2026 UPDATE

​​​

Ruby is s a 3-state laser that requires enormous light input to lase and is erratic when lasing. Q-switching is mandatory. Inrad pockels 202-090 BBVIS (400-700nm) is suitable. Needs a fast turn on and a slower fall time. Only suitable for single pulse as heat from multi-pulse will degrade optical performance.

​

Google AI:

Elimination of Thermal Lensing: Because the ruby rod has minutes to completely dissipate heat between flashes, thermal gradients across the crystal drop to zero. The severe shot-to-shot mode hopping and beam divergence fluctuations common in ruby lasers are completely eliminated.

​​

Active Q-switching will solve the primary erratic behavior of a ruby laser, turning it from an uncontrollable mess of spike trains into a highly predictable, single-pulse utility suitable for Sum Frequency Mixing (SFM).

​

However, Q-switching does not fix the underlying thermal and material flaws inherent to ruby; it simply alters how the laser cavity stores and releases its energy. [1]

​

1. What Q-Switching Fixes: Eliminating the "Spike Train"

Left to its own devices in a free-running (normal) pulsed mode, a ruby laser exhibits highly erratic relaxation oscillations. Instead of delivering a clean, predictable pulse, it fires a chaotic "spike train" lasting about 1 millisecond. [1, 2]

 

The Chaos: This train consists of hundreds of microsecond-spaced micro-pulses with random amplitudes and unpredictable timing.

 

The Q-Switching Cure: By placing an optical shutter (like an electro-optic Pockels cell) inside the cavity, you block the mirrors while pumping. The crystal stores a massive population inversion. When the shutter snaps open, all that stored energy dumps out in one single, clean "giant pulse" lasting 10 to 40 nanoseconds. [1, 2, 3]

​

True single shot libs

​

Operating in a true single-shot regime fundamentally alters the physical challenges of your ruby laser:

 

Elimination of Thermal Lensing: Because the ruby rod has minutes to completely dissipate heat between flashes, thermal gradients across the crystal drop to zero. The severe shot-to-shot mode hopping and beam divergence fluctuations common in ruby lasers are completely eliminated.

 

Flawless Spatial Beam Profile: Your ruby beam will achieve its maximum possible spatial uniformity and wavefront cleanliness, drastically increasing your SFM conversion efficiency in the cooled BBO crystal.

 

No Thermally Induced Birefringence: The Inrad 202-090 cell will experience zero thermal loading, preventing any unwanted polarization rotation or leakage while the cavity is closed.

​

Inrad 202-090 BBVIS

​

Performance Analysis of the Inrad 202-090 BBVIS at 694.2nm

The Inrad Optics Dual-Crystal KD*P Pockels Cell is a highly reliable choice for this configuration, though its design demands precise optical alignment:

 

The BBVIS Coating Advantage: The broadband visible coating provides high transmission across the 400 nm–700 nm range. At the 694.2 nm ruby line, it sits near the upper edge of its performance curve, offering low insertion loss and minimizing internal optical reflection.

 

Low Half-Wave Voltage (\(V_{\pi }\)): Because this is a dual-crystal design (crystals arranged optically in series but electrically in parallel), its operating voltage is halved. At 694.2 nm, expect a quarter-wave voltage (\(V_{\pi /2}\)) of roughly 1.5 kV to 1.8 kV (for a double-pass cavity layout) or a half-wave voltage (\(V_{\pi }\)) of 3.0 kV to 3.6 kV (for a single-pass layout). [1]

 

The Single-Shot Piezoelectric Challenge: KD*P is highly piezoelectric. When fired at multi-kilohertz rates, acoustic ringing can cause polarization distortion. At one pulse every few minutes, piezoelectric ringing is a non-issue, as the crystal completely settles long before the next shot.

​

Timing

​

Because you are mixing two completely separate laser systems (a Ruby and a 266 nm Nd:YAG), your timing window is incredibly narrow. The pulses must overlap inside the cooled BBO crystal within a few nanoseconds. An avalanche transistor string delivers the exact electrical performance needed to make this happen

​

your master timing circuit must bridge two entirely different laser media. Ruby requires a massive flashlamp pump time compared to Nd:YAG. Your synchronization workflow will follow this sequence:

​

 

T = 0μs  ───► Fire Ruby Flashlamp (Requires ~1,000 μs to pump

T ≈ 800μs  ─► Fire Nd:YAG Flashlamp (Requires ~200 μs to pump)

T ≈ 1000μs ─► Fire Nd:YAG Q-Switch (Emits 266nm pulse)

[Delay Line] ──► Trigger Avalanche String ──► Fire Ruby Q-Switch

​

The Ruby Head Start: The master clock fires the Ruby flashlamp first because a ruby rod requires roughly 1,000 microseconds of pumping to reach peak population inversion.

 

The Nd:YAG Inversion: Around 800 microseconds later, the Nd:YAG flashlamp fires (as Nd:YAG only requires about 200 microseconds of pumping).

 

The Coincidence Window: At the 1,000-microsecond mark, both crystals are fully charged.

