Ultra Rare Gemstone · Science – Gemmology Scientific edition

Rhodizite

From rare-element pegmatites to the londonite–rhodizite series: formation, structure, colour and gemmological interpretation.

Rhodizite is studied within the context of pegmatites and the rhodizite–londonite series. Structure, properties, deposits and identification methods inform observations of the specimens presented here.

Rare light, a deep history

A small yellow crystal can encapsulate a long history of chemical concentration. Rhodizite belongs to the beryllium aluminium borates; its crystals combine cubic architecture, high hardness and very limited gem-quality occurrence. Madagascar plays a central role in this story.

Its interest extends beyond the rarity of a cut stone. Together with londonite, it provides a remarkable example of alkali substitution within one structural family. The eye encounters colour and lustre; the mineralogist investigates atomic-site occupants, growth conditions and the reactions that made crystallisation possible.

Three scales guide this interpretation: the granitic massif, which supplies and concentrates elements; the pegmatite, where assemblages evolve; and the crystal, whose chemistry and defects record part of that evolution. Gemmology then relates this information to transparency, colour and the stone's behaviour in light.


Typical rough crystal · computer-generated illustration.
Typical rough crystal · computer-generated illustration.

Rhodizite 1.55 carat · original photograph.
Rhodizite 1.55 carat · original photograph.

Crystallography and the rhodizite–londonite series

A cubic framework with several occupants

The ideal end-member formulae are KAl₄Be₄(B₁₁Be)O₂₈ for rhodizite and CsAl₄Be₄(B₁₁Be)O₂₈ for londonite. This notation shows four beryllium atoms on one set of sites and an additional beryllium atom associated with the eleven boron atoms. It does not represent an analysis of every natural crystal.

The symmetry is cubic, space group P4̅3m, no. 215 , class 4̅3m, primitive cubic Bravais lattice. The refinement by Pring and colleagues gives a = 7,318(1) Å for their sample. Atomic positions from that work are used in the model below. Partial occupancies of the alkali site and the shared B–Be site are retained in the model data. [1]

Boron- or beryllium-bearing tetrahedra connect groups of AlO₆ octahedra; alkalis occupy cavities within this arrangement. The framework remains three-dimensional: an external crystal face represents neither an isolated layer nor the shape of an ion. [2]

Reference framework Rhodizite
Ideal end member KAl₄Be₄(B₁₁Be)O₂₈
System / Bravais lattice Cubic · primitive P
Class / space group 4̅3m · P4̅3m (no. 215)
Reference unit cell a = 7,318(1) Å · α = β = γ = 90° · Z = 1
Strunz / Dana6.GC.05 / 25.08.02.01
Coordinate reference Pring et al., 1986 · AMCSD 0014483

Primitive unit cell · 49 sites


Primitive unit cell · 49 sites

Local coordination · AlO₆


Local coordination · AlO₆

Local coordination · BeO₄


Local coordination · BeO₄

Reference-based model: 49 geometric positions in the primitive conventional unit cell, including an alkali site and mixed-occupancy sites. Spheres represent sites, not 49 fully occupied atoms. Graphic radii are not proportional to ionic radii.

The continuous series: what changes and what remains

The substitution K⁺ ↔ Cs⁺ is isovalent: one monovalent ion replaces another. In an ideal binary section, rhodizite lies on the K-dominant side and londonite on the Cs-dominant side. A change in species name therefore does not require a visible change in colour or habit.

Here we define X Cs = n(Cs)/[n(K) + n(Cs)]. The boundary X Cs = 0,50 applies to this two-component projection. The quantities n are atomic amounts: directly comparing weight percentages of K₂O and Cs₂O would give an incorrect criterion because their molar masses differ greatly.


K–Cs substitution series: atomic proportions, not a colour range.
K–Cs substitution series: atomic proportions, not a colour range.

A substitution series is not a colour range. Two equally yellow crystals may have different compositions; conversely, domains within a single crystal may differ chemically. A continuous diagram does not in itself demonstrate complete miscibility at every temperature or the existence of all pure end members in nature.

