# Nikon Eclipse Ti2-E with AX R MP Confocal/Multiphoton System

## System Reference for Imaging & Microscopy Workflows

This document describes a fully motorised inverted microscope platform (Nikon Eclipse Ti2-E) coupled with an AX R multiphoton/confocal scan head, a dual-output tuneable IR laser, six-line visible laser unit, NSPARC super-resolution detector, FLIM capability, and environmental control. It is intended as a machine-readable reference for designing acquisition workflows, analysis pipelines, and experiment planning with LLM-based tools.

---

## 1. Microscope Stand

### 1.1 Main Body — Eclipse Ti2-E (MEA54000)

- Inverted configuration.
- Motorised built-in encoded Z-drive: 10 nm step (open loop), 20 nm step (closed loop).
- Motorised light-path change with positions: Left 100 %, Eyepiece 100 %, Right 100 %, Left/Eye 80/20 %.
- Built-in Bertrand lens (encoded).
- Encoded 1×/1.5× tube lens.
- Field Number 25 at left/right ports.

### 1.2 Eyepiece Assembly

| Component | Code | Description |
|---|---|---|
| S Tube | MEB52340 | Inclination 35°, interpupillary distance 50–75 mm |
| BC Port Eyepiece Tube Base | MEB55855 | 16 mm FOV camera port; accepts 0.55× C-mount directly or C-mount with LV-TV adapter; low-weight/small-footprint cameras |
| Eyepieces (×2) | MAK10110 | CFI 10×22 with dioptre adjustment |

### 1.3 Nosepiece — Perfect Focus System

| Component | Code | Description |
|---|---|---|
| PFS + Motorised Nosepiece for MP | MEP59395 | TI2-N-NDM-P Perfect Focus Unit v4 with altered notch filter for multiphoton use; two-part PFS4 for stability; integrated liquid-leak protection tray; overflow tube |

### 1.4 Stage & Accessories

| Component | Code | Description |
|---|---|---|
| Motorised Stage with Encoders | MEC56120 | XY travel: ±57 mm (X) × ±36.5 mm (Y) = 114 × 73 mm total; speed 25 mm/s; minimal step 100 nm; repeatability ±500 nm; connects directly to Ti2-E controller |
| Stage-Up Kit | MED53210 | Enables dual-layer configuration (Ti2-E only; not compatible with Ti2-A/Ti2-U) |
| Stage Joystick | MEF55706 | TI2-S-JS-SS |

### 1.5 Condenser

| Component | Code | Description |
|---|---|---|
| Motorised Condenser Turret | MEL51920 | TI2-C-TC-E |
| LWD Lens System | MEL56200 | NA 0.52, WD 30 mm, objectives 4–100× |

### 1.6 DIC Optics

| Component | Code | Description |
|---|---|---|
| DIC Slider 20×III | MBH76222 | Compatible with MRD77200 and MRD70270 |
| DIC Slider 40× I | MBH76240 | — |
| LWD Dry DIC Module | MEH51520 | TC-C-ML-N2D |
| DIC Slider 60× II | MBH76260 | — |
| Analyser Cube (Large FOV) | MEN51985 | Required for DIC with FOV > 18 mm |
| Intelligent Polariser | MEN51944 | Requires motorised or intelligent condenser |
| LWD DIC 4× Module | 1DM37150 | 1.5 mm aperture, 37 mm diameter; fits inside phase slots |
| LWD DIC 20× Aperture Module | 1DM37940 | 9.0 mm aperture, 37 mm diameter |

---

## 2. Widefield Fluorescence (Epi)

### 2.1 LED Illumination

| Component | Code | Description |
|---|---|---|
| D-LEDI System | MBF83000 | 4-channel LED: 385 nm, 475 nm, 550 nm, 621 nm; controllable via NIS-Elements or manual controller; brighter when attached directly (vs. fibre) |
| D-LEDI Ti2 Connection Cable | MXA22196 | Required for Ti2-E/A |
| AC Adapter (D-LEDI) | MQF52057 | Required power adapter |

### 2.2 Filter Turret & Filter Cube

| Component | Code | Description |
|---|---|---|
| Motorised Epi Filter Turret | MEV51030 | Holds six 32 mm or 25 mm filter cubes |
| Quad-band Filter Cube | MXR00252 | DAPI/FITC/TRITC/Cy5 BrightLine® "Full Multiband", ZERO™ Pixel Shift, LED-optimised; Exciter FF01-378/474/554/635-25; Emitter FF01-432/515/595/730-25-Z; Dichroic FF409/493/573/652-Di02-25×36-Z; 25 mm; mounted in TE2000 cube |

