Loupewise / Guide / How to Use a Refractometer
Refractometer: a wedge-shaped body with a hinged flap over the prism window and a cylindrical eyepiece at the low end

How to Use a Refractometer for Gemstone Identification

Last updated: April 2026

Calibration, reading technique, doubly refractive stones, and interpreting your results at the bench.

The refractometer measures how much a gemstone bends light — the refractive index (RI). RI is the most reliable single diagnostic constant in gemology. Unlike color or appearance, RI is a fixed physical property determined by crystal chemistry: sapphire is always near 1.762–1.770, spinel near 1.712–1.736, and synthetic cubic zirconia (2.15–2.18) is so far past the instrument’s ceiling that its over-the-limit reading is itself the clue. No treatment, no lighting condition, and no gem dealer's description changes those numbers.

A well-calibrated refractometer reading, combined with specific gravity, identifies most faceted gems without further testing. The reading takes under two minutes once you have the technique. This guide walks through calibration, stone placement, reading doubly refractive (DR) stones, and interpreting what you see in the eyepiece.

Start here: how to read these pictures Every other figure in this set assumes you have seen this one. Each label below lights up with the part it names. 1.40 1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80 1.401.451.501.551.601.651.701.751.80 This circle is what you see Put your eye to the eyepiece and this round field is the whole of it. Nothing else is visible. Above the boundary is darkThat is not a shadow on theglass. It is the part of thescale the light cannot reach. The coloured line IS the readingWhere dark meets light is thenumber you write down — here1.724. A good one is dead straightand hair sharp. The numbers are refractive indexThey increase upward, and the printeddivisions are 0.01 apart — the same ruler inevery figure here. TWO THINGS THAT APPEAR THROUGHOUT DO / SEE / MEANS Under each figure: what to do with your hands, what you should then be looking at, and what it tells you. Read them in that order. A cross in the corner That figure shows a FAULT — something to recognise and avoid, never a target to aim for. Example is spinel, computed from gems.json. The orientation — index increasing upward, dark field above the boundary — follows the sourced wording in the Layer 1 dossier; the edge-quality bar is eurogemology's. Leader terminators follow ISO 128-22: a dot where the leader ends within an outline, an arrowhead where it ends on one.

What is one good reading worth? In Loupewise's own engine, a refractive index alone narrows 132 species to a median of 22 candidates, and settles 0 outright on its own — a stone above the instrument's 1.81 limit gives no reading at all, which is itself a strong signal. Add the birefringence the same instrument gives you on a doubly refractive stone and the two figures together do most of an identification. These are simulated perfect readings — an upper bound the technique below is how you approach.

Types of Refractometers

Not all refractometers are equal. Make sure yours covers the gem species you commonly work with.

What You Need

Gemological refractometer — standard (to 1.81) is sufficient for most bench work. Brands: Eickhorst, GIA, Kassoy, OPL.

Refractometric contact liquid (RI 1.81) — methylene iodide based. Keep it tightly sealed and shielded from light when not in use.

Monochromatic light source — sodium vapor lamp or LED yellow lamp filtered to ~589 nm. A white light source gives blurry, color-fringed readings.

Two known control stones — a quartz variety (reads 1.544–1.553) and a synthetic corundum (1.762–1.770). Stones with book values are how a gem refractometer is verified at the start of a session, and the same pair doubles as reading practice.

Soft lint-free cloth — wipe the stone free of fingerprints before each reading, and wipe the hemicylinder gently after each session. Never let contact liquid dry on the glass: dried liquid can scratch it, so re-wet any residue with a fresh drop rather than scraping.

The Calibration Check

Gem-A quotes the faceted-stone error at no more than 0.005 — a miscalibrated instrument eats that entire budget. A gem refractometer's scale is set at the factory and has no user adjustment; what you calibrate is the session, by reading stones whose values you already know. Do it at the start of every session, and again if you move the instrument between environments: Cargille's datasheet for its 1.80-series contact liquid gives a temperature coefficient of about -0.0007 RI per °C, so a 10 °C room change alone shifts the fluid by roughly 0.007.

1

Read a known stone

Take a stone whose values you trust — a quartz variety is the canonical choice — and take a full reading exactly as described in the next section. Quartz should show its two edges at 1.544 and 1.553.

2

Compare against the book values

Write down what you actually read and subtract the book values. A difference is a signed offset — note whether it is high or low, not just how big it is.