 

The Avalanche Drop: The Nd:YAG Q-switch fires, emitting the 266 nm pulse. A fast delay generator receives a pick-off signal from this pulse and triggers your avalanche string. The string dumps the voltage on the Inrad Pockels cell, letting the 694.2 nm ruby pulse escape exactly as the 266 nm pulse arrives at the BBO crystal.

​

Critical Feasibility Metrics for the Avalanche Driver

To ensure your conceptual driver design aligns with the Inrad 202-090 cell, you must design for three specific electrical realities:

​

Capacitive Loading: The Inrad 202-090 has an internal capacitance (typically around 4 to 8pF depending on the housing and connector choices). Your avalanche string must be able to rapidly charge or discharge this capacitance without distorting the voltage wavefront.

 

Pulse Tail Control: For single-pulse LIBS, you want a clean, isolated pulse. Avalanche strings naturally produce a very fast leading edge. However, you must design a passive RC decay network or a pulse-forming line (PFL) to clip the pulse width to match your Nd:YAG profile (typically 5–10ns), preventing a long 'tail' of ruby light from passing through after the 266nm light has gone.

 

Voltage Overhead: Because KD*P requires roughly 3.0 to 3.6kV for a single-pass half-wave switch at 694.2nm, your string will require roughly 10 to 12 transistors in series (FMMT415 or 2N5551 variants).

​

HOWEVER BBO IS THE ISSUE

​

And we don't have to use 193nm as 196.7nm is perfectly adequate:

​

1. The Physics of LIBS Below 200nm: 193nm vs. 197nm

From an analytical chemistry perspective, there is no meaningful performance gap between 193nm and 197nm for LIBS. The critical threshold is simply staying below 200nm to operate within the 'Vacuum UV' boundary.

​

=====================

​

In 2026 Google AI revealed, as I suspected, LIBs does not have to go as low as 193nm to achieve improved fractionation from molecular vs thermal disintegration.

 

The choice of 193nm is a perfect intersection of a hard physical threshold and engineering practicality. It does not have to be exactly 193nm, but it must be below 200nm to achieve truly athermal ablation across all material matrices. Argon Fluoride (ArF) excimer lasers became the industry standard because they were the most reliable, high-power commercial technology capable of crossing that specific deep-ultraviolet (DUV) boundary

 

The Physics Constraint: Why It Must Be Below 200nm [1]

Wavelengths above 200nm lack the single-photon energy required to cleanly disrupt chemical bonds in transparent or resilient materials. [1]

 

Harmonic limitations: Solid-state Nd:YAG lasers easily reach 266nm or 213nm using nonlinear frequency conversion. [1, 2]

 

​Thermal fracturing: Studies comparing 213nm and 193nm ablation show that 213nm sits on the fence, still causing micro-cracking and melting residues on tough matrices like silicates and calcium carbonates. [1, 2, 3]

 

Aerosol particle size: Wavelengths below 200nm yield an aerosol composed strictly of ultra-small sub-150nm particles. [1]

 

Perfect atomisation: Only these micro-particles can be fully and stoichiometrically converted into ions without elemental fractionation. [1, 2]

 

Why 200nm

​​​

The Oxygen Absorption Edge: Ambient oxygen (O₂) features a major molecular absorption band (known as the Schumann-Runge band) that begins abruptly at 200nm. Light with a wavelength shorter than 200nm is rapidly absorbed by air, which is why the region below 200nm is officially termed the Vacuum Ultraviolet (VUV) spectrum. [1, 2]

 

The 6.2 eV Photon Energy Wall: A wavelength of 200 nm corresponds precisely to a photon energy of 6.2 eV. Most tightly bound matrix bonds (such as Silicon–Oxygen bonds in quartz and silicates, which sit around 5 eV to 6 eV) require single-photon energy greater than 6.2 eV to undergo direct, non-thermal photolytic cleavage.

 

The 213nm vs 193nm Gap: The nearest major solid-state laser wavelength is 213nm (the 5th harmonic of Nd:YAG). Because 213nm sits just above the 200nm threshold, it still triggers thermal melting and fractional crystallization in tough matrices. Crossing below 200nm to 193nm breaks past this physical barrier, providing a clean transition into athermal ablation.

 

Why 204nm is Sometimes Cited

​

The number 204nm appears in specific industrial and spectroscopic contexts but is not the physical threshold for athermal ablation

​

For Laser-Induced Breakdown Spectroscopy (LIBS) and laser ablation, 200nm remains the absolute physical line in the sand.

​

Conclusion

There is no scientific advantage to forcing exactly 193nm for a single-pulse LIBS setup. The historical choice of 193nm is no longer relevant given modern solid-state laser capabilities.

​

========================

​

Therefore anything below 200nm is ok.

​

193nn BBO has to be cooled for multi-shot BUT NOT FOR SINGLE SHOT.

197nm BBO does not have to be cooled for multi-shot.

​

I can make 196.7nm by SFM 266nm + 755nm. 

​

========================

​

I have a 10.5mm x 8.5mm dia Ti:S rod.​

​​

Ti:S is also problematic with a very short pulse that does not lend itself to dual pulse LIBS if I should want to experiment that route. The (typically 532nm) end-pumped crystal I have is tiny (8.5mm dia x 10.5mm long) and yields low power. However not a lot is required for LIBS.
 