### 2.3 LAPP (Laser Application) Illumination Path

| Component | Code | Description |
|---|---|---|
| Fixed Main Branch | MEE54860 | For LAPP system; single module; no sub-branch; no field stop |
| Stage-Up Lens | MEV53004 | For use with LAPP + stage-up kit |
| EPI-FL Module | MEE54741 | TI2-LA-FL-3; NIR-compatible; 25 mm FOV; connects directly to D-LEDI; requires MXA22205 C-FIBA adapter for third-party fibre light sources |

### 2.4 Transmitted-Light Path

| Component | Code | Description |
|---|---|---|
| Pillar for Diascopic Illumination | MEE59925 | Holder for two fixed 45 mm filters; one position for PFS filter |
| AX LED Lamp House | MEE55855 | For use with DUT (MHE50350) |

---

## 3. Confocal / Multiphoton Scan Head

### 3.1 AX R MP Scan Head & Controller (MHA52300)

- Point-scanner with both confocal and multiphoton (MP) capability.
- 22 mm Field of View.
- **Galvano** mirror pair: high-resolution acquisitions up to **8192 × 8192**.
  - Max frame rate: 10 fps at 512 × 512 (bidirectional); 240 fps at 512 × 16 (band scan).
- **Resonant** mirror: high-throughput gentle imaging up to **2048 × 2048**.
  - Max frame rate (2K mode): 30 fps at 2048 × 512; 15 fps at 2048 × 1024; 7.5 fps at 2048 × 2048; 720 fps at 2048 × 16 (band scan).
  - Max frame rate (1K mode): 30 fps at 1024 × 512; 720 fps at 1024 × 16 (band scan).
- Scan modes: line scanning, bidirectional scanning, averaging.
- Simultaneous acquisition: max 5 channels (including diascopic detector channel).
- One VIS laser input, one IR laser input.
- First dichroic mirrors: 405/488/561/640 nm.
- Beam splitter: 20/80.
- IR composite dichroic mirror, gold mirror.
- Variable pinhole: 6–153 µm.
- Zoom: 1–1000× continuously variable.
- AI-powered acquisition parameter selection and visualisation.
- Upgradeable to: NSPARC super-resolution detector, confocal detection (with VIS laser + detectors), FLIM.

### 3.2 Scan-Head Accessories

| Component | Code | Description |
|---|---|---|
| Optional Output Laser Port | MHV56300 | 3rd outport AUX; required for third-party FLIM add-on |
| Resonant 2K Conversion Unit | MHE60310 | AD-R2K; connection board for resonant scanner up to 2048 × 2048 |
| Ti2 Adapter Set | MHV55710 | AX-TI2 |
| Ti2 Interlock Switch | MHV55730 | AX-TI2-ISW |

---

## 4. Detectors

### 4.1 Non-Descanned Detector (NDD) — Epi Direction (MHE60510)

- Collects emitted visible light in epi direction from IR-excited dyes (no confocal sectioning).
- Positioned directly after the objective for maximum light collection.
- Specific to inverted Ti2 + AXMP scan head.
- Detectable wavelength range: **400–750 nm** (1300 nm system configuration).
- Pre-equipped with **2 × GaAsP PMTs**; upgradeable to **4 channels** simultaneously (3 GaAsP + 1 multi-alkali also possible).

### 4.2 Additional NDD PMTs

| Component | Code | Qty | Description |
|---|---|---|---|
| GaAsP PMT (MP NDD) | MHE60530 | 2 | High-sensitivity, low-noise; green–orange wavelength range |

### 4.3 NDD Filter Cubes

| Filter Cube | Code | Bandpass |
|---|---|---|
| A1-F NDD 641/75 732/68 | MHE57821 | 641/75 + 732/68 |
| A1-F NDD 450/70 | MHE57801 | 450/70 |
| A1-F NDD 550/88 | MHE57811 | 550/88 |

