FTIR ATR Crystal Selection: ZnSe vs Diamond vs Ge

Nine times out of ten the sample decides the crystal, not the budget. Diamond is the correct ATR crystal selection for a mixed, unknown or high-throughput sample stream because it is the only common crystal that tolerates both pH 1–14 and full clamping pressure. Choose ZnSe when samples are neutral, soft and low-volume, and choose germanium when the sample absorbs so strongly that a deeper penetration depth would saturate the spectrum — carbon-black-filled rubber and dark polymers being the classic cases.

The short answer. Order diamond first unless your sample list is narrow and chemically mild. Diamond is the one ATR crystal you can hand to any analyst without a compatibility question, and its cost is recovered by not replacing it. Add a Ge or ZnSe crystal plate only when a specific application demands what diamond cannot do: Ge for surface sensitivity and strongly absorbing dark samples, ZnSe for low-cost routine work and for the low-wavenumber range that a standard diamond puck does not reach.

What follows is a specification aid for an FTIR ATR accessory, not a product tour: the physics that sets penetration depth, the crystals compared on eight properties, the diamond absorption band that causes more rework than any other fact here, the exception that makes germanium a far-IR crystal, and the economics of swapping crystal plates over an instrument’s life. If you have not yet settled whether ATR or transmission is the right sampling mode, start there — this page assumes that decision is already made.

How ATR sampling sets the constraints on every crystal

An FTIR ATR accessory never transmits light through the sample. The beam travels inside a crystal of high refractive index, reflects at the crystal–sample interface, and a decaying evanescent field leaks a fraction of a micron into the sample, where functional groups absorb at their characteristic frequencies 1. For the underlying optics of how an FTIR spectrometer works, the linked guide covers the interferometer and optical train. Two consequences explain almost every recommendation below.

First, the sampling depth is small and wavelength-dependent. The penetration depth d_p is the distance at which the evanescent field falls to 1/e of its value at the surface:

where n1 is the crystal refractive index, n2 the sample refractive index, theta the angle of incidence (usually 45° or 42°), e lambda the wavelength 23. Because d_p scales with wavelength, a band at 1000 cm⁻¹ is sampled roughly four times as deeply as one at 4000 cm⁻¹. That inequality is why ATR spectra are not scaled copies of transmission spectra, and why the ATR correction routine — multiplying absorbance by wavenumber/1000 — exists at all 4.

Second, the crystal must have a higher refractive index than the sample. Most organic materials sit near n ≈ 1.5, and the common ATR crystals span n = 2.4 to n = 4.0, which normally leaves enough contrast [2]. When it does not, the spectrum distorts: peak symmetries diminish, baselines develop shoulders, and in bad cases derivative-shaped artefacts appear because anomalous dispersion changes the effective refractive index across a strong absorption band [2][4]. That failure is easy to miss, because the spectrum still looks like a spectrum.

The three crystals, spec by spec

The table below consolidates figures from Bruker, PIKE Technologies, Specac and Thermo Fisher — it is the short version of the ZnSe vs diamond ATR comparison that most procurement documents need. Where suppliers disagree, both numbers are shown, because the spread is real and worth negotiating against [2][3][4]56.

Three rows of that table carry most of the decision. The pH row eliminates ZnSe for any sample stream that is acidic or alkaline. The penetration-depth row is why Ge exists at all. The intrinsic-absorption row is the one that quietly invalidates results.

The diamond absorption band, and why it causes rework

Diamond is not transparent across the mid-infrared. The lattice absorbs strongly from roughly 2300 para 1900 cm⁻¹, with transmission falling to about 5 % in that window [2][4][6]. Functional groups that absorb there — nitriles, isocyanates, isothiocyanates, diimides, azides and ketenes — are therefore difficult or impossible to measure confidently on a diamond crystal.

The interaction with reflection geometry is counter-intuitive and worth understanding. A multiple-reflection diamond accessory puts more energy into the sample, which should help a weak band; but it also traverses more diamond, so the lattice absorption multiplies. In a published application study of a nitrile glove, a multiple-reflection diamond ATR returned only noise across the 2300–1900 cm⁻¹ window while a high-throughput single-reflection diamond accessory resolved the C≡N stretch at 2238 cm⁻¹ 7.