A LIBS ~197nm setup only needs one Nd:YAG laser:

​​

You can use the same primary Nd:YAG laser to pump 532nm into the Ti:S as well as produce tuned 755nm , but it requires a specific architecture. Using a standard single-pulse laser will fail because the pulses will entirely miss each other in time.

​

Temporal Alignment

When a 532nm nanosecond pulse hits a Titanium-doped Sapphire (Ti:S) crystal, the crystal absorbs the green light and stores that energy in its upper lasing state (which has a lifetime of ~2.6 to 3.2µs).

​

The Ti:S laser does not fire instantly. It takes anywhere from 30ns to several hundred nanoseconds of cavity round-trips for the 703nm pulse to build up and escape the cavity (gain switching).

​

Meanwhile, if you split a portion of that exact same Nd:YAG pulse to make 266nm, that 266nm pulse is only 5 to 10ns wide. It will arrive at your SFM crystal instantly and terminate long before the 755nm pulse ever leaves the Ti:S cavity.

​

 

The Solution: A Double-Pulse Nd:YAG Laser

To make a single Nd:YAG laser work for a single-shot LIBS event, you must use a dual-pulse (or twin-pulse) Q-switched Nd:YAG system (very common in dual-pulse LIBS architectures).

 

Flashlamp/Diode Fire: The laser head is energized once.

 

Pulse 1 (532nm): The first Q-switch fires a 532nm pulse to pump the Ti:S crystal. The Ti:S stores this energy for its ~2.6µs lifetime.

 

Delay: You wait roughly 1.5 to 2.5µs.

​

Pulse 2 (266nm): The second Q-switch fires. This pulse passes through the fourth-harmonic generation (4HG) crystals to become 266nm.

 

Cavity Dump / Synchronization: At that exact microsecond mark, an active electro-optic Q-switch inside your Ti:S cavity releases the 755nm pulse. Both pulses arrive at the BBO crystal at the exact same moment.

​

------------------------

​

​what about timing - NEED A 2 pulse Nd:YAG

​

1. to gen 532 > TiS

2. to sum TiS 755 + 266nm

​

EG&G 9650A will trigger two distinct Q-switch event pulses separated by a few nanoseconds: [1, 2]

 

[Flashlamp Fires] ──> Wait 180µs (Max Inversion) │ ├──> Trigger Pulse 1 (EG&G Ch B) ──> Fires 532nm Only ──> Pumps Ti:Sapphire │ │ │ (Ti:Sapphire builds up over ~20ns) │ │ └──> Trigger Pulse 2 (EG&G Ch C) ──> Fires 266nm ─────────────┴─> Perfect Mix in BBO (Delayed by exactly ~20 ns)

 

The 532nm First Pulse (Pump Pulse): You trigger the Pockels cell with an abbreviated voltage step (e.g., partial voltage drop or a very fast gating slice). This releases just enough stored energy from the YAG rod to create a strong 532 nm pulse. It hits your Ti:Sapphire rod, starting the ~20ns cavity build-up time. [1, 2, 3, 4]

 

The Electronic Delay: While the Ti:Sapphire is building up its 755nm pulse internally, your EG&G 9650A counts down exactly 20ns on Channel C.

 

The 266 nm Second Pulse (Mixing Pulse): Channel C dumps the remaining stored population inversion out of your Nd:YAG rod. This second 1064nm flash hits your harmonic conversion crystals, instantly generating a 4 ns wide 266 nm pulse. [1, 2, 3]

 

The Intersection: Because you managed the 20ns delay electronically inside the YAG cavity, the 266nm pulse exits the laser head and arrives at the BBO crystal at the exact picosecond the 755nm pulse breaks out of the Ti:Sapphire resonator

​

Issues:

​

While this eliminates physical bench delays, shifting the burden to the electronics introduces three specific design rules for your DIY laser head:

 

Pockels Cell Driver Recovery Time: Standard high-voltage Pockels cell switches (like classic Marx generators or simple thyratrons) require microseconds to recharge their capacitance before they can fire a second time. To pull off a 20ns double pulse, you must use a modern high-speed MOSFET-based Pockels cell driver designed specifically for burst-mode or double-pulse configurations (e.g., a push-pull circuit topology). [1, 2]

 

Harmonic Crystal Cross-Talk: When the first pulse (532nm) passes through your fourth-harmonic generator crystal (BBO/KDP), it will technically convert a small slice of itself into 266nm. You must make sure your first pulse isn't so energetic that this unintended UV leak pre-triggers or targets your BBO mixing crystal prematurely.

 

Energy Balancing: The second pulse will naturally have slightly less gain available in the YAG rod because the first pulse depleted part of the population inversion. You will need to carefully tune your partial-voltage gate on the first trigger pulse so that both pulses exit with well-balanced energies. [1]

​

========================

​

Alexandrite is a better choice.

​

In 2026 I acquired a refurbished (as new) Alexandrite 5mm x 117mm rod.