Beyond the two-alkali model

Analyses of natural material also reveal Rb, Na, Li, vacancies and coupled substitutions. Pekov and colleagues showed that historical Ural specimens were Cs-dominant and should be classified as londonite. Their study also describes crystal chemistry more complex than a binary substitution. [2]

The mineral touretite , the lithium analogue with ideal formula LiAl₄Be₄(B₁₁Be)O₂₈, was approved by the IMA and announced in 2024 from a pegmatite at Ambalabe, Madagascar. This isostructural species expands the group; it is not a new name for rhodizite. The K–Cs diagram remains a useful educational section within this broader diversity. [3]

Classification: three-dimensional framework borates, tektododecaborates; Strunz 6.GC.05, Dana 25.08.02.01. These codes describe mineralogical classification, not gem-quality grades.

From oxide analysis to species boundaries

Dominance is calculated from atomic amounts. For an analysis expressed as oxide weight percentages, the relative potassium amount is proportional to 2w(K₂O)/M(K₂O), and the caesium amount to 2w(Cs₂O)/M(Cs₂O). The molar masses are approximately 94,20 and 281,81 g·mol⁻¹. Equal oxide masses therefore supply almost three times fewer Cs atoms than K atoms.

In Pring's published example, occupancies K₀,₄₆ and Cs₀,₃₆ give X Cs ≈ 0,439 when only these two alkalis are projected. Rb, Na and vacancies nevertheless remain essential to a complete description of the site. The binary projection illustrates substitution; the structural formula allows the actual crystal to be discussed.

Chemical continuity and zoning

A solid solution preserves the same structural arrangement while its occupants vary. Retaining this architecture does not mean that every property varies linearly with X Cs . Unit-cell parameter, density and vibrations also depend on occupancies and defects. A straight line joining the end members would be a modelling assumption, not an analytical calibration.

In a zoned crystal, analysis points distributed from core to rim relate growth sectors to composition. Here, ‘continuous series’ describes substitution and documented intermediate compositions; it is not a complete experimental diagram of temperature-dependent equilibria. Classifying a stone requires representative analysis and consideration of its heterogeneity.

Geological formation and morphology

Primary vein deposits: Antsongombato

The source material is a differentiated melt of granitic composition, whose rare-element constituents are distributed among pegmatitic assemblages. At Antsongombato, veins cut dolomitic marbles of the Itremo group. Field descriptions distinguish the metamorphic host, the intrusive body and the domains concentrating crystals: here, the gem-bearing rock is pegmatite, not a transported deposit.

The productive vein described is approximately 400 metres long and 10 to 60 centimetres thick. Rhodizite–londonite is concentrated in coarse-grained columns following the dip; some domains associate the crystals with fine-grained pollucite and spodumene. Quartz, feldspars and tourmalines form part of the described assemblage. This distribution explains why rich mineralisation can remain confined to a small portion of the vein. [11]

Internal zones and pegmatite cavities: Antandrokomby and associated localities

The description of londonite places Antandrokomby, Ampanivana and Antsongombato within highly evolved granitic pegmatites rich in red tourmaline. Minerals of the series occur in internal zones and miarolitic cavities, where growth can develop free crystal faces. In contrast, a crystal enclosed within the assemblage retains constraints imposed by contact with neighbouring minerals.

The association notably includes feldspars, quartz, tourmaline, spodumene, danburite and rare Be, Nb and Ta minerals. It indicates the system's chemical specialisation without guaranteeing an identical assemblage around every crystal. Antandrokomby lies in the Manandona valley and is distinct from Antsongombato. An occurrence of the series does not automatically imply production of transparent gems. [8]

Reaction domains and replacement: the Manjaka example

At Manjaka, studying spodumene alteration products offers another perspective on pegmatite evolution. Observations concern mineral assemblages and contacts in a body hosted by calc-silicate rocks. Relationships with boron- and beryllium-bearing phases help explain residual-melt evolution and local exchanges.

These reaction domains belong to the primary deposit. They do not represent a second independent origin that could be inferred from a cut stone's colour. The details of this interpretation and the role of water activity are discussed below, distinguishing crystallisation from replacement. [7]

Weathering and transport: distinguishing release from crystallisation

Weathering of a pegmatite may release a crystal; subsequent transport does not change the rock in which it crystallised. The examples documented here concern primary deposits. Attribution to a colluvial or alluvial concentration would require separate field documentation, distinct from that of the vein.