3

Read a second known stone

Repeat on a stone in a different part of the scale — a synthetic corundum (1.762–1.770) pairs well with quartz. One stone tells you something is off; two tell you what.

4

Interpret the offsets

The same signed offset on both stones means the instrument is biased — a small, constant bias can be mentally subtracted from every reading. An offset on one stone but not the other means the problem is technique or contact on that stone, not the instrument. Re-read it.

5

Begin the session

Wipe the stage gently with a soft cloth and proceed to your unknowns. Repeat the check if you change light sources, move rooms, or notice readings drifting between known stones.

Where the light goes The lamp sits behind the instrument and shines into the rear window. Nothing is read through the stone. Light does NOT travel up through the stone — which is exactly why an opaque stone stillgives a reading. it bounces here THE READING HAPPENS HERE At the contact surface where thepolished facet meets the glass, througha film of liquid. The stone only has tobe flat and polished — not clear. THE REAR WINDOW All the light goes in here —never at the top, and neverthrough the stone itself. TO YOUR EYE The internal scale and theshadow edge are both readthrough this one lens. slide the lamp until thefield is brightest andmost even Sourced: light enters at the rear window (GP cutaway; IGS "behind the instrument, directed into the light window"). The reading is total internal reflection at the contact surface, not transmission through the stone (Hughes/Lotus; the GP wiki explicitly denies the light-through-the-stone model; TR612’s first sentence admits opaque stones with a flat polished facet). No surveyed source states a lamp distance or angle, so the figure gives the observable to adjust for rather than a number to trust; the measurement is kit-session item 1. Body proportions are drawn from Krüss ER605 and Gem-A dimensions (1 : 2 : 3.5), tall at the stage end and stepping down toward the ocular; the ocular is drawn perpendicular to the sloping face it is mounted on, which is how all three photographed instruments carry it. The internal mirror POSITION is simplified because no manufacturer cutaway is published, and the hemicylinder is drawn oversized against the stage plate so the contact surface is legible at this size.

Taking a Reading

The stone must have at least one flat, polished facet. The table facet works well for most cuts. Curved surfaces (cabochons) yield a spot reading, not a shadow edge — useful for an approximate RI but insufficient to determine optic character.

1

Apply contact liquid to the stage

Place a very small drop on the stage — smaller than you think you need. Excess liquid spreads under and around the stone and can migrate into the instrument.

Applying the liquid Four steps, on the flat face of the hemicylinder seen from above. The stone is never pressed and never dropped on. 1 One small drop The smallest drop you canplace — a pin head. Centreof the glass. 2 Stone beside it Set it on the metal besidethe glass, largestpolished facet face down. 3 Slide it on Nudge it across with afingernail until it coversthe drop. Never pressdown. 4 The film spreads The liquid pulls out intoa thin film under thefacet. That film is thecontact. AND THE TWO WAYS IT GOES WRONG TOO LITTLE No optical contact, so the scalereads blank — exactly as if therewere no stone on the glass at all. TOO MUCH The stone floats on the liquid andthe reading skews; on a cabochonthe spot bloats past two or threedivisions and grows a dark ring. Sourced: drop size "pin head" (GO); centre placement (GP); stone set beside the glass and slid on, table down (GP course, quadruply sourced); "slide gently", never press (TR612, checked verbatim — defect D7). Failure signatures: blank scale as if no stone were present (GO-Torraca); stone lifts and the reading skews, spot bloats past two or three divisions with a dark ring (GP/Hoover, IGS). Drop diameter in millimetres is unmeasured — kit session item 2 — so these panels show relative size only.
2

Place the stone on the stage

Set the stone table-down (or any flat polished facet down) onto the liquid, then slide it gently into position — Gem-A's instruction sheet says to slide the stone down the glass table until the shadow edge appears. Do not press: the glass is very soft, and the slide alone makes the optical contact. The stone should sit flat — any tilt will shift the reading.

3

View under monochromatic light

Look through the eyepiece with your light source in position. The field of view shows a scale. You are looking for a distinct light/dark boundary — the critical angle shadow edge.

4

Find the shadow boundary

Identify where the illuminated (light) area meets the dark area. This boundary corresponds to the RI on the scale behind it. If the boundary is blurry or shows color fringes, your light source is not monochromatic enough — or the stone surface or liquid is contaminated.