### 4.4 Descanned Detector — DUX-VB4 4-Channel Unit (MHE60000)

- Tunable and high-sensitivity low-noise detectors.
- **4 detection paths**: 4 dyes simultaneously, up to 66 channels sequentially.
- 2 tunable detection paths + 2 filter-wheel detection paths.
- Detection range: **400–920 nm** (per-laser breakdown: 400–650 nm with IR laser, 400–750 nm with visible laser; full 400–920 nm reflects the combined operational range of the installed unit).
- **Wavelength resolution: 5 nm**; wavelength range variable in **1 nm steps** (continuously tunable, not discrete).
- **Detection width: 10 nm to 320 nm** (continuously variable per channel).
- Maximum pixel size: 8192 × 8192 (galvano); compatible with resonant scanner.
- PMT flexibility: GaAsP / IRGA / GaAs / MA.

### 4.5 DUX-VB4 Filter Cubes

| Filter Cube | Code | Bandpass / Use |
|---|---|---|
| A1-F 452/45 | MHE57011 | 452/45 |
| AX-F 700/75 | MHE57120 | Em Cy5 for DUX-VB4 |

### 4.6 Diascopic (Transmitted) Detector — AX-DUT-MP (MHE60580)

- Transmitted images with 22 mm FOV.
- Excitation laser up to IR; detection up to **920 nm**.
- Second Harmonic Generation (SHG) signal detection possible (with appropriate items).

### 4.7 DM/IR Cut Filter Kit (MHE60450)

- A1-U-DM1300 for inverted stand.

### 4.8 Confocal AX Detector PMTs

| Component | Code | Qty | Description |
|---|---|---|---|
| GaAsP PMT | MHE60220 | 3 | High-sensitivity, low-noise; green–orange range |
| Multi Alkali PMT | MHE60210 | 1 | High-sensitivity, low-noise; blue + red range |

### 4.9 First Dichroic Mirror Turret

The AX R MP scan head (MHA52300) includes a motorised first dichroic turret with multiple positions. The quotation documents the following dichroics — the first four are included with the scan head, the fifth is a separately purchased pentaband add-on:

| Position | Dichroic | Code | Laser lines reflected | Primary use |
|---|---|---|---|---|
| 1 | **Quad-band 405/488/561/640** | Included with MHA52300 | 405, 488, 561, 640 nm | **Standard confocal** — use for DAPI, GFP/AF488, TRITC/mCherry/AF568, Cy5/AF647 |
| 2 | **Pentaband 405/445/515/561/640** | MXU95086 | 405, 445, 514/515, 561, 640 nm | When using 445 nm and/or 514 nm laser lines (e.g. CFP, YFP) |
| 3 | **BS 20/80** | Included with MHA52300 | Any (reflects 20 %, transmits 80 %) | Low-reflection beamsplitter; useful for reflected-light or when maximum transmission to detector is needed |
| 4 | **IR composite dichroic** | Included with MHA52300 | IR (multiphoton wavelengths) | Two-photon / multiphoton imaging — reflects IR excitation, transmits visible emission |
| 5 | **Gold mirror** | Included with MHA52300 | IR (broadband) | Multiphoton mode — maximum IR reflection efficiency |

⚠️ **Critical for LLM workflow routing:** The dichroic position determines which laser lines can excite the sample and which emission wavelengths reach the descanned detectors. Selecting the wrong dichroic will either fail to deliver excitation light or block emission signal. The rules below must be applied when recommending a confocal acquisition configuration.

#### 4.9.1 Dichroic selection rules for LLM routing

| Laser line | Use quad-band (pos 1) | Use pentaband (pos 2) | Notes |
|---|---|---|---|
| 405 nm | ✅ | ✅ | Both dichroics reflect 405 nm |
| 445 nm | ❌ Not reflected — laser passes through | ✅ | **Only** the pentaband reflects 445 nm |
| 488 nm | ✅ | ❌ Not reflected — laser passes through | **Only** the quad-band reflects 488 nm |
| 514 nm | ❌ Not reflected — laser passes through | ✅ (as 515 nm notch) | **Only** the pentaband reflects ~514/515 nm |
| 561 nm | ✅ | ✅ | Both dichroics reflect 561 nm |
| 640 nm | ✅ | ✅ | Both dichroics reflect 640 nm |

**Key consequence:** Alexa Fluor 488, GFP, FITC, and any fluorophore excited at 488 nm **require the quad-band dichroic (position 1)**. The pentaband does not reflect 488 nm. Conversely, CFP (445 nm excitation) and YFP (514 nm excitation) **require the pentaband dichroic (position 2)**. The quad-band does not reflect 445 or 514 nm.