Concentrating dispersed elements

A granitic pegmatite commonly combines quartz, feldspars and micas. Rare-element varieties develop more specialised assemblages. The LCT family is characterised by lithium, caesium and tantalum enrichment, accompanied in some systems by rubidium, beryllium and boron. This is a geochemical family: not every member necessarily contains rhodizite. [4]

During crystallisation, minerals do not incorporate every element in the same proportions. An element poorly incorporated by early phases may accumulate in the remaining melt. Joint availability of B, Be, Al and alkalis is necessary for rhodizite formation, but is not sufficient. Competing-phase stability, melt composition and exchanges with the host rock determine which species actually form.

A calculation to explain fractionation

The Rayleigh model is written C L/C₀ = FD−1 , where F is the remaining melt fraction and D the bulk solid/melt partition coefficient. In the purely educational example D = 0,1 and F = 0,1, the melt is enriched by approximately 7,9 times relative to its initial concentration. This calculation assumes a closed system and effective crystal separation. It is neither an analysis of Antsongombato nor an estimate of the gem's composition.

Caesium lends itself to this interpretation of fractionation, but D changes as the crystallising assemblage evolves. The appearance of a phase that readily incorporates Cs changes its accumulation in the melt. A petrogenetic model must therefore connect mass balance with a mineralogical sequence, rather than extend a single curve indefinitely. [5]


Formation factors in the rhodizite–londonite series: differentiation, assemblages, growth and interfaces.
Factors contributing to formation of the series; arrows do not imply a universal sequence.

Why large crystals do not prove slow cooling

Pegmatitic texture also depends on kinetics. London's experimental work and models emphasise undercooling, nucleation, diffusion and a boundary melt enriched in constituents rejected by growing crystals. A small number of efficiently supplied nuclei can produce large crystals; their size does not directly establish cooling duration. [6]

Internal zoning may thus result from the advancing crystallisation front and chemical evolution of the adjacent melt. Water and fluxing elements influence transport and crystallisation conditions, but a large pegmatite is not necessarily a cavity entirely filled with aqueous fluid. Silicate melt, exsolved fluid and later transformations must be distinguished.

The role of host rock and late reactions

At Manjaka in the Sahatany valley, Novák and colleagues describe spodumene alteration products, including boralsilite, in a pegmatite hosted by calc-silicate rocks. They propose evolution involving a residual melt rich in Na, K, B and Be, spatially and temporally associated with rhodizite–londonite assemblages. Their interpretation emphasises the local importance of low water activity. [7]

This case clarifies the geological reasoning: two pegmatites containing boron and beryllium may develop different associations depending on host-rock reactions. ‘Late mineral’ indicates a position within a sequence, not a universal temperature. No thermobarometric information is available here for the collection specimen.

Saturation: a question of chemical activities

High boron concentration does not automatically cause a borate to crystallise. Boron may remain in the melt or enter tourmalines; beryllium may enter beryl, phenakite or other phases. Rhodizite appears when the combination of its constituents becomes thermodynamically favourable relative to these competing reservoirs. Chemical activities, rather than bulk-rock concentrations alone, control this competition.

This distinction is essential in a zoned pegmatite: bulk sampling can dilute the signal of a reaction domain only a few centimetres across. A rhodizite crystal reflects local conditions at its growth interface. Contacts with feldspars, tourmalines and beryllium-bearing phases must be studied at this scale before a formation reaction is proposed.

From residual melt to replacement reactions

Two processes may follow one another or overlap. Magmatic crystallisation removes elements from the melt and enriches incompatible constituents in the residue. Replacement subsequently dissolves an existing phase and precipitates another, with material transfer. Replacement textures are recognised through relics, irregular fronts and cross-cutting relationships; they cannot be inferred simply from the presence of a rare mineral.

The sequence of events is reconstructed through petrography: a phase enclosed in another is generally older than the enclosing portion, whereas a veinlet cutting it is younger. Exceptions, such as recrystallisation, require several observations to be combined. Chemistry interpreted within these relationships helps establish whether K–Cs variation accompanied growth or later transformation.

