1535 nm vs 905 nm: Eye Safety and Range Compared
*The two wavelengths used in laser rangefinders and LiDAR differ in a way that decides what pulse energy a
design can use. This article explains the mechanism behind 1535 nm eye safety, what it does and does not allow,
and where the trade-offs actually fall.*
In short: at 1535 nm the light is absorbed in the front of the eye and spread through the thickness of the
cornea; at 905 nm it reaches the retina and is focused onto a small spot. That single difference is why a
1535 nm source may emit far more energy within Class 1 — and it is the reason most field instruments that must
be safe around people use 1535 nm.
The mechanism
The relevant property is how the eye's own tissues absorb light at each wavelength. The eye is not uniform:
different parts have different compositions and therefore different absorption spectra.
At 1.5 to 1.55 µm, absorption in the crystalline lens is negligible. Around **75% is absorbed in the
cornea and 25% in the aqueous humour**. Almost none of the beam reaches the retina. Because the absorption
happens through the thickness of the cornea rather than at a single point, the energy deposited per unit volume
stays low, and the cornea — which is about as resistant to radiation as skin, and regenerates well — is not
readily damaged. This is why the 1.5 µm region is described as an ocular-safety domain in the laser-safety
literature. (Optique Ingénieur, laser safety)
At 905 nm, the situation is reversed. Absorption in the cornea and lens is low, so the beam passes through
to the retina. There the eye's own optics focus it: a collimated beam a few millimetres across is concentrated
onto a spot on the order of tens of micrometres. The hazard is not the total energy so much as the
concentration of it.
The wavelengths that present the highest retinal hazard are those between roughly 400 nm and 1400 nm — the
range in which the eye focuses light onto the retina and the ocular media do not block it. Beyond about
1400 nm, the hazard shifts from the retina to the front of the eye.
A note on the wording, because it is often repeated incorrectly: it is not the vitreous humour that absorbs the
beam. The absorption happens in the cornea and the aqueous humour, in front of the lens.
What this means for accessible emission
Laser classification under IEC 60825-1 sets an accessible emission limit for each class, and that limit is a
function of wavelength. Because the eye's damage threshold differs so much between 905 nm and 1535 nm, the
Class 1 limit at 1535 nm is far higher than at 905 nm.
For a product designer the practical reading is this: at 1535 nm you may emit considerably more pulse energy
and still be Class 1, than you could at 905 nm. If the instrument must be safe to use where people are present
— a surveying site, a construction area, a vehicle in traffic, an aircraft overflying an area — that difference
changes what energy is available to the link budget.
The sources in the ERDI range are specified as Class 1 at 1535 nm, from 100 µJ up to 10 mJ. What that
energy is worth in range is worked through in
How to choose pulse energy for a laser rangefinder.
The comparison that matters less than it is made to sound: detection
A claim sometimes seen is that 1535 nm is worse for detection because silicon detectors are less sensitive
there. That is true of silicon, which responds well at 905 nm and poorly beyond about 1100 nm.
It is not true of the detectors actually used at 1535 nm. InGaAs photodiodes and APDs respond in this band, and
they are the standard choice. The receiver is a design decision, not a limitation imposed by the wavelength.
The fair comparison is between two complete receiver chains, not between two detector materials.
Atmospheric transmission
Both wavelengths are transmitted by the atmosphere, and both suffer scattering and absorption that grow with
range and with the amount of water, dust or haze in the path. The two do not behave identically: over a given
path in haze or rain, the losses differ.
This is a condition to be stated in a specification rather than a general advantage for either wavelength. A
range figure quoted without a visibility condition is not comparable to another quoted under different
conditions — which is a common reason two datasheets appear to disagree.
Where 905 nm still wins
Being honest about the trade-off makes the choice easier.
Silicon detectors are cheaper and well established. For a short-range instrument built in volume, a 905 nm
source with a silicon APD can be the more economical design, and the eye-safety constraint may not bind if the
energy required is low and the instrument is used away from people.