​

Alexandrite can be tuned from 710nm to 820nm but has a major emission at 755nm.

755nm + 266nm = 196.7nm.

​

​By designing a conceptual layout utilizing an Alexandrite laser tuned to 755nm mixed with the fourth harmonic of an Nd:YAG (266 nm), you will generate a 196.6nm single pulse. This configuration delivers identical deep-UV plasma mechanics while drastically lowering the risk of crystal damage, providing a highly reliable platform for your high-stakes, single-shot system.

 

Summary

​

No Dual-Pulse Stress on Electronics: Your DIY Nd:YAG laser only has to fire one single pulse, which means your Pockels cell driver doesn't need ultra-fast nanosecond recovery times.

 

Massive Energy Storage: The 117mm Alexandrite rod will easily absorb enough energy from a standard flashlamp to output the 20mJ to 50mJ needed to drive your BBO crystal efficiently.

 

Stable 755nm Base: With your Birefringent Filter (BRF) installed, the Alexandrite laser stays locked onto its native 755nm peak, keeping your 196.7nm output completely stable from shot to shot.

 

By using the flashlamp to power the Alexandrite, you leverage the crystal's long 260µs lifetime as a natural timing buffer.

​​​

Alexandrite has a unique physical trait: its stimulated emission cross-section increases at higher temperatures. Allowing the coolant loop or cavity body to warm to approximately 50°C to 60°C will naturally boost the laser's gain and efficiency.

​

BRF

​

Wavelength Control: Placing an intra-cavity birefringent filter (BiFi) or a high-power prism tuner inside the resonator will allow you to lock the wavelength onto the required 755nm line.

​

You do not strictly need a Birefringent Filter (BRF) to get an Alexandrite laser to emit at 755nm.

By default, an untuned Alexandrite laser will naturally free-run and lase at its peak gain wavelength, which is right around 755nm. Because an 80% Output Coupler (OC) provides a high amount of output coupling (meaning high cavity loss per pass), the laser cavity will automatically favor the wavelength that provides the highest possible stimulated emission gain to overcome that loss. That peak is 755nm. [1, 2]

However, whether you can get away with omitting the BRF entirely depends on your specific application requirements:

​

Even if you only want 755nm, you may still need a BRF if your application demands the following:

 

Spectral Line Narrowing: An untuned, free-running Alexandrite laser has a relatively broad linewidth (typically 2nm to 5nm wide). A BRF acts as a line-narrowing element, compressing that spectral width down to fractions of a nanometer for precision applications. [1, 2]

 

Temperature Stability: Alexandrite’s peak gain wavelength is highly temperature-dependent. As the crystal heats up (often purposefully up to 70°C–100°C to increase efficiency), the peak gain shifts slightly toward the red (e.g., 760 nm+). If your application requires exactly 755nm regardless of crystal temperature fluctuations, a BRF is required to 'lock' and hold that exact wavelength.

​

If you mix an untuned Alexandrite laser with a 266nm source using Sum Frequency Mixing (SFM), the output wavelength will be centered around 196.7nm, with a likely operating range of 196.5nm to 197.0nm depending on your crystal's temperature.

 

Omitting the Birefringent Filter alters the deep ultraviolet (DUV) output in two distinct ways:

 

Output Bandwidth: Without a BRF, your Alexandrite fundamental has a broad free-running linewidth of about 2nm to 5nm. When mixed with a stable 266nm line, this translates to an ultraviolet spectral breadth of roughly 0.3nm to 0.4nm.

 

Center Wavelength Drift: As the Alexandrite crystal heats up under high pump power, its peak gain naturally shifts redder toward 760nm. Without a BRF to anchor it, your DUV output will slowly drift upward from 196.7nm toward 197.0nm.

 

LACK OF A BRF IS BAD FOR LIBS:
 

If your laser output drifts back and forth between 196.5nm and 197.0nm from shot to shot, it can completely corrupt your analytical LIBS results.

​

While the ablation target does not care about a 0.5nm difference (the physical plasma will still ignite), the plasma emission data collected by your spectrometer will suffer drastically for two major reasons:

​

 

1. The Calibration Error (Fake Peak Shifting)

LIBS relies on pinpointing the exact atomic emission lines of elements, many of which are dense and tightly packed in the deep-UV (DUV) region (such as Carbon, Phosphorus, Sulfur, and various metals).

 

If your excitation wavelength shifts, the exact spatial position, temperature, and plasma density of the micro-plume will shift slightly from shot to shot.

 

More critically, if any stray excitation light leaks into your spectrometer, a 0.5nm jitter will create moving artifacts across your baseline. This obscures real elemental peaks and makes automated peak-identification software highly unreliable.

​

2. Radical Changes in Ablation Efficiency and Plasma Temperature

Nonlinear sum frequency mixing (SFM) efficiency is incredibly sensitive to wavelength. If the Alexandrite laser drifts, your 197nm pulse energy will fluctuate violently from shot to shot because the BBO crystal will fall in and out of its strict phase-matching angle.

