Er:Yb Glass vs Nd:YAG for Eye-Safe Laser Sources
*Most pulsed laser sources an engineer has worked with are Nd:YAG at 1064 nm. Moving to a 1.5 µm eye-safe
wavelength is not a matter of filtering the output — it needs a different gain medium. This article compares the
two, explains why erbium–ytterbium co-doped phosphate glass is the practical choice for compact 1535 nm sources,
and is honest about where Nd:YAG still wins.*
In short: Nd:YAG emits at 1064 nm, which lies inside the retinal hazard region and therefore has a low
permissible exposure limit. Er:Yb glass emits directly at 1535 nm, where the eye's front tissues absorb the
light before it reaches the retina. The trade is that Nd:YAG stores more energy per unit volume and conducts
heat far better, which is why it still dominates applications that do not need eye safety.
Why the gain medium is the decision
The output wavelength of a solid-state laser is set by the energy levels of the active ion in its host. It is not
a property of the cavity, the coatings or the pump — those can be optimised, but they cannot move the emission
line.
So when a specification says 1535 nm, that is a statement about which ion is doing the lasing. Nd³⁺ lases at
1064 nm. Er³⁺ lases at around 1535 nm. Those are different ions in different hosts, and they bring different
properties with them.
The consequence is that the choice of wavelength is made once, at the level of the gain medium, and everything
else follows from it.
Nd:YAG at 1064 nm
Neodymium-doped yttrium aluminium garnet is the workhorse of pulsed solid-state lasers, and deservedly so.
It stores energy well. Nd:YAG has a long upper-state lifetime — around 230 µs — which means it can be
pumped over a comparatively long interval and then released in a short, energetic pulse. That is why it reaches
joules per pulse in a modest package.
It conducts heat well. YAG is a crystal with high thermal conductivity, so waste heat leaves the gain
medium efficiently. That makes high average power practical, and it makes the thermal design of a
repetitively-fired laser tractable.
The technology is mature. Pump diodes, coatings, Q-switches and manufacturing processes are all
well-established, and there are many suppliers at every level.
Its limitation is the wavelength. At 1064 nm, light passes through the cornea and the lens and reaches the
retina, where the eye's own optics focus it onto a spot a few tens of micrometres across. The permissible
exposure limit is correspondingly low, and a Class 1 product cannot emit very much at all.
That is not a defect of Nd:YAG — it is simply what 1064 nm means for eye safety. For an application where the
beam is enclosed, or where the energy required is small, Nd:YAG at 1064 nm remains an excellent choice. The
problem appears when a system must emit substantial energy in open air where people may be present.
What about shifting the wavelength?
The wavelength can be moved by nonlinear conversion — an optical parametric oscillator, or Raman shifting — to
produce 1.5 µm from a 1064 nm pump. That approach is used, and it works.
It also adds a nonlinear crystal, a second resonator to keep aligned, and additional losses at each conversion
step. The result is a larger, more complex and more expensive transmitter than one that emits at the target
wavelength directly. For a compact, rugged instrument, avoiding the conversion is worth a great deal.
Er:Yb glass at 1535 nm
Erbium-doped glass lases directly at 1535 nm, which is the property that matters. Getting it to work
efficiently, in a small package, is where the design effort goes.
Why ytterbium is co-doped
Erbium alone absorbs pump light poorly. Its absorption around 980 nm is weak and narrow, which makes direct
diode pumping inefficient — a serious problem for a compact source, where every wasted pump photon becomes heat
in a small volume.
Ytterbium solves this. Yb³⁺ absorbs strongly and broadly across the 940–980 nm band, which is exactly where
high-power pump diodes are readily available. The ytterbium absorbs the pump energy and transfers it to the
erbium, which then lases at 1535 nm.
The two ions are therefore doing different jobs: ytterbium is the absorber, erbium is the emitter. This
sensitisation scheme is what makes diode pumping of a small erbium source practical.
Why a phosphate glass host
Erbium's solubility in YAG is poor, so a crystal host cannot hold enough of it in a short length. Phosphate
glass accepts very high erbium concentrations, which means a few millimetres of material can absorb a useful
fraction of the pump.
The high doping concentration also shortens the gain length, and short gain length is what allows the whole
assembly to be small.
The trade is thermal. Glass conducts heat far less well than YAG — roughly an order of magnitude worse — so
waste heat does not leave the gain medium easily, and the repetition rate is limited as a result. This is
visible in the specifications: the high-repetition models in this range top out at a few kilohertz, not the tens
of kilohertz a Nd:YAG design might reach.
Why the microchip construction
In a conventional laser, the gain medium sits in a cavity formed by separate mirrors that must be aligned and
held in alignment through shock and temperature change. That alignment is a large part of what makes a
conventional laser bulky and expensive to make rugged.
In a microchip laser, the mirrors are deposited directly onto the faces of the gain medium, and a saturable
absorber — the Q-switch — is bonded to it as well. The cavity is the gain medium. There is nothing to align
because there are no separate parts.