Some designs are already committed. If a receiver chain, an optical design and a supply chain are in place
at 905 nm, changing wavelength is a redesign, not a component swap.
At very short range, less energy is needed. An instrument measuring a few metres may need so little energy
that the Class 1 limit at 905 nm is not a constraint at all, and the wavelength question becomes less
important.
Where 1535 nm wins
When energy is the constraint. Long range against a diffuse target needs a large returned signal, and the
return falls with the fourth power of range. If the receiver is already fixed, energy is what you have to
increase — and at 1535 nm far more of it is available within Class 1.
When people are in the beam path. This is the decisive case for most field instruments. It is what allows a
terrain-mapping system to operate over an area, or a handheld rangefinder to be used without a warning label
that limits where it can be pointed.
When the receiver is a free choice. Starting fresh at 1535 nm means choosing an InGaAs detector without
having to work around an existing silicon design.
When the product line must scale. Every model in this range shares one wavelength, one beam quality and one
safety class, so a single optical and receiver concept can be reused across energy tiers. That is worth more
than it sounds: it is one design and one set of qualification evidence instead of several.
A note on the phrase "eye-safe"
"Eye-safe" is used loosely and is worth being precise about. The label means the product falls within a
classification whose accessible emission is below the level considered safe under defined conditions — not that
the beam cannot cause harm under any circumstance.
Three qualifications matter in practice:
Classification assumes defined conditions of distance, duration and beam geometry. A source that is Class 1
as a product may not be if the beam is focused by external optics.
The near field is not the far field. A collimated beam is more hazardous close to the aperture, where it has
not yet spread.
Optical instruments change the answer. Viewing any laser through binoculars or a telescope concentrates the
collected light and can defeat the classification. This is what the Class 1M and 3R categories address.
A 1535 nm source is safer than a 905 nm source at the same energy; it is not a licence to disregard the beam.
The range of sources here
Every model in the ERDI range shares the same 1535 nm wavelength, Class 1 classification and TEM00 beam quality,
so a receiver and signal-processing chain designed for one can, in principle, serve the others. Only the energy
tier, the repetition rate and the drive change.
| Tier | Models | Pulse energy | Repetition rate | Drive current | Package |
|---|---|---|---|---|---|
| Micro package | MINI-100, ER100 | 100 µJ | 1–10 Hz | 8–12 A | from 14.5×6.3×4.37 mm³ |
| Compact | ER200, ER300, ER400, ER500 | 200–500 µJ | 1–10 Hz | 10–20 A | 21×8×6.8 to 35×8×6.8 mm³ |
| Standard | ER800, ER900 | 800–900 µJ | 1–10 Hz / 1–5 Hz | 30 A | 40×8×7.5 mm³ |
| High energy | ER2000, ER4000, ER8000, ER10000 | 2–10 mJ | 1–5 Hz | 80–100 A | 60×25×13.5 to 110×40×24.5 mm³ |
| High repetition | ER10, ER20, ER40, ER70 | 10–70 µJ | 500 Hz – 5 kHz | 5 A | 17×8×7 mm |
Full specifications for each are on the Laser Sources page, and the
1535 nm technology page covers the Er:Yb glass microchip construction.
Choosing between them
Work it in this order:
- What energy does the link budget require? Range, target reflectance, aperture, detector sensitivity and
the margin the detection logic needs.
- Are people in the beam path? If yes, the Class 1 limit is a design constraint, and it is far more
generous at 1535 nm.
- Is the receiver already chosen? If it is silicon, the wavelength may already be decided.
- What are the atmospheric conditions? State them, because a range figure without them is not comparable.
- What envelope is available? Energy tier and package size move together.
If the answer to step 2 is yes and step 1 calls for more than a few hundred microjoules, 1535 nm is usually the
only workable choice.
Next steps
For a specific requirement, the Contact / RFQ form takes range, target, update rate
and envelope, and we will identify which parameters need to be fixed before design freeze. Datasheets and 3D
models for every model are published on the Laser Sources page, and the customisation scope is
described under OEM / ODM.