 

The Shot-to-Shot Chaos: One shot might give you a robust 5mJ pulse at 196.7nm (perfect phase match), while the next shot drifts to 197.0nm, misses the phase-match window, and drops to an anemic 0.5mJ. [1]

 

Plasma Fluctuations: Because LIBS line intensity scales exponentially with plasma electron temperature, a massive drop in pulse energy means the plasma will be much cooler on the second shot. The emission lines you are trying to measure will abruptly dim or disappear entirely, destroying your quantitative data. [1]

​

THEREFORE A BRF IS MANDATORY FOR NON-NATIVE NM LIBS (i.e. Alexandrite 755nm)

​

​​​========================

 

TEMPERATURE

​

Running Alexandrite hot at 70c 80C maximizes its 755nm output energy.

​

========================
 

what is the relative power output of a nominally doped Nd:YAG rod the same size as the Alexandrite?

The fundamental output energy of the Nd:YAG rod will be roughly 2x to 3x higher than the Alexandrite rod
 

========================

​

Inrad pockels

​

Fast high-voltage switching creates a massive amount of Electromagnetic Interference (EMI).

 

Driving two Inrad Pockels cells simultaneously means dumping several kilovolts (3kV - 5kV) in less than 5 ns using hydrogen thyratrons or fast MOSFET Marx banks.

 

This high-dV/dt noise can easily feed back through the BNC cables and glitch or reset your EG&G 9650A timer.

 

The Fix: Use double-shielded RG-223 or RG-214 coaxial cables for the Pockels cell triggers, keep the EG&G unit on a completely separate mains power strip from the high-voltage capacitor banks, and consider adding ferrite cores to your signal lines.

​​​

========================

​

EG&G 9650A timing

​

Because the upper-state lifetime of Alexandrite (~260µs) is slightly longer than Nd:YAG (~230µs), your flashlamp triggers must be staggered, while your Pockels cells must fire almost simultaneously.

You can configure your EG&G channels using this precise sequencing:

​

Channel A (T = 0 µs): Fires the Alexandrite Flashlamp SCR/Ignitron.

 

Channel B (T ≈ +30µs): Fires the Nd:YAG Flashlamp SCR. (Adjust this window so both rods reach peak population inversion at the exact same moment).

 

Channel C (T ≈ +260µs): Fires the Alexandrite Pockels Cell driver.

 

Channel D (T = Channel C ± Jitter): Fires the Nd:YAG Pockels Cell driver.

 

Engineering Tip: You will need to fine-tune the delay between Channel C and D by a few nanoseconds using the 9650A's fine-delay knobs to compensate for the different optical path lengths from each laser cavity to the BBO crystal.

​

================================

 

BBO SFM

​

Crucial Engineering Design Parameters for BBO at 197nm

To ensure success with this setup, three critical engineering constraints must be managed:

 

Keep the Crystal Thin (0.1mm - 0.5mm)

At 197nm, BBO still exhibits high dispersion and an intense spatial walk-off angle. If the crystal is too thick, the fundamental and the UV beams will physically separate inside the crystal, dropping your conversion efficiency to zero. A very thin crystal (0.2mm is ideal) minimises walk-off and internal absorption. [1, 2]

 

The Polarization Clean-Up

Because BBO is highly birefringent, the 197nm UV light generated via Type-I phase-matching will emerge with a polarization vector rotated 90 deg relative to the incoming pump beam. You will need a simple Pellin-Broca prism or a fused-silica atmospheric separator to split the generated 197nm beam from the residual fundamental pump light before it hits your LIBS sample target. [1]

 

Moisture is the Enemy [1]

BBO crystals are notably hygroscopic.

Keep the crystal in a sealed housing with a packet of silica gel, or keep a small, low-power 50C heating element wrapped around the crystal mount to ensure it stays bone dry when not in use.

​​

​================================

​

Crystal polarization

​

For Type-I SFM in BBO, the 266nm and 755nm input beams must have the same polarization (e.g., both vertically polarized). The generated 197nm output will emerge orthogonally (horizontally polarized), making it very easy to separate using a simple prism.

​

Ask me if you would like to map the exact Type-I phase-matching angle profile for BBO at 196.6nm.

​

================================

​

Suprasil B Pellin Broca #1 to combine 266nm + 755nm

​

Scenario B: Using the Prism to Combine the Input

You can use the prism in reverse to combine the 755nm and 266nm beams into the BBO crystal.

 

However because Suprasil B  has high dispersion in the deep UV, the refraction angle difference between 266nm and 755nm inside the prism is massive. Aligning this to get perfect spatial overlap inside a tiny BBO crystal is incredibly difficult.

 

A dichroic combiner is much friendlier for the input stage. TOO EXPENSIVE

 

Use PB & these tricks instead:

Achieving perfect spatial overlap between two narrow, nanosecond laser pulses inside a tiny BBO crystal using a prism can be tedious. You can make this much easier using two simple workshop tricks:

 

Use a HeNe or Green Diode Alignment Laser: Feed a cheap, continuous-wave green pointer laser backward through your BBO crystal and into the output face of the Pellin-Broca prism. The prism will naturally split that green beam into different exit paths. Mark where those paths hit your bench. If you shoot your 755nm and 266nm lasers precisely back down those exact exit lines, they are guaranteed to merge perfectly inside the prism.