Positioning your eye Hunt for the edge first by moving your head. Then hold still — every reading after that is taken from the same place. SQUARE AND CENTRED Looking straight down themiddle of the eyepiece, atright angles to it. close enough to see thewhole scale at once Self-check: you can see the whole scalewithout moving your head, and the scalelines are single, not doubled. OFF TO ONE SIDE Or drifting between onereading and the next. This shifts the reading by about 0.002 —a quarter of ruby’s entirebirefringence. It is the differencebetween resolving two edges andreporting one. Moving your head to FIND the edge is correct and expected. Holding it still once you have foundit is what makes two readings comparable. Sourced: perpendicular through the centre of the eyepiece, head position locked between readings, and the whole-scale / no-doubling self-check (GP course). Hunting for the edge by moving your head first is correct practice (DCGIA). The 0.002 parallax spread is eurogemology's worked figure; ruby's birefringence of 0.008 is from gems.json (this diagram states it; it does not compute it). Only one source offers an eye-to-eyepiece distance, so the figure gives the self-check to adjust for instead of a number; the measurement is kit-session item 3.
Is that edge good enough to read? The bar is specific, and it is the whole difference between a measurement and a guess. CORRECT Straight and hair sharp 1.40 1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80 One boundary, dead straight, with no softness on either side of it. This is the sight you are hunting for. NOT A READING Fuzzy, or curved 1.40 1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80 A soft band instead of a line, or a boundary that bows. Do not read it — the contact is bad or the drop is flooded. Wipe and start again. Sourced: the quality bar is eurogemology’s — a usable edge is “100 percent straight and hair sharp”; fuzzy or curved indicates bad optical contact or too much fluid (GP/Hoover, IGS). This is why the correct-sight figures render a crisp line: softening it would draw the fault as the target.
5

Record the RI value

Read the scale value at the shadow edge. This is your first RI reading (omega or epsilon ray for uniaxial stones; one of the two principal vibration directions for biaxial stones). Note it to three decimal places.

6

Rotate the stone 360 degrees and observe

While watching the eyepiece, slowly rotate the stone on the stage through a full 360-degree rotation. Two distinct outcomes are possible:

  • Shadow edge stays fixed — singly refractive (SR): isometric crystal system or amorphous material. Examples: garnet, spinel, glass, synthetic CZ, opal.
  • Two shadow edges move independently — doubly refractive (DR): non-isometric crystal. The two edges approach and separate as you rotate. Record both the minimum and maximum RI values at their widest separation.
7

Calculate birefringence for DR stones

Birefringence = RI max − RI min. A high birefringence value (e.g., calcite at 0.172, zircon at 0.059) is itself diagnostic. Low birefringence (beryl at 0.006, corundum at 0.008–0.010) can be hard to see but is measurable with care.

The gap between the two edges is the birefringence LOW EXTREME Sugilite DR_U- · birefringence 0.002 1.40 1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80 magnified ×22 0.2 scale divisions apart HIGH EXTREME Calcite DR_U- · birefringence 0.172 1.40 1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80 17.2 scale divisions apart BEYOND THE INSTRUMENT Malachite DR_B- · birefringence 0.254 1.40 1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80 1.909 only 15.5 of its 25.4 divisions fit below the ceiling Sugilite's two edges are 0.002 apart — smaller than this instrument's own stated error of 0.005, so it cannot report them as two. Calcite has the widest gap in our 132 species with both edges still on the scale. Wider gaps run off the top, as malachite's does. Computed from gems.json — the three species are the measured extremes, not chosen by eye. Error figure: TR612. Scale divisions are 0.01.

Interpreting the Result

The combination of RI value(s) and optic character (SR or DR) narrows identification dramatically. Add birefringence for DR stones and you have eliminated most candidates.

Singly Refractive (SR)

One fixed shadow edge that does not move during rotation. Crystal is isometric or stone is amorphous.

Examples: Garnet, spinel, glass, synthetic CZ, diamond, opal

Doubly Refractive — Uniaxial

Two shadow edges; one stays fixed (ordinary ray), one moves. Hexagonal or tetragonal crystal system.

Examples: Corundum (ruby, sapphire), quartz, tourmaline, apatite

Doubly Refractive — Biaxial

Two shadow edges, both move independently during rotation. Orthorhombic, monoclinic, or triclinic system.