If an experiment requires both 488 nm and 445 nm excitation, the acquisition must be split into sequential runs with a dichroic change between them (or use the BS 20/80 at the cost of 80 % excitation loss).

For multiphoton imaging, use position 4 (IR composite) or position 5 (gold mirror). The NDD sits before the scan head and is not affected by the first dichroic — NDD-collected emission bypasses the turret entirely.

#### 4.9.2 Quad-band dichroic (405/488/561/640) — emission transmission windows (approximate)

| Band | Wavelength range (approx.) | Behaviour | Fluorophores passed |
|---|---|---|---|
| Notch 1 | ~395–415 nm | **Reflects** (405 nm) | — |
| Window 1 | ~415–475 nm | Transmits | DAPI, Hoechst, BFP emission |
| Notch 2 | ~475–500 nm | **Reflects** (488 nm) | — |
| Window 2 | ~500–550 nm | Transmits | GFP, AF488, FITC, YFP emission |
| Notch 3 | ~550–575 nm | **Reflects** (561 nm) | — |
| Window 3 | ~575–625 nm | Transmits | TRITC, mCherry, AF568 emission |
| Notch 4 | ~625–655 nm | **Reflects** (640 nm) | — |
| Window 4 | >~655 nm | Transmits | Cy5, AF647, far-red emission |

#### 4.9.3 Pentaband dichroic (405/445/515/561/640) — emission transmission windows (approximate)

| Band | Wavelength range (approx.) | Behaviour | Fluorophores passed |
|---|---|---|---|
| Notch 1 | ~395–415 nm | **Reflects** (405 nm) | — |
| Window 1 | ~415–435 nm | Transmits | DAPI, Hoechst emission (narrower window than quad-band) |
| Notch 2 | ~435–455 nm | **Reflects** (445 nm) | — |
| Window 2 | ~455–505 nm | Transmits | CFP, GFP emission (note: 488 nm excitation NOT reflected by this dichroic) |
| Notch 3 | ~505–525 nm | **Reflects** (515 nm) | — |
| Window 3 | ~525–550 nm | Transmits | YFP, AF532 emission |
| Notch 4 | ~550–575 nm | **Reflects** (561 nm) | — |
| Window 4 | ~575–625 nm | Transmits | TRITC, mCherry, AF568 emission |
| Notch 5 | ~625–655 nm | **Reflects** (640 nm) | — |
| Window 5 | >~655 nm | Transmits | Cy5, AF647, far-red emission |

⚠️ **All edge wavelengths in §4.9.2 and §4.9.3 are approximate**, derived from standard multiband dichroic design. The actual cut-on/cut-off positions of the installed dichroics may differ by several nm. For precise fluorophore planning, request the measured transmission spectra from Nikon or FILM staff, or verify on-site.

#### 4.9.4 General fluorophore selection rules

- Avoid fluorophores whose emission peak falls within ±10 nm of any reflected laser line on the active dichroic.
- For the tunable DUX-VB4, set detection windows within the transmission bands of the active dichroic.
- When using the NDD for multiphoton imaging, the first dichroic turret is bypassed — NDD notch restrictions do not apply.
- If a fluorophore's emission straddles a dichroic notch, expect signal loss in that region. Consider switching to a different dichroic position or using sequential acquisition.

---

## 5. NSPARC Super-Resolution Detector (MHE62100)

- **AX-ISM-MP NSPARC Unit MP**.
- Image Scanning Microscopy (ISM) with IR lasers.
- Resolution improvement up to **100 nm lateral**, **300 nm axial**.
- Extreme low-noise; ideal for dim samples.
- Works with resonant and galvano scanners, up to 22 mm FOV.
- Includes: main unit, electrical boards, cables, optical switching unit, collimator lens, IR cut filter.
- Standard filter positions:
  - Pos 1: Quad band 446/523/600/677 (4-dye sequential imaging).
  - Pos 2: Single BA 450/50 (Ex 405).
  - Pos 3: Single BA 525/50 (Ex 488).
  - Pos 4: Single BA 585/65 (Ex 561).
  - Pos 5: Single BA 700/75 (Ex 640).