 

Long Focal Length Lens: Place a long focal length lens (e.g., f = 500mm or 1000mm) before the prism on both laser lines. This creates a longer beam waist, meaning the beams stay overlapped over a longer distance, giving you a much wider margin for error inside the BBO crystal.

- - - - - - - - -  - - - - -  - - -

​​

Use the 213nm Pellin Prism to Separate the Output

Bring your 755nm and 266m beams together collinearly using a standard dichroic mirror before the BBO crystal. Then, place the Pellin-Broca after the crystal.

 

A Pellin-Broca provides a constant 90° deviation for the target wavelength while dispersing all others. By setting the prism angle for 197 nm, your generated LIBS beam will shoot out cleanly at a perfect 90° angle to your bench, while the residual, dangerous 266nm and 755nm pump beams will exit at wildly different angles into your beam dumps.

​

NOT TRUE: A PB DISPERSES FOCUS ACROSS THE ENTIRE SPECTRA

All spectra exiting from the PB will have a different focus point, and any lens trying to focus spectra into it will induce chromatic abberations leading to differing focus at the entrance as well. PB doesn't have parallel i/o faces so spectra will enter and exit at differing angles and will also travel different distances inside the prism. The only recourse is to use a parabolic mirror at the entrance and another at the exit.

​

All of which which makes me want to return to a simple mirror setup to filter out DUV and NIR.

​

The main issue is getting rid of the reheating NIR. Realistically I think that's going to be 1064nm and if this is insurmountable then ultimately I need to find a way to get the intensifier gate to work,

​

-----------------------------

​

SFM (266nm + 755nm) -> 197nm

​

​

-----------------------------

​

COLD ABLATION

​

197nm breaks the molecular bonds producing a very clean ablation and improved signal but because it is no longer a thermal disintegration, the atomic signal is weaker.

​

To counter this, 1064nm will also be fed to the target. The output lens will focus the 197nm and the 1064nm will be out of focus but this is desirable as otherwise ablation would again be thermal. Instead, the heat of the 1064nm will excite the atomic lines, amplifying them as before.

 

1064nm notch (blocking) filters are very expensive if they even exist, so if I go this route I will need to gate the FO as before.

​

Better: feed the 1064nm separately into the plasma and once passed over the target, into a beam dump.

 

AND use mechanical triggering as before. 

​

Two ARC 193nm +/-18nm 45 AOI mirrors will be placed in 'Z' orientation between the target and the filter to deflect 197nm reflections off the target into beam dumps. The Z cancels out the offset introduced by each mirror as the now ~211 - 1000nm atomic signal passes through to the notch filter and then FO. The mirrors are spaced so there is 40mm between the target and the FO, which is just enough air to degrade 197nm beam deflections without severely degrading the spectra.

​

Note: Argon will not now be added as although this amplifies the spectra, it also significantly amplifies the Bremsstrahlung radiation which then threatens the PCM-401 intensifier MCP. 

​

-----------------------------

​

PROTECTING THE PCM

​

To completely isolate your delicate image intensifier from the highly destructive 197nm excitation beam while gathering clean LIBS data, you must deploy two distinct optical isolation strategies.

 

 

1. Geometric Filter (The 90° Trick)

​

Never place the fiber optic collection probe or the intake slit of the PCM-401 directly in the path of the laser beam.

 

The Setup: Position your 197nm laser firing straight down at a 90-degree vertical angle relative to your sample surface.

 

The Collection: Place your PCM-401 collection optics at a 45-degree oblique angle looking sideways into the plasma plume. Because laser reflection is specular (angle of incidence equals angle of reflection), the destructive 197 nm bounce will shoot harmlessly away from your spectrometer intake.

 

2. [June 2026] 197nm focused 100um molecular 'cold ablation'

Molecular disassembly is better than thermal but the cold much reduces atomic spectra intensity.

I therefore added an unfocused beam of 1064nm fundamental to reheat the plasma to increase intensity.

The final DUV lens would render the 1064nm defocused but NIR on the target would again thermally ablate.

The ideal route is 197nm to the target, and 1064nm into the plasma ~0.5 to 2mm above the target.

The 1064nm then runs into a beam dump.

This necessitates 2 separate beams, parallel being optimum so the FO can be at 90 degrees to them.

However the ungated PCM-401 FO gets reflections of 197nm from the target and 1064nm from the plasma.

These reflections will damage the intensifier and must be filtered out.

The PCM-401 range is 200nm to 1000nm. Numerous ideas to remove the DUV & NIR dismissed:

 

Notch filter far too expensive / dielectric mirror 1064nm AR coating stops UV tx /

1064nm line mirror blocks lower harmonics too / pinhole spatial mirror: UV ~>230nm but NIR mirror poor /

Si wafer at Brewster's 74.3 degrees to relflect spectra and absorb or tx P-pol 1064nm won't work with plasma reflected 1064nm which will scramble the polarisation (and reflection % is fairly low) 

​

3. DUV / NIR Optical Filter

​

Final choice: DUV Pellin Boca to gate the central spectra from the 197nm & 1064nm.