The operational rule is therefore: if a sample’s diagnostic band falls in the diamond blind spot, do not solve it with more reflections. Specify a ZnSe or Ge crystal plate for that application, or confirm the band with transmission. Ask the supplier to run your specific compound before you accept the quotation. Note that a diamond puck’s low-wavenumber cutoff is frequently set by its focusing lenses rather than the diamond itself — Thermo Fisher quotes a 650 cm⁻¹ cutoff for its diamond ATR Smart iTR because the accessory uses ZnSe lenses [6]. If you need to reach below that, the lens material is part of the specification.

Germanium: the exception that buys you the far-IR range

The standard objection to germanium is correct as far as it goes: its high refractive index produces a shallower penetration depth (0.65–0.67 µm) and therefore weaker absorbance than ZnSe or diamond, and its useful range stops around 600 cm⁻¹ [2][3][6]. For general screening, that is a genuine disadvantage.

But germanium has a property the other two lack, and it is systematically under-reported. In PIKE Technologies’ own crystal table, the long-wavelength cutoff is approximately 780 cm⁻¹ for Ge, 525 cm⁻¹ for ZnSe, e 250 cm⁻¹ for KRS-5 [2]. Because lower wavenumber corresponds to longer wavelength, the material that cuts off at the highest wavenumber reaches the furthest into the far-infrared. Germanium therefore covers the far-IR region that ZnSe cannot, and a standard mid-IR diamond puck does not reach either.

The second legitimate reason is surface specificity. Because Ge samples only the outermost fraction of a micron, it reports the surface layer rather than the bulk — useful for coating thickness, surface treatment and adhesion studies, and necessary for dark, highly absorbing materials where a deeper probe would absorb everything and flatline the spectrum [4][6].

ZnSe: the value crystal, and its two real limits

ZnSe is the most common ATR crystal in routine laboratories, for understandable reasons. Its refractive index (2.41–2.43) and penetration depth are practically identical to diamond’s, so the spectra look the same, and it is substantially cheaper [3][6]. For neutral liquids, soft polymers and screening work, it is the rational choice.

Its limits are specific and both are avoidable. Chemically, the acceptable pH window is 5–9. Acidic samples below about pH 5 etch the crystal surface, and in the worst case react to release hydrogen selenide (H₂Se), a toxic and flammable gas; strong alkalis above pH 9 also attack it [2][5][6]. Complexing agents such as ammonia and EDTA erode the surface even at moderate pH [2]. Physically, ZnSe is soft — 120–130 kgf/mm² against diamond’s 9,000 — so abrasive or hard samples scratch it, and the etched surface progressively loses throughput [3][5][6].

There is also a consumables detail that shortens crystal life more than most laboratories expect: cleaning materials. Only cotton-based wipes should be used on ZnSe and Ge crystals, because paper-based laboratory wipes gradually abrade the surface [4]. That is a one-line sentence in a standard operating procedure and it is worth writing into yours.

Reading a crystal specification without getting misled

Four numbers decide most of the argument, and each is easy to misread.

Depth of penetration. Always check the assumed conditions before comparing figures. O 1.66 µm and 2.03 µm values quoted for diamond are the same crystal measured under different assumptions (n₂ = 1.5, 45°, 1000 cm⁻¹) by different suppliers [2][6]. A quoted d_p without its incidence angle and sample refractive index is not a comparable number.

Refractive index. What matters is not the absolute value but its ratio to the sample’s. An n₂ ≈ 1.5 organic sample against n₁ = 2.4 is comfortable; a carbon-black-filled rubber or a mineral filler can push n₂ high enough that a 2.4 crystal no longer delivers enough contrast, and the spectrum distorts [2][4]. That is a refractive-index problem, not a hardness problem, and it is solved by a higher-index crystal.

Hardness. The Knoop figures (9,000 for diamond; 550 for Ge; ~120–130 for ZnSe) predict which crystal survives your sample preparation, not which produces the best spectrum [2][3]. Hardness decides whether high-pressure clamping is available to you, since diamond is the only common crystal suitable for high-pressure accessories [3].