Examples: Topaz, chrysoberyl, alexandrite, orthoclase, peridot

Spot Reading (no edge)

A fuzzy dot of light rather than a shadow edge. Indicates a curved surface (cabochon). Gives approximate RI only; optic character cannot be determined.

Examples: Any cabochon-cut stone, some heavily rounded beads

Key RI Values — Commonly Confused Species

Use this table to cross-reference your reading. Overlapping ranges between species are where birefringence and specific gravity become decisive.

Gem Species RI Range Optic Character Birefringence Notes
Sapphire (corundum) 1.762–1.770 DR, uniaxial − 0.008 RI often read as 1.762/1.770 pair
Ruby (corundum) 1.762–1.770 DR, uniaxial − 0.008 Same crystal as sapphire; color only difference
Spinel 1.712–1.736 SR (isometric) None SR distinguishes from ruby immediately
Tourmaline 1.624–1.644 DR, uniaxial − 0.020 Strong birefringence; visible doubling in deep stones
Tsavorite Garnet 1.730–1.760 SR (isometric) None SR and high RI separate from emerald
Emerald (beryl) 1.565–1.602 DR, uniaxial − 0.006 Low RI separates from tsavorite and demantoid
Aquamarine (beryl) 1.577–1.583 DR, uniaxial − 0.006 Same species as emerald; RI range overlaps heavily
Topaz 1.609–1.643 DR, biaxial + 0.008 Biaxial — two moving edges; SG 3.49–3.57 confirms
Tanzanite (zoisite) 1.691–1.700 DR, biaxial + 0.009 Strong trichroism; biaxial birefringence visible
Amethyst / Quartz 1.544–1.553 DR, uniaxial + 0.009 Consistent RI regardless of color; SG 2.65

Readings above 1.81 (off-scale or at the edge): This indicates a high-RI stone that exceeds the standard refractometer's range. Use specific gravity and fluorescence to distinguish: demantoid garnet (SG 3.82–3.88, and typically inert to UV), zircon (SG 4.60–4.80, strong birefringence visible as back-facet doubling), and cassiterite (SG 6.8–7.1, rare). Diamond (RI 2.42) and moissanite (RI 2.65) are also off-scale and require dedicated testers.

Loupewise accepts your RI reading and immediately ranks matching candidates — no manual table lookup required. Enter RI min and max, toggle optic character, and the engine filters 132 gem species in real time.

Open Loupewise

Studying for the FGA or GIA practical? Loupewise drills instrument-based identification — refractometer, SG, and UV readings — and runs timed mock exams built around the real exam formats.

Practice exams →

Where This Comes From

Technique and failure modes on this page are re-verified against published sources (2026); the elimination figures are computed from Loupewise's database and engine and are guarded by tests.

Frequently Asked Questions

How do I know if my reading is off?

Read a known stone before reading unknowns — a quartz variety should show 1.544 and 1.553, and if it does, the instrument is reading true. Ambient temperature changes the RI of the contact liquid — Cargille publishes a coefficient of about -0.0007 RI per °C for its 1.80-series liquid, and bounds the intended use of its 1.81 gem fluid to 20–26 °C — so re-calibrate if you have moved between significantly different environments (e.g., a cold storage room to a heated workshop). A reading that is systematically high or low across multiple known stones also indicates calibration drift.

Can I get a reading off a curved surface?

The other method: reading a cabochon A curved back has no facet to give you a shadow edge, so you sit back and read a spot instead. EYEPIECE VIEW, MAGNIFIED ×7 EYE WITHDRAWN A dot, floating 1.64 1.65 1.66 1.67 1.68 Sit back from the eyepiece — start about a foot away and adjust — and the scale blurs while a small droplet-shaped dot floats in the field. That dot is the reading. It is not a fault. THE READ POINT Half dark, half light 1.64 1.65 1.66 1.67 1.68 Move your head slowly. The dot's centre goes from dark to light, and where it is half and half is the reading — 1.66 here. Two decimals. The spot method does not give you a third. TOO MUCH FLUID The dot bloats 1.64 1.65 1.66 1.67 1.68 Wider than the two or three divisions it should be, with a dark ring around it and a cutoff that bows instead of running straight. Wipe it off and use less. 75 OF THE 132 SPECIES HERE ARE DOUBLY REFRACTIVE — THE SPOT METHOD CANNOT TELL YOU THAT ABOUT ANY OF THEM Sourced: eye withdrawn — the GP course says start about 30 cm, and the surveyed range across sources is 15–45 cm, so the figure says start and adjust rather than fixing a number. The correct sight is a droplet-shaped spot bisected half dark and half light at the read point (GP course, GO-Frank, and the Liddicoat-style "center changes from dark to light"). The spot must span no more than two or three scale divisions; wider, with a dark ring and a curved cutoff, is too much fluid (IGS, Liddicoat-style). Precision is 0.01 and never a third decimal — TR612 puts the spot method's stated error at 0.01 against 0.005 or better for the faceted method. Shown magnified ×7: at true size a scale division is about three pixels deep, so a correct spot would be seven pixels tall. The 75-of-132 figure is counted from gems.json; the spot method reports refractive index only, with no birefringence and no optic character.