 

MY 1st Google AI session told me to use Thorlabs tubes but I have Newport LP-1-XYZ mounts, so the arrangement below will change:

 

Use Newport RRA-1?? [see 67 notes dated 15/06/26] retainer rings to hold pinhole (between O rings) & lens in Newport LP-1-XYZ than can hold up the 1/2" thick optic.
 

Add a black plastic tube to the outside 11/4"? dia of the Newport

To lock these optics securely inside the Newport mounts without them falling out, you just need 1-inch Newport retaining rings, not bulky external tubes.

 

If your mounts did not come with them, you can use these exact parts:

 

Newport LP-1-RR: This is the official Newport 1" Retaining Ring designed specifically for the LP series mounts.

 

Newport LT-1-WR: Alternatively, a standard Newport 1" Wave Washer combined with a retaining ring can prevent over-tightening and cracking on delicate optics like ceramic pinholes.

​

Suprasil B Pellin Broca #2 in Aly box bolted to breadboard​

​An off-Axis Parabolic (OAP) Mirror coated in UV-enhanced aluminum reflects all wavelengths at the exact same angle, eliminating chromatic aberration entirely

rest of optics in a Thorlabs 1" dia x 2" (or longer) tube after the PB

A 50mm space below the PB widens the gap between DUV & NIR

A ceramic pinhole with black carbon loaded ESD foam on the surface to absorb the 197nm & 1064nm.

A F-20mm 1" lens concentrates the pinhole passed spectra into the PCM-401 FO.

The tube is mounted in a Newport XYZ to perfectly align the pinhole X,Y and ~50mm Z to the PB.

Google AI calcs below:

​

              +--------------------------+
PARABOLIC     ¦ ALUMINIUM ENCLOSURE BOX  ¦
AL MIRROR     ¦ +----------------------+ ¦
  IN -----------> Pellin-Broca Prism   ¦ ¦
              ¦ +----------------------+ ¦
              ¦            ¦             ¦
              +------------|-------------+
                           |

                 PARABOLIC |

                 AL MIRROR | OUT

                           v

50mm below PB:
197nm is 0.21mm from 200nm vs 1000nm is 0.04mm from 1064nm
distance between 200nm and 1000nm is 5.06mm
distance between 197nm and 1064nm is 5.31mm

 

Thorlabs tube mounted on Newport XYZ stage:

​

 <-- X,Y-AXES -->          ¦    ^
Wavelength Tuning          ¦    ¦ Y-AXIS
197nm & 1064nm Rejection   |    v Focal Length ~50mm to match pinhole
                           v
    +----------------------------------------------+
    ¦ Thorlabs 1" dia Extension Barrel Tube        ¦
    ¦                                              ¦
    ¦ +------------------------------------------+ ¦
    ¦ ¦ 50mm Empty Air Space / Spacers           ¦ ¦ <-- PB wavelengths expand
    ¦ +------------------------------------------+ ¦     across 5.06mm
    ¦                      ¦                       ¦
    ¦                      v                       ¦
    ¦   Thorlabs SM1BAP Sleeve around pinhole      ¦ <-- Shims out the tube's 1.035" internal gap

    ¦ +------------------------------------------+ ¦
    ¦ ¦ 1" Dia 5mm pinhole + carbon trap         ¦ ¦ <-- 197nm & 1064nm fall into carbon foam:
    ¦ +------------------------------------------+ ¦     pinhole passes spectral lines
    ¦                      ¦                       ¦
    ¦                      v                       ¦
    ¦ +------------------------------------------+ ¦
    ¦ ¦ Thorlabs SM1RR Retaining Ring            ¦ ¦ <-- TOP LENS LOCKING RING
    ¦ +------------------------------------------+ ¦     Clamps outer edge of lens (spanner SPW602)

    ¦ +------------------------------------------+ ¦
    ¦ ¦ LENS: 1" Dia F=20mm (CURVED face DOWN)   ¦ ¦ <-- Focuses clean
    ¦ +------------------------------------------+ ¦     broadband LIBS pool
    ¦ +------------------------------------------+ ¦
    ¦ ¦     Thorlabs SM1RR Retaining Ring        ¦ ¦ <-- 2.0 mm SPACER SHELF
    ¦ +------------------------------------------+ ¦     (Clears lens bulge / Stops tilt)
    ¦                      ¦                       ¦
    ¦                      v                       ¦
    ¦ +------------------------------------------+ ¦
    ¦ ¦ Thorlabs SM1SMA Fiber Optic Adapter      ¦ ¦
    ¦ +------------------------------------------+ ¦
    +----------------------------------------------+
                           ¦
                           v
  [ SMA0905 to PCM-401 2m long 1mm UVFS Solarised FO ]
                           ¦
                           v
            [ PCM-401 22µm Spectrometer Slit ]

​

Thorlabs tubes

https://www.thorlabs.com/lens-tube-systems

Thorlabs spanners:

https://www.thorlabs.com/spanner-wrenches?tabName=Overview

​

 

THE UNEVEN FOCUS IS DUE TO NONLINEAR SPACING

​

A PB output has a non-linear throw with DUV lines wider apart than NIR:

197nm is 0.21mm from 200nm  vs  1000nm is 0.04mm from 1064nm

​

Doubling the PB pinhole distance to 100mm only increases NIR spacing from 0.04mm to 0.06mm

​​​

​

LIMITING RANGE TO 200nm TO 900nm 

​

The PCM-401 spectrometer range is 200nm to 1000nm however most spectra can be detected between 200nm to 900nm and most LIBS systems cap NIR to 900nm so this is viable.