Pressure, temperature and the limits of reconstruction

Emplacement depth, water content and melt composition shift the solidus and stability fields. A temperature range borrowed from another pegmatite therefore cannot be assigned to every rhodizite deposit. Well-characterised fluid inclusions, assemblage equilibria and appropriate experiments can constrain the trajectory; no pressure or temperature value specific to the gem presented here is available.

This approach distinguishes the robust scenario — differentiation, local concentration, crystallisation and possible reworking — from a quantified history requiring measurements. The formation diagram presents relationships between these processes without turning an educational section into a map of an actual mine.

Forms governed by growth rate

Crystals are generally equant, with dodecahedral {110} faces and modifying forms. The description of londonite records {221}, {211}, {111} and, more rarely, {100}. These indices denote crystallographic orientations. The six models below are idealised combinations of crystal forms; their synthetic textures suggest growth irregularities without reproducing six natural specimens. [8]

Dodecahedron {110}


Dodecahedron {110}

Modified dodecahedron


Modified dodecahedron

Combination {110}–{211}

Combination {110}–{211}

Modification {111}

Modification {111}

Developed {221} faces


Developed {221} faces

Complex combination


Complex combination

Six theoretical morphologies in the rhodizite–londonite series. Forms {110}, {111}, {100}, {211} and {221} are documented in the series; their relative development is idealised here. External form does not distinguish the K and Cs end-members. Sources: Handbook of Mineralogy, Rhodizite and Londonite.

Physical and optical properties

The values below are published reference points for the species or gemmological series; they do not replace measurements on an individual stone. [9]

Refractive indices n ≈ 1,689–1,693 · published variation
Birefringence Ideally zero; local anomalies possible
Optic character and sign Isotropic; no uniaxial or biaxial optic sign
2V angle Not applicable
Pleochroism Absent in the isotropic model
Dispersion 0,018 (B–G), published value reported by Arem and cited by Laurs et al.; not measured in this study
Specific gravity Approximately 3,22–3,44; studied gem material: 3,34–3,42
Hardness 8–8,5 on the Mohs scale
Cleavage / parting Indistinct cleavage reported; this should not be taken to mean universally absent cleavage
Fracture / tenacity Conchoidal to uneven; brittle
Lustre Vitreous to subadamantine

Refraction, reflection and transparency

At an air–gem interface, refraction bends light according to Snell's law. Inside, certain angles of incidence produce total internal reflection. For an illustrative refractive index n = 1,69, the gem–air critical angle is arcsin(1/n), approximately 36,3°. Cutting must organise these paths: a light area seen through the table may depend on facet angles and lighting rather than being an inclusion.

As the stone rotates, the facets reflecting the illuminated surroundings towards the observer change. Moving flashes can create an impression of fluidity. The historical term ‘water’ for gemstones refers to their limpid appearance and transparency; it does not establish the presence of mobile water. [10]

Hardness measures scratch resistance. Toughness concerns resistance to fracture: a hard stone may be brittle. Setting design should therefore protect edges and avoid concentrated stress.

Colour and fluorescence

Yellow and transmitted light

In a yellow gem, the spectrum reaching the eye contains relatively less of certain violet and blue components. Perception depends on lighting, the path through the stone, cutting and visual adaptation. Describing this spectral balance does not identify the absorbing centre: a trace element, its charge state and crystal environment must be characterised before a precise attribution can be made.

Material in the series notably ranges from colourless to yellow and yellow-green. Colour changes under irradiation and light exposure are documented. A more intense hue therefore proves neither higher Cs content nor better stability.

Emission under ultraviolet excitation

Fluorescence is light emission following energy absorption. It adds to reflection and transmission; it is not simply a brighter version of daylight colour. The lamp wavelength must be specified: short-wave UV at 254 nm and long-wave UV at 365 nm may excite the centres present differently.

22.62 ct · pear · lagoon-blue fluorescence. Original Ultra Rare photograph under UV.
Long-wave UV · 365 nm. Pale blue fluorescence observed by Ultra Rare; original photograph.

On the left, the photograph shows the 22.62 carat pear under UV; on the right, the response at 365 nm belongs to the 1.55 carat cushion. It is not a universal rhodizite signature and does not identify a particular emission activator. UV photographs also reflect lamp, exposure and sensor effects; photoluminescence measurement would be needed to locate emission maxima.