### 5.1 NSPARC Additional Filter

| Component | Code | Description |
|---|---|---|
| Triple-band filter | MXU96226 | 445/515/561 nm |

---

## 6. Laser Sources

### 6.1 Multiphoton Laser — Spectra Physics Insight X3+ Dual (1THNEWPC)

| Parameter | Main Output | Dual Output |
|---|---|---|
| Tuning Range | 680–1300 nm | 1045 nm (fixed) |
| Average Power @ 700 nm | > 1.5 W | — |
| Average Power @ 800 nm | > 2.4 W | — |
| Average Power @ 900 nm | > 3.0 W | — |
| Average Power @ 1000 nm | > 2.7 W | > 3.5 W |
| Average Power @ 1100 nm | > 2.4 W | — |
| Average Power @ 1200 nm | > 2.1 W | — |
| Average Power @ 1300 nm | > 1.5 W | — |

### 6.2 Visible Laser Unit — LUA-S6 (MHF49020)

- Six lines: **405 / 445 / 488 / 514 / 561 / 640 nm**.
- Complex Monolithic Optic (CMO) combiner: permanent alignment, maximum throughput, no manual calibration.

### 6.3 IR Optical Path

| Component | Code | Description |
|---|---|---|
| Dual IR Alignment Unit | MHV55530 | Combines main beam (up to 1300 nm) and sub beam (up to 1040 nm); motorised shift mirror + alignment sensor; auto-aligns to AX R MP scan head |
| Incident Optical Unit (Main) | MHE61000 | Motorised focusing for IR pulsed laser up to 1300 nm into scan head |
| Incident Optical Unit (Sub) | MHE61010 | Focusing for 2nd IR pulsed laser up to 1040 nm |

### 6.4 Laser Connection & Safety

| Component | Code | Description |
|---|---|---|
| LUA–AX Connection Kit A | MHF46800 | Connects LUA to AX controller |
| Laser Indicator | MHF49210 | LU-IND; required for LUA-based system |
| Interlock Cable | MXA22194 | IL cable for Ti2 and NI-TT2 heads |
| Interlock Box | MHF49200 | LU-INTL; hub for controllers and interlocking |
| MP Interlock Switch Unit | MHF52000 | A1-MP-TI2S; for multiphoton and N-SIM on Ti2-E |
| Laser Interlock Lid | 1OKH201D | Safety switch on chamber lid; 30 mm glass window; Koehler illumination compatible |

---

## 7. FLIM — PicoQuant 2-Detector System (1PQNEWPC)

- **FLIM, FRET, and FCS** in one turn-key system.
- Customised for Nikon LSMs.
- Up to **8 detection channels**.
- Fluorescence lifetimes: **100 ps to µs**.
- Filters for: CFP, GFP, YFP, DsRed, mCherry, Alexa Fluor 647.
- Software: **SymPhoTime 64** + **NovaFLIM** (4-year licence).
- Includes installation and training.

### 7.1 FLIM Software Integration

| Component | Code | Description |
|---|---|---|
| NIS-A FLIM Module | MXS58000 | Required for confocal + PicoQuant FLIM integration in NIS-Elements |

---

## 8. Cameras

| Camera | Code | Description |
|---|---|---|
| Hamamatsu ORCA-Flash4.0 v3 | 1HMA1444 | USB option; sCMOS |
| Basler MED acA2440-75uc | MXBA0006 | Colour camera |
| USB 3.0 Cable (Basler) | MXBA0016 | MicroB 90° to A, 3 m |
| SMA-BNC Trigger Cable | 1HMATC01 | Trigger cables |
| Basler Trigger Cable | 1BA34087 | GP-I/O 6p/open, 10 m |

---

## 9. Objective Lenses

### 9.1 CFI Plan Apochromat Lambda D Series

| Magnification | Code | NA | WD (mm) | Immersion | Cover Glass | Chromatic Correction | FN | Notes |
|---|---|---|---|---|---|---|---|---|
| 4× | MRD70040 | 0.20 | 20.0 | Air | 0 / 0.17 | 405–850 nm | 25 | Part of 4×/10×/20× set (MRD70PAD); correction collar not required |
| 10× | MRD70170 | 0.45 | 4.0 | Air | 0.17 | 405–850 nm | 25 | Part of 4×/10×/20× set; DIC, POL, DF compatible |
| 20× | MRD70270 | 0.80 | 0.80 | Air | 0.17 | 405–850 nm | 25 | Part of 4×/10×/20× set; DIC, POL compatible |
| 60× Oil | MRD71670 | 1.42 | 0.15 | Oil | 0.17 | 405–850 nm | 25 | Spring loaded; DIC, POL compatible |

