates interpretive pitfalls. A strong fluorescence background can obscure a Raman feature; changing the laser can radically alter the PL picture.

A spectrum is therefore interpreted not only by peak position but also by which physical process produced the signal.

Bulk versus local problems

In routine gemology, FTIR often behaves as a relatively bulk method: it provides integrated information along the optical path. Raman and confocal PL can examine a particular zone much more locally.

This difference becomes critical in:

• a zoned diamond;
• a CVD multi-run structure;
• a small seed;
• a surface layer;
• an inclusion;
• a treatment that is not uniformly distributed.

“FTIR indicates type IIa” and “PL reveals a particular center in a small zone” are not contradictory results if the methods sample different volumes.

[VISUAL 70.3: Sampling volume—FTIR bulk path versus a local Raman/PL spot]

Calibration, reference, and library discipline

Spectroscopy is not merely a matter of overlaying a sample spectrum on an image from the internet. It requires:

• a validly calibrated axis;
• known instrumental resolution;
• an appropriate reference;
• a record of temperature and excitation;
• control of the laser and focusing;
• a relevant reference library;
• awareness that a new generation of laboratory-grown material may fall outside historical patterns.

A reference library is a living evidentiary resource, not an immutable atlas.

Multimethod workflow

For an unknown colorless diamond:

1. confirm that the material is diamond;
2. use FTIR to determine diamond type;
3. do not automatically declare a type II or other suspicious context laboratory-grown;
4. add PL and/or luminescence imaging according to the question;
5. integrate microscopy and other data.

For a microscopic inclusion:

1. describe its morphology;
2. use Raman to identify the phase if the geometry permits;
3. leave the petrogenetic conclusion to mineralogy and the broader context.

For a suspected treatment:

1. define which treatment scenario is physically possible;
2. use FTIR/PL/UV-Vis-NIR according to its markers;
3. add microscopy and imaging when the spatial relationship carries information;
4. do not reconstruct the treatment history in more detail than the data permit.

[VISUAL 70.4: Multimethod workflow—FTIR + Raman + PL + microscopy/imaging]

Table of functional distinctions

Method	Primary question	Typical strength	Principal limitation
FTIR	bulk impurities and diamond type	nitrogen/boron/hydrogen/defect context	type is not an origin verdict
Raman	which phase is present at the measurement location?	local material fingerprint	the diamond peak does not distinguish natural/LGD
PL	which optical centers emit?	very high sensitivity	the signal depends strongly on conditions and is not a one-peak barcode

Three levels of inference

For every spectroscopic method, separate:

Direct measurement: peak, band, baseline, intensity, position.

Physical interpretation: probable defect, impurity, or material phase.

Historical diagnosis: natural/laboratory-grown origin, treatment, or growth history.

Each successive level requires more context than the one before it. Most spectroscopic errors arise when interpretation jumps directly from the first to the third.

Chapter summary

• FTIR, Raman, and PL answer different questions.
• FTIR is key to diamond type and the bulk impurity/defect context.
• The type Ia/Ib/IIa/IIb classification is not a standalone natural/LGD verdict.
• Nitrogen aggregation carries thermal information, but it is not a simple clock.
• Diamond’s Raman feature at approximately 1332 cm⁻¹ confirms phase, not geological origin.
• Raman is particularly valuable for local identification of inclusions and surface phases.
• PL can detect very small concentrations of optically active centers.
• NV, SiV, and nickel-related centers must be interpreted contextually.
• SiV is not present in every CVD diamond, nor is it absolutely absent from natural diamonds.
• Excitation wavelength, detector range, and temperature are integral parts of a PL result.
• Peak intensity is not concentration without an appropriate model and control of conditions.
• The strongest origin and treatment conclusions arise from a combination of multiple complementary methods.