Inclusions as a record of growth

Internal discontinuities provide information on growth, strain and possible late fluid circulation. Gemmological literature notably describes partially healed fractures and zoning; a polariscope may also reveal anomalously birefringent domains. These observation categories must remain distinct: optical contrast does not always indicate a foreign body.


Computer-generated illustrations of documented internal structures; none is a photograph of the gem offered for sale.
Computer-generated illustrations of documented internal structures; none is a photograph of the gem offered for sale.

Reading an image without overinterpreting it

A healed fracture may retain a network of small cavities; an angular surface may also be a fracture plane or growth interface. Identifying a solid inclusion requires more than its outline: relief, optical behaviour, Raman spectra and sometimes chemical analysis provide complementary evidence. A known pegmatite assemblage does not prove that every constituent mineral is included in every gem.


Original Ultra Rare photomicrograph of the 22.62 ct pear-shaped rhodizite · High magnification · internal interfaces
Rhodizite · 22.62 ct
High magnification · internal interfaces. Original Ultra Rare photograph.

Original Ultra Rare photomicrograph of the 22.62 ct pear-shaped rhodizite · Macro view · internal architecture
Rhodizite · 22.62 ct
Macro view · internal architecture. Original Ultra Rare photograph.

The photomicrographs above show angular forms and strongly contrasting interfaces. They invite examination of the specimen's internal geometry; the image alone cannot identify an included species or establish a liquid inclusion. They do not depict the eye clean 1.55 carat rhodizite.

Deposits – Atlas

Mineralogical occurrences and sources of cuttable stones answer different questions. A country may have yielded a collector's crystal without sustaining regular gem production. This atlas locates the major documented regions for the series; colours express their documentary role, not extraction volumes.

Rhodizite · Deposits – Atlas

Location key

Madagascar Gemmological reference source for the rhodizite–londonite series.

Urals · Russia Historical rhodizite occurrences.

Wisconsin · United States; Devon · United Kingdom Other documented occurrences, without equivalent gem production.

Understanding colours and symbols Countries and documented locations. Blue intensity indicates qualitative significance, without measured volumes; not every point marks an active mine.

Non-exhaustive reference atlas · backgrounds and sources retained from the study. An occurrence establishes neither current mining nor a stone's provenance.
Region Published context
Central Madagascar Pegmatites of the Betafo–Antsirabe region and Sahatany; a major reference source for crystals and gems of the series.
Urals, Russia Historical occurrences; some specimens formerly labelled rhodizite are classified as londonite following analysis.
Devon, United Kingdom Listed historical occurrences; their inclusion does not imply current gem production.
Wisconsin, United States Published mineralogical occurrence; does not imply availability of gem-quality material.

A world map of the series preserves the history of identifications while retaining the modern distinction between species. Historical localities are not presented as currently active mines. [9] [2]

Madagascar: a natural laboratory for pegmatites

Madagascar · geology and documented areas


Map of rhodizite–londonite pegmatites in the Betafo–Antsirabe region, including Antsongombato, figure 3 in Laurs et al. 2002

Betafo–Antsirabe · rhodizite–londonite pegmatites. Laurs et al. (2002), figure 3, © GIA. Locations in the publication concern the mineral series; none is assigned to the Ultra Rare specimens. [11]

Location key

Betafo · Antsirabe Central pegmatites; Antsongombato area.

Antandrokomby · Manjaka Separate localities in central Madagascar; Antandrokomby lies in the Manandona valley.

Geological background and symbols Colours: basement domains, cover rocks and volcanism. Halos: indicative areas; generalised boundaries. The study distinguishes deposits, occurrences and geographical reference points.

Cartographic sources and credits: references retained in the Sources section. No precise site is assigned to a specimen without documentation.

Ancient host rocks, younger intrusions

Central Madagascar juxtaposes metamorphic terrains and granitic intrusions. A pegmatite is younger than the rock it cuts; its age therefore cannot automatically be inferred from basement age. Mapping distinguishes the regional setting, host rocks and pegmatite bodies, often too small to appear at national scale.