Spectral range. Two different limits hide in this number, and suppliers rarely separate them. The intrinsic cutoff is a property of the crystal material; the system cutoff is what your accessory and spectrometer actually deliver, and it is usually set by the lenses, beamsplitter or detector. A range printed as “4000–400 cm⁻¹” is a claim about a system, not about a crystal.

The economics of a crystal plate you will replace

Crystal selection decisions are often made on the accessory price and revisited, expensively, two years later. The right comparison is total cost of ownership across the instrument’s life. Published accessory pricing shows the spread: a single-reflection ZnSe crystal plate assembly around US1,185, a Ge plate around US1,595, a single-reflection diamond plate around US6,311, and a diamond/KRS-5 plate — which extends the cutoff to 250 cm⁻¹ — around US6,470 8. Market surveys put diamond replacement kits in the US800–2,500 band and multi-reflection ZnSe or Ge crystals in the US400–1,200 band 9.

The arithmetic that matters is replacement frequency. Industry cost analyses estimate ZnSe replacement cycles of 2–4 years and Ge cycles of 3–5 years depending on duty, against a diamond crystal that typically outlives the instrument; the same analysis puts the productivity penalty of more careful handling and longer cleaning at 15–30 % for the softer crystals 10. A facility running aggressive or abrasive samples can therefore exceed diamond’s purchase premium within three to five years, while a facility running neutral liquids should buy ZnSe deliberately rather than apologetically [10].

Why interchangeable crystal plates beat a sealed accessory

Buy interchangeability, not a single crystal: an accessory with tool-free, pre-aligned crystal pucks lets one instrument serve several applications, e o BRT FTIR range supports tool-free plug-and-play ATR crystal modules for exactly this reason. Models such as the BRT-890 e BRT-860 are the ones most often specified with more than one crystal plate, while the BRT-850 pairs naturally with a single ZnSe plate for routine teaching and QC work. Price the second plate into the original order, since accessory prices rise, and confirm the housing material: plates for exceptionally caustic samples come in Hastelloy rather than stainless steel, at a premium [8].

What a spare crystal plate actually buys you

A second crystal plate is not redundancy; it is capability. A Ge plate turns the same instrument into a surface-analysis tool, and a ZnSe plate rescues any method whose diagnostic band sits in the diamond blind spot. For laboratories that identify samples at the goods-in bay rather than the bench, the portable BRT-8 Plus e BRT-800 accept modular ATR units, so the crystal decision travels with the instrument. Treat the plate as a consumable with a service life, not an accessory with an indefinite one.

Matching the crystal to the sample stream

The sample-stream decision in one table

What to put in the specification document

An ATR crystal specification is only useful if it survives procurement. State the crystal material and the accessory geometry separately, name the housing material, and give every parameter with its test conditions. A quotation that says only “diamond ATR” leaves the lens material, the cutoff, the plate interchangeability and the replacement pathway undefined — and each becomes a change order later.

The verification step is worth more than the specification. Send the supplier three representative samples, including the worst case on pH and the most abrasive, and require the returned spectra with search scores. Where a standard-method identification is claimed, require the method reference and confirm that the ATR correction applied matches what the reference data expects, since ASTM E573 and ASTM E1252 both treat internal reflection as a distinct technique requiring documented corrections rather than a drop-in equivalent to transmission [1][4].

Common crystal-selection mistakes

Verifying the crystal at acceptance

Put these to the supplier in writing before the order is placed: which crystal material is supplied, in which housing, with which lens material; the measured system cutoff with the delivered configuration; the pH and temperature limits in writing; the replacement part number and current lead time for a spare crystal plate; whether a heated or extended-range puck is available for the same accessory; and what is required to swap crystals without a service visit.

Conclusion: the selection order we would follow

1. List the sample types, and mark any that are acidic, alkaline or abrasive. If any exist, diamond is the crystal.
2. Check whether the diagnostic bands of your key methods fall between 2300 e 1900 cm⁻¹. If they do, add a ZnSe or Ge plate to the order.
3. Check whether any sample is dark, highly absorbing or high-refractive-index. If it is, add germanium as a surface-analysis tool rather than a backup crystal.
4. Confirm the low-wavenumber cutoff you need, and whether it is set by the crystal or by the accessory lenses.
5. Specify the second crystal plate, the housing material and the replacement part number in the original purchase order.
6. Require your own worst-case samples to be run, with spectra and search scores, before acceptance.