Three consequences follow, and they explain most of the specifications in this range:
It can be very small. The MINI-100 is 14.5×6.3×4.37 mm³ and 2.5 g, which is only possible because the
cavity is a few millimetres of glass.
It is robust. With no free-space optics to shift, the assembly survives vibration that would move a
conventional resonator out of alignment. The range is specified to 20–2000 Hz at 20 G, and the MINI-100 is
stated as meeting MIL-STD-810G test standards.
The pulse width is set by the cavity, not the driver. This is a point that causes confusion. The optical
pulse is produced by the Q-switch releasing the stored energy, and its duration follows from the cavity length
and the absorber — not from how long the driver is switched on. Increasing the drive pulse does not lengthen the
optical pulse; it risks damaging the source. This is why every model publishes a maximum drive pulse width, and
why drive pulse width and optical pulse width are different quantities.
The comparison, side by side
| Property | Nd:YAG | Er:Yb glass |
|---|---|---|
| Emission wavelength | 1064 nm | 1535 nm |
| Eye safety at 1064/1535 nm | Low permissible exposure | Orders of magnitude higher |
| Upper-state lifetime | ~230 µs | ~1 ms class, but limited by transfer |
| Thermal conductivity | High (crystal) | Low (glass), roughly an order lower |
| Energy storage per volume | High | Lower |
| Maximum repetition rate | Tens of kHz feasible | A few kHz in practice |
| Doping tolerance | Low | Very high |
| Construction | Often discrete optics | Can be monolithic microchip |
| Best suited to | Enclosed systems, high energy, high average power | Open-air operation, compact packages, eye safety |
Where Nd:YAG still wins
An honest comparison has to say this clearly.
Higher energy per pulse. Nd:YAG stores more energy per unit volume and can release it in a shorter pulse.
For a laboratory system in an enclosed beam path, that is simply more capable.
High average power. The thermal conductivity of YAG allows continuous or high-repetition operation that a
glass host cannot match. Industrial marking, cutting and welding are Nd:YAG territory — usually at 1064 nm,
sometimes frequency-doubled or tripled.
Materials processing. At 1064 nm and its harmonics, the absorption of metals and other materials is
favourable for machining. 1535 nm is not the wavelength those processes want.
Cost and supply chain maturity. More suppliers, more standard parts, more accumulated process knowledge.
Systems where the beam is enclosed. If no one can be exposed, eye safety is not a constraint, and the
question becomes purely one of optical performance.
Where Er:Yb glass wins
When the beam must be emitted into open air where people may be present. This is the decisive case, and it
covers most field instruments: handheld rangefinders, vehicle-mounted systems, airborne LiDAR, surveying
equipment, industrial sensors in shared spaces.
When the package must be small and light. A monolithic cavity is what makes a 2.5 g transmitter possible.
A discrete Nd:YAG resonator with the same output energy would be substantially larger and heavier.
When it must survive vibration and temperature swing. No alignment to lose means the optical performance does
not drift with mechanical stress.
When the receiver can be chosen freely. InGaAs detectors work well at 1535 nm, and the receiver design
follows from the requirement rather than from a legacy silicon choice.
A note on 1064 nm in this product range
The ERDI range is specified at 1535 nm throughout. Programs that require 1064 nm are handled separately, since
the gain medium, the optics, the coatings and the safety classification all differ. If a system is committed to
1064 nm — because the receiver, the optics or the qualification evidence already exists at that wavelength — say
so at the requirement stage rather than after a model has been proposed, because it changes which platform is
relevant.
For the 1535 nm range, the technology page covers the construction in more
detail, and the comparison with 905 nm covers the
eye-safety mechanism and the trade against silicon detectors.
Choosing between them
- Where does the beam go? If it is emitted into open air where people may be present, 1535 nm is the
practical choice.
- How much energy is needed? If the link budget needs more than a 1535 nm source can deliver in the
available envelope, reconsider the receiver before reconsidering the wavelength.
- How small must it be? Below a certain size, a microchip construction is the only realistic route.
- What repetition rate? Above a few kilohertz, the thermal limits of a glass host become the constraint.
- Is the wavelength already fixed? If a receiver and qualification evidence exist at 1064 nm, that may
settle the question before the optical requirements do.
Related reading
- 1535 nm vs 905 nm: eye safety and range compared
- Pulse width and range resolution
- How to choose pulse energy for a laser rangefinder
Next steps
The range runs from the MINI-100 at 2.5 g and 100 µJ to the
ER10000 at 10 mJ, all at 1535 nm and all Class 1. Full specifications are on the
Laser Sources page, and each model page carries its datasheet and a 3D model for mechanical
integration. For a requirement that a standard model does not cover, the customisation scope is described under
OEM / ODM, and Contact / RFQ is the route for a specific
enquiry.