A curved surface (cabochon, bead) gives a spot reading — a dot of light in the field of view rather than a clean shadow edge. This yields an approximate RI (center of the spot) but cannot determine optic character, birefringence, or the individual RI rays. For cabochons, the spot reading is useful as a rough confirmation but should be combined with SG and other tests. A flat polished facet is required for a full refractometer reading.

Why are my two shadow edges not crisp?

Three common causes: (1) The contact liquid is drying — the boundary blurs as the liquid evaporates, especially in warm environments. Re-wet and re-read. (2) The stone face or refractometer stage is contaminated with oil, grease, or residue from a previous stone. Wipe the stone with a soft lint-free cloth and the stage gently with a soft cloth — a little water if needed — and never let liquid dry on the glass. (3) The light source is not truly monochromatic — white light produces colored fringes (dispersion) around the shadow edge. A proper sodium lamp or 589 nm LED gives a sharp, achromatic boundary.

What stones cannot be read on a standard refractometer?

The scale is dark all the way down This is the one sight that looks like an answer and is not. Darkness on its own permits no conclusion at all. No boundary anywhere, top to bottom — and no faint line from the liquid either. Three things cause that, and nothing in the eyepiece separates them. The stone is over the limitIts index is higher than yourcontact liquid's, so theinstrument has nothing to showyou. You used too much liquidIt floods the glass and swampsthe boundary. The commonest ofthe three, and the cheapest toundo. The set-up failedNo optical contact, the lampmisplaced, the lid open, or youreye in the wrong place. SO ELIMINATE THE TWO YOU CONTROL, IN THIS ORDER 1 Wipe both dryThe glass and the stone.Soft cloth, no solvent onthe hemicylinder. 2 Re-drop, minimallyThe smallest drop you canplace. If cause two was it,the edge appears now. 3 Prove the instrumentRead a stone you alreadyknow, well inside therange. If that reads, theset-up is good. 4 Only now, log itA dark field that survivesall three is a measurement.Write it down as one. A negative reading is a result, not a failure With a 1.81 contact liquid it rules out 110 of the 132 species here and leaves 22. Which 22 depends on the bottle, not the instrument: on 1.79 liquid the count is 23. Read the label before you name the candidates. Sourced: the three co-equal causes and the wording “dark all the way down to the bottom” are TR612's fault table (p.4) with DCGIA step 7; the wipe / minimal re-drop / prove-on-a-known-stone sequence is bank RF-11. The ceiling is the contact liquid's index, not the glass's, and grades run 1.78–1.81 — so the candidate counts are computed here against 1.81 and shown against 1.79 as well. This figure deliberately prints no list of over-the-limit species: RF-11 records that any fixed list is wrong by construction, because the set is a function of the bottle. In this database the two grades differ by a single species.

Stones with RI above 1.81 exceed the scale: demantoid garnet, zircon, cassiterite, sphalerite, diamond, moissanite, and synthetic rutile. Stones without a flat facet (cabochons) give only a spot reading. Facets smaller than about ⅛ inch (3 mm) across are difficult — though not impossible — to get a reading from; the limit is the facet size, not the stone. Opacity itself is not a barrier: the reading comes from total internal reflection at the polished surface, not from light passing through the stone, and Gem-A's leaflet lists "transparent, translucent and opaque gemstones with a flat polished facet" among what the instrument measures. A polished turquoise or lapis reads on-scale; malachite shows only its lower edge, because its upper indices sit past the 1.81 fluid ceiling. What genuinely cannot be read is a rough, unpolished surface — no polish, no optical contact, no edge.