Done in that order, the crystal question resolves in a single meeting. If you would rather not run it alone, send us your sample list and the methods you are working to — we will come back with a crystal recommendation and the configuration that delivers it. Laboratories specifying a bench and an accessory at the same time will also find our UV/Vis spectrophotometer buying guide useful for the parallel specification logic in optical spectroscopy.

About the author. This guide was prepared by the BRT Testing Applications Team (Shenyang Baoruitong Automation Equipment Co., Ltd.), which supports FTIR and materials-testing laboratories with instrument selection, method transfer and acceptance documentation. Contact us at BRT Testing or review the full FTIR spectrometer range; our company profile covers the wider materials-testing portfolio.

Perguntas frequentes

Is ZnSe or diamond better for FTIR ATR?

Diamond is better if your sample stream includes acids, bases, abrasives or unknowns, because it covers pH 1–14 and is hard enough for high-pressure clamping. ZnSe is better if samples are neutral, soft and routine, because it matches diamond’s refractive index and penetration depth at substantially lower cost — but it is limited to pH 5–9 and scratches easily. The deciding question is chemical and mechanical exposure, not performance.

When should I choose germanium over diamond or ZnSe?

Choose germanium when the sample absorbs too strongly for a deeper probe, such as carbon-black-filled rubber or dark polymers, or when you need surface-only information from a coating or treated layer. Its refractive index of 4.0 gives a penetration depth near 0.65 µm, roughly three times shallower than diamond or ZnSe. Germanium also reaches further into the far-infrared than ZnSe.

What is the depth of penetration of ZnSe, diamond and Ge ATR crystals?

Quoted at 1000 cm⁻¹ and 45° incidence for a sample of refractive index 1.5, diamond and ZnSe both give roughly 1.66–2.03 µm, while germanium gives about 0.65–0.67 µm. These figures are not absolute: penetration depth scales with wavelength and depends on the incidence angle and both refractive indices. Always compare depth-of-penetration values at the same stated conditions.

Does diamond ATR have an absorption band that limits its use?

Yes. The diamond lattice absorbs strongly between roughly 2300 e 1900 cm⁻¹, where transmission falls to about 5 %. Functional groups absorbing in that window — nitriles, isocyanates, isothiocyanates, azides and ketenes — are difficult to measure on diamond. If your diagnostic band falls there, specify ZnSe or Ge for that method rather than adding reflections.

What pH range can each ATR crystal tolerate?

Diamond and germanium both tolerate pH 1–14, while ZnSe is limited to approximately pH 5–9. Acids below pH 5 etch ZnSe and can release toxic hydrogen selenide gas; strong alkalis above pH 9 also attack it, and complexing agents such as ammonia and EDTA erode the surface. If your sample stream is acidic or alkaline at all, diamond or germanium is the safer specification.

How deep into the sample does an FTIR ATR accessory actually measure?

Typically only 0.5 para 2 µm at mid-infrared wavenumbers, which is why ATR reports the surface layer rather than the bulk of a material. The exact depth is governed by the penetration-depth equation and varies across the spectrum, so a band at 1000 cm⁻¹ is sampled far more deeply than one at 4000 cm⁻¹. This wavelength dependence is why ATR spectra need correction before library matching.

How much does an ATR crystal plate cost to replace?

Published accessory pricing puts a single-reflection ZnSe plate near US1,185, a germanium plate near US1,595, and a single-reflection diamond plate near US$6,311. Diamond looks expensive until replacement frequency is counted: ZnSe and Ge plates are typically replaced every two to five years depending on duty, while a diamond crystal usually outlives the instrument. Price the spare plate into the original order.

Can I change the ATR crystal myself, or does it need a service visit?

On an FTIR ATR accessory built around interchangeable, pre-aligned crystal pucks, the crystal plate swaps without tools and without optical realignment, so an operator can change it between samples. On accessories with the crystal permanently mounted, replacement is a factory or service procedure. Confirm which design you are buying, and ask for the replacement part number and lead time in writing.