Antsongombato and the Betafo region

Antsongombato was known before 1998. That date marks the recognition of promising new bodies in the modern history of mining. The 2002 gemmological publication describes pegmatites within a metamorphic assemblage notably including gneisses, quartzites, schists and marbles. It reports approximately 300 carats of faceted stones for the period studied: this is a historical, local figure, not cumulative worldwide production. [11]

Antandrokomby and the Manandona valley

Antandrokomby is the locality from which londonite was described. It must be distinguished from Antsongombato rather than combining every name into a single mine. The described association includes feldspars, quartz, tourmalines and various minerals characteristic of pegmatite evolution. Species diversity reflects several concentration mechanisms and stages of crystallisation. [8]

Manjaka: reactions recorded in textures

Manjaka demonstrates why petrography matters as much as a mineral inventory. Fine alteration products develop near spodumene and within feldspars. Their arrangement helps establish a reaction sequence and residual-melt properties. A mineral list without contacts and replacement relationships would lose this information. [7]

Connecting rock and crystal

To reconstruct a rhodizite's history, the ideal protocol combines position within the vein, host-rock description, association sequence, chemical mapping and structural characterisation. K–Cs variations can then be compared with growth zones. Crystal composition in isolation is neither an absolute clock nor a geographical address.

The rhodizite weighing 1.55 carat presented by Ultra Rare is stated to come from Madagascar. No precise district is assigned to this stone: the localities discussed provide national mineralogical context.

Antsongombato: concentration at vein scale

Laurs and colleagues describe veins in dolomitic marbles of the Itremo group. The productive body studied reaches approximately 400 metres in length but only 10 to 60 centimetres in thickness. Crystals concentrate in evolved domains; many are enclosed within the assemblage rather than free in a large pocket. Pollucite and spodumene accompany some of the best domains. These observations describe the deposit studied in 2002, not its current mining status. [11]

An informative association, without imposing a single reaction

Pollucite is a caesium reservoir; spodumene indicates lithium specialisation. Their association with rhodizite–londonite clarifies alkali distribution between phases. It does not mean that all pollucite transforms into londonite or that Cs follows only one phase. The balance must include feldspars and other minerals, their abundances, compositions and growth order.

Contact between silicate melt and marble creates compositional gradients and pathways for exchange. Calcium, magnesium, silica and volatile constituents may participate in local reactions. Describing these gradients allows testable hypotheses; writing a specific balanced reaction would require identification and analysis of the phases actually in contact.

Treatments and colour preservation

Irradiation intended to intensify yellow is documented for material in the series. Experiments have also shown colour fading in light, in irradiated material and some naturally coloured samples. Behaviour within an experimental series must not be converted into a universal fading rate.

Absence of indications of treatment and future stability are different pieces of information. A laboratory conclusion concerns the examinations performed; it is not a prolonged photostability test. For a collector's gem, avoiding extended sunlight and UV exposure and strong heat is consistent with this literature. Stability should not be tested through deliberate exposure that might alter the stone.

Gentle cleaning with lukewarm soapy water, followed by careful drying, limits stress. Impact, thermal changes and jewellery-making treatments should be discussed with the workshop. Eye clean clarity does not remove the need to protect the crystal during setting.

Rarity and collecting criteria

Occurrence

Rhodizite requires an unusual combination: a sufficiently differentiated pegmatitic environment, available constituents, a favourable assemblage and preserved crystals. A species' geological rarity cannot be measured solely by the carats visible on the market.

Known occurrences, extracted volumes and available gems do not share the same denominator. A quantitative geological comparison would require comparable global inventories, which are not assembled here.

Rarity within rarity

Selection narrows when documented mineral identity, attractive colour, eye clean clarity, well-proportioned cutting and a weight above one carat are required together. Each criterion excludes some initial material. They are not independent: selecting a cleaner area may substantially reduce the final cut weight.

A fine gem of this family is therefore an object of knowledge as much as adornment. Its report, dimensions, faithful images and preserved history give rarity verifiable substance without promising an increase in value.

Identification: combining the evidence

The combined gemmological evidence

Refractive index, specific gravity, optic character, microscopic examination and UV response guide identification. None of these tests alone necessarily separates every member of the series. Anomalous birefringence does not exclude a cubic mineral; it may reflect strain or growth domains.