### 9.2 CFI Plan Apochromat Lambda S Series (Silicone Immersion)

| Magnification | Code | NA | WD (mm) | Immersion | Cover Glass | Chromatic Correction | FN | Notes |
|---|---|---|---|---|---|---|---|---|
| 25×C Sil | MRD73250 | 1.05 | 0.55 | Silicone oil | 0.17 | 405–950 nm | 25 | Correction collar; part of 25×/40× set (MRD73SIL); refractive index ~1.40 matches live cells |
| 40×C Sil | MRD73400 | 1.25 | 0.30 | Silicone oil | 0.17 | 405–950 nm | 25 | Correction collar; part of 25×/40× set; ideal for deep live-cell / tissue imaging |

### 9.3 CFI Plan Apochromat LWD Lambda S Series (Water Immersion)

| Magnification | Code | NA | WD (mm) | Immersion | Cover Glass | Chromatic Correction | FN | Notes |
|---|---|---|---|---|---|---|---|---|
| 20×C WI | MRD77205 | 0.95 | 0.93 | Water | 0.11–0.23 | 405–950 nm | 25 | Correction collar for cover glass thickness; long WD for thick samples |
| 40×C WI | MRD77410 | 1.15 | 0.60 | Water | 0.15–0.19 | 405–950 nm | 25 | Correction collar; combines high NA with long WD for deep imaging |

### 9.4 Immersion Media

| Item | Code | Description |
|---|---|---|
| Immersion Oil F | MXA22168 | 30 cc; for BF, DF, Phase, DIC, UV/Vis fluorescence, PFS, super-resolution; medium viscosity, weak odour |
| Silicone Oil | 1ME37837 | 250 mL, 500 cSt (Sigma-Aldrich 378380) |

### 9.5 Water Immersion Dispenser (MEV54006)

- Supplies pure water to the objective tip for long-term water immersion.
- Operating environment: +20 °C to +38 °C, ≤ 60 % RH (no condensation), altitude ≤ 2000 m.
- Includes 2 × nozzle assemblies (MEV54006-A002).

---

## 10. Piezo Z Stage

| Component | Code | Description |
|---|---|---|
| Mad City Labs Nano-Z500-N2 | MXMCL004 | Z-axis piezo nanopositioning; 500 µm travel range; closed-loop with Nano-Drive controller; USB control |

---

## 11. Environmental Control

| Component | Code | Description |
|---|---|---|
| BOLD LINE 3 Cage Incubator Bundle | 1OKCAG01 | Premixed gas; compatible with Nikon manual/motorised XY stages and MCL piezo; holders for multiwell plates, 35 mm petri dishes, 1″×3″ chamber slides; additional holders available separately |
| Laser Interlock Lid | 1OKH201D | Lid with safety switch (turns off laser when lifted); 30 mm central glass window; Koehler-compatible |

---

## 12. Controller & Cabling

| Component | Code | Description |
|---|---|---|
| Ti2-E Controller (TI2-CTRE) | MEF55037 | Integrated stage + LAPP controller (except H-TIRF); 4 digital I/O via S-TI2-EXT cable; up to 8 I/O with two cables |
| External Trigger Cable (×2) | MXA22149 | 4 digital I/O per cable; two connect directly to Ti2 controller |
| Daisy Cable L | MXA22147 | Connects filter turret to main body |
| C-Mount Adapter (×2) | MQD42005 | 1× C-Mount TV Adapter A |
| LV-TV Tube Adapter | MBB63435 | — |
| UPS | 1WASMUPS | APC Smart-UPS SMT1000I; 1 kVA / 670 W; 220–230 V AC; 8 × IEC 60320 C13 |
| USB 2.0 Cable A-B | MXF60001 | — |

---

## 13. Workstation & Display

### 13.1 HP Z4 G5 Workstation (MPX12007)

- CPU: Intel Xeon W3-2435, 4.30 GHz, 8 cores.
- RAM: 128 GB (4 × 32 GB) DDR5-4800 ECC.
- GPU: NVIDIA RTX 4000 Ada, 20 GB.
- Storage: 2 TB M.2 SSD (OS) + 2 × 2 TB M.2 SSD (data) + 8 TB HDD 7200 RPM.
- Network: HP 10GBase-T.
- OS: Windows 11 Pro 64 for Workstations.