​​​

​

ALTERNATIVE 1500nm REHEAT

​

Spectrosil B is limited to 2000nm; 1500nm is safely within that:

​

Replacing the 1064nm with a 3HG 355nm OPO generating 1500nm increases NIR spacing by 0.25mm:

This assumes what I was sold is actually an OPO (see above)

​​

50mm below PB:
197nm is 0.21mm from 200nm vs 1000nm is 0.29mm from 1500nm
distance between 200nm and 1000nm is 5.06mm
distance between 197nm and 1500nm is 5.56mm
 

​

LESS ENERGY

​

A down side is 1500nm needs ~2x the energy to heat the plasma than 1064nm 

(the longer the wavelength the lower the heating efficiency)
 

The unfocused 1064nm will likely need ~30-50mJ 1064nm plasma reheat

1500nm will likely need ~60-100mJ.

​

This may not be achievable given it needs SFM:(1064nm + 532nm) PLUS an OPO with both NLOs sapping power,

although the Nd:YAG will be Q-switched.

​

1500nm is a standard comms wavelength so I imagine lenses are available and one will be needed to soft focus the output of the OPO, potentially offering some thermal improvement.

​​​

However I doubt I could deflect sufficent Nd:YAG energy to make enough 1500nm

​

=============================

https://www.nist.gov/pml/atomic-spectra-database

=============================

​​​​

Another session:

​​

Alexandrite Resonator Cavity Boundary (Optical)

This is the total optical track bounded by the two mirrors.

The Back Wall:  the 1000 mm Concave High Reflector.

The Front Wall: the Flat Output Coupler.

 

 

[HR Mirror] ════════════════════════════════════════════════════ [OC Mirror]

                                                                    │

    └─> [Pockels Cell] ──> [ PUMP CHAMBER ] ──> [ BRF Pill ] ───────┘ (Rod & Lamp)

​

Position 0mm: 1-Meter Concave HR Mirror Mount >99.5%

Position 15mm to 55 mm: Inrad 202-090 Pockels Cell [2]

Position 60mm: Rear Entrance of your 117 mm Pump Chamber [1]

​​​

The rear High Reflector (HR) ideally needs to be >99.5% to 99.9% reflectivity.

​​​​​

Resonator Cavity Alignment Setup Sequence

​

 

[HeNe / Green Diode] ───> ⊕ (Pin 1) ───> [Cavity Elements] ───> ⊕ (Pin 2) │ [Returned Back-Reflection Bounce] <────────────────────────────┘

 

 

Step 1: Establish the Optical Anchor Axis

1. Mount a cheap visible alignment laser (such as a green 532 nm or red 633 nm pointer) firmly at one end of the bench.

2. Set up two fixed pinhole iris apertures or alignment targets (Pin 1 and Pin 2) spaced roughly 350 mm apart along the table.

3. Adjust the alignment pointer until its beam passes cleanly through the dead center of both pinholes. This beam profile now acts as the rigid "anchor line" for your entire laser cavity.

 

 

Step 2: Position the Pump Chamber & Optics

1. Slide your static water-cooled pump chamber block onto the track. Adjust its positioning until the alignment beam passes straight through the center of your 5 mm Alexandrite rod without clipping the edges.

2. Install your 1-meter concave rear mirror mount (holding the EKSMA HR) behind the chamber. Do not insert the Pockels cell yet.

3. Look back at the face of your alignment laser. You will see a small back-reflection dot bouncing off the concave HR surface. Adjust the kinematic screws on the HR mount until this reflection dot bounces directly back into the output aperture of the alignment laser.

 

 

Step 3: Align the Output Coupler

1. Drop your new 65% flat EKSMA Output Coupler mount into position at the front of the track.

2. Ensure Face 1 (S1, the 65% coated face) is pointed inward toward the rod.

3. Adjust the output coupler's kinematic screws until its secondary back-reflection dot matches up exactly with the HR dot, disappearing back into the alignment laser's aperture. Your core resonator cavity is now collinear.

 

 

Step 4: Inject the Intra-Cavity Elements

1. The Pockels Cell: Place the Inrad cell between the HR and the chamber. Rotate it mechanically until its internal crystal faces do not skew or step the alignment beam off target.

2. The BRF Pill: Clamp your 1.8" Spectra-Physics birefringent filter into its rotary yoke mount between the chamber and the output coupler. Tilt it precisely to Brewster’s angle (approx. 57°) relative to the track. You will know you hit the correct angle when the bright secondary reflection dots off the quartz plates vanish completely from the room.

At this stage, your optical cavity is fully stabilized and matched to your exact wavelength parameters.

​