Naming the species from its chemistry

Distinguishing rhodizite from londonite requires assessment of alkalis in atomic proportions, ideally with spatial information in a zoned crystal. X-ray fluorescence is useful for several elements but cannot adequately characterise very light elements such as Li, Be and B on its own. Suitable LA-ICP-MS analysis or specialised techniques can complement the chemistry; ablation is microdestructive and must remain justified.

An average measurement may conceal different domains. Research compares several points, zoning images and analytical uncertainties before applying a dominance rule. Diffraction confirms the structure, but a shared structure alone does not determine the species name.

Why qualitative analysis is insufficient

Detecting both K and Cs does not establish their atomic dominance. Raw X-ray line intensities depend on matrix, geometry and instrumental sensitivities; calibration and correction are needed to obtain concentrations. Light elements B and Be present specific difficulties: calculated stoichiometry may be useful but must not be presented as direct measurement of every constituent.

Backscattered-electron mapping helps locate compositional contrasts: under comparable geometry and acquisition conditions, a domain containing more heavy elements may appear brighter. This contrast guides analysis points; it does not replace quantification. Point measurements across several zones avoid assigning one domain's composition to an entire stone.

Spectroscopy: distinguishing colour, structure and emission

UV–Visible–Near infrared

A spectrophotometer measures wavelength-dependent intensity. Transmission T = I/I₀ is related to absorbance A = −log₁₀(T). In the simplified case of a homogeneous path, absorbance increases with thickness; in a faceted gem, internal reflections and measurement geometry complicate interpretation.

The diagram below illustrates only how increased absorption towards violet–blue can favour yellow transmitted light. Its curve is constructed to explain the mechanism: it does not locate diagnostic bands measured in rhodizite.


Absorption and transmission: an explicit educational model, not a spectrum measured on the specimen.
Absorption and transmission: an explicit educational model, not a spectrum measured on the specimen.

Raman and infrared

Raman spectroscopy analyses frequency shifts associated with vibrations; infrared spectroscopy measures vibrational absorption. They explore structure under different selection rules. Their axes are generally expressed in cm⁻¹, distinct from the nanometres used for visible light.

Frost and colleagues published a vibrational study of a sample from Antandrokomby. It discusses modes associated with B-, Be- and Al-bearing groups and records no OH-stretching Raman band in their sample. This does not prove universal absence of water or inclusions throughout the series. [12]

Photoluminescence

Emission measurements must specify excitation wavelength, lamp filtering, detector response and acquisition conditions. A visually pale blue response under 365 nm guides the protocol; it does not locate a line at an exact wavelength. No chromium or rare-earth line is assigned here without spectral evidence.

Absorption, Raman, infrared and photoluminescence therefore answer complementary questions. An emission curve does not replace alkali analysis, and chemical identification alone does not predict colour.

Published vibrational reference features

The series studied by Laurs et al. notably shows Raman signals near 430 and 470 cm⁻¹, accompanied by weaker bands near 294, 544, 651, 803 and 857 cm⁻¹. Their proximity in rhodizite and londonite reflects a shared framework; these positions do not replace K–Cs analysis. [11]


Published Raman position markers for the rhodizite–londonite series
Published positions; conventionally equal heights, without reproducing a measured spectrum.

Choosing the method for the question

Method Information sought Points to consider
UV–Vis–NIR Electronic absorption and the origin of colour Optical path and stray light
Raman Framework vibrations; identification of certain inclusions Orientation, background fluorescence, resolution
FTIR Vibrational absorptions, structural or inclusion-related contributions Thickness and band saturation
Photoluminescence Spectral distribution of emission under defined excitation Filtering excitation and correcting detector response
Quantitative microanalysis Chemical occupancies and alkali dominance Calibration, light elements and zoning

Four specimens, from colourless to vivid yellow

Colourless rhodizite · oval · 1.42 carat

This oval from Madagascar stands out for its lack of perceptible colour, transparency and flashes of light. Its product page lists Eye clean VVS clarity and dimensions of 7.1 × 5.2 × 5 mm. Transparency offers a different view of facets and light paths from that of the yellow specimens.

The stone is presented as untreated. The reported yellow fluorescence is an observation specific to this specimen. Its portrait, media and characteristics are assembled on its product page; the report for the 1.55 carat cushion is not attributed to it.

Pale yellow rhodizite · pear · 22.62 carats

At 22.62 carats and 22.9 × 14.9 × 10.2 mm, this pear has a very different volume from the other three stones. Its product page describes a transparent, included, light yellow gem from Madagascar, untreated. Its substantial size reveals an internal architecture visible in macro photographs.

The lagoon-blue fluorescence and both photomicrographs in this article belong to this stone. These documentary views explore different lighting conditions and observation scales; they complement the white-light portrait on its product page.

Light yellow rhodizite · oval · 1.52 carat

This 1.52 carat oval offers a delicate yellow and measures 8.8 × 6.3 × 3.9 mm. Its product page describes a slightly included, natural, untreated stone from Madagascar. Its elongated outline and depth differ from those of the vivid yellow cushion.

The photograph reveals facets and variations in light within this gem. Its individual page brings together its data and media; neither the fluorescence of other specimens nor their laboratory documentation is transferred to it.

Vivid yellow rhodizite · cushion · 1.55 carat

The report LFG no. 406040 dated 1 December 2023 identifies a transparent, vivid yellow rhodizite with a cushion shape and brilliant / step cut. It records dimensions of 6.84 × 6.32 × 4.64 mm and concludes ‘No indications of treatment’.

Ultra Rare presents this specimen as originating from Madagascar and eye clean. The reported observation under long-wave UV at 365 nm is pale blue. Its stronger colour and cushion shape offer a fourth perspective on the collection. The stone's own report and media are accessible from its product page.

Method, conclusion and references

This overview combines text supplied by Ultra Rare, mineralogical publications and the specimen's documentation. Growth models, the binary section and absorption profile are educational. The unit cell uses published coordinates. Photographs are distinguished from computer-generated rough-crystal and jewellery visualisations.

Rhodizite connects a geology of concentration with a mineralogy of atomic sites. Its relationship with londonite does not diminish its identity: it shows why a gemstone name sometimes requires detailed analysis. From pegmatite to cut stone, transparency thus preserves a history that the eye alone cannot fully read.

Main sources and further reading

  1. Pring et al. (1986). The crystal chemistry of rhodizite: a re-examination. Mineralogical Magazine 50, 163–172. Coordinates: AMCSD 0014483.
  2. Pekov et al. (2010). Londonite from the Urals, and new aspects of the crystal chemistry of the rhodizite–londonite series. Canadian Mineralogist 48, 241–254.
  3. Bosi et al. (2024). CNMNC Newsletter 80. European Journal of Mineralogy 36, 599–604. Touretite, IMA 2023-003a.
  4. Černý & Ercit (2005). The classification of granitic pegmatites revisited. Canadian Mineralogist 43, 2005–2026.
  5. London (2022). A Rayleigh model of cesium fractionation in granite–pegmatite systems. American Mineralogist 107, 82–91.
  6. London (2014). A petrologic assessment of internal zonation in granitic pegmatites. Lithos 184–187, 74–104.
  7. Novák et al. (2015). Boralsilite and Li,Be-bearing “boron mullite”, breakdown products of spodumene from the Manjaka pegmatite, Sahatany Valley, Madagascar.
  8. Simmons et al. (2001). Londonite, a new mineral species: the Cs-dominant analogue of rhodizite from the Antandrokomby granitic pegmatite, Madagascar. Canadian Mineralogist 39, 747–755.
  9. Handbook of Mineralogy. Rhodizite and Londonite. Mineralogical Society of America.
  10. Larousse. Eau: its jewellery usage, referring to a stone's limpid appearance and transparency.
  11. Laurs et al. (2002). Rhodizite–Londonite from the Antsongombato Pegmatite, Central Madagascar. Gems & Gemology 38, 326–339.
  12. Frost et al. (2014). The molecular structure of the borate mineral rhodizite — A vibrational spectroscopic study. Spectrochimica Acta A 128, 291–294.

Keep this article

The scientific study in PDF format.

Download the article

Ultra Rare Gemstone · Luxury · Beauty · Rarity · Science · Nature · Heritage

Drag to explore the figure.

Atlases and maps