### 13.2 Monitor

| Component | Code | Description |
|---|---|---|
| HP S7 Pro 734pm | MPX50030 | WQHD, 3440 × 1440 |

---

## 14. Software — NIS-Elements AR v7.01.00

### 14.0 Installed Licence & Module Summary

| Module | Code | Qty | Description |
|---|---|---|---|
| NIS-Elements C-ER | MHS51001 | 1 | Active confocal licence (see §14.1) |
| NIS-Elements AR-SP | MQS31501 | 1 | Passive analysis licence (see §14.2) |
| NIS-A General Analysis | MQS43110 | 2 | Processing/analysis for active and passive licences |
| NIS-A Upgrade to GA3 | MQS43150 | 2 | Node-based analysis engine (see §14.3) |
| NIS-A 2D/3D Deconvolution | MQS42700 | 1 | Blind and non-blind deconvolution (see §14.4) |
| NIS-A JOBS Editor | MQS43130 | 1 | Automated experiment design (see §14.5) |
| NIS-A JOBS Viewer | — | 1 | Browse/review JOBS results (see §14.5) |
| NIS-A FLIM Module | MXS58000 | 1 | Confocal + PicoQuant FLIM integration (see §14.6) |
| NIS.Ai Module | MQS43200 | 1 | AI image processing (see §14.7) |

### 14.1 NIS-Elements C-ER — Confocal Active Licence (MHS51001)

The active acquisition licence. Controls all motorised hardware on HCF4 and provides the full confocal and multiphoton acquisition UI.

**AX Confocal Control** — the primary acquisition interface for the AX R MP scan head. Documentation: [NIS-Elements AX](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/ax.html)

- Experiment presets: create, duplicate, and edit confocal/multiphoton experiment configurations with named dye selections.
- Multichannel and spectral experiments: simultaneous and sequential multi-channel acquisition; spectral (lambda) scanning with DUX-VB4 tunable detectors.
- Multiphoton experiments: DUX-VB4 spectral detection, NDD detection, and NSPARC detection — each with dedicated preset workflows.
- NDD filter change management: software-guided filter cube swaps inside the NDD unit.
- AX Pad: real-time control of scanning parameters (resolution, speed, averaging, bidirectional), channel settings (laser power, detector gain, offset), pinhole size (1 AU indicator), IR laser wavelength/power, and NSPARC.
- AX Scan Area: rectangular, ROI-based, and polyline (1DT) scan regions; shading correction.
- Device Setup: alignment tools, sampling calculator (Nyquist), pinhole calibration, piezo control, AutoSignal.ai (automated signal optimisation), AutoFocus, AX-MP IR laser settings, NSPARC configuration, and I/O trigger settings.
- Automatic correction collar: AX + Ti2-E motorised correction collar control for silicone and water immersion objectives with correction collars.

**Core ND Acquisition** — multi-dimensional acquisition engine:

- Time-lapse, multipoint (XY), Z-series, multi-channel, and large-image (tiling with auto-stitching) acquisition.
- Combined ND experiments (any combination of the above dimensions).
- Inputs/Outputs control for hardware triggering.
- Ring buffer capture for continuous streaming.
- Stimulation and photoactivation (ROI-targeted laser bleaching/activation via scan head).
- File Simulator for offline testing of acquisition sequences.

**Image Display & Analysis Core:**

- Supported formats: ND2 (native), TIFF, JPEG, PNG, BMP, AVI, and others.
- Image layers: fluorescence channels, binary layers (masks), annotations.
- Volume Viewer for 3D rendering of Z-stacks.
- Large-image display with seamless tiling navigation.
- LUT control (non-destructive brightness/contrast/gamma).
- Organizer for image archiving and metadata-based browsing.
- Image segmentation (thresholding, spot detection, binary operations).
- Measurement: manual, automated, time-based; object counting; field and object features; histogram; export to Excel/CSV/report.
- Movie creation from ND2 datasets (AVI export, Volume View rotation movies).

### 14.2 NIS-Elements AR-SP — Passive Analysis Licence (MQS31501)

An offline analysis licence. Identical analysis and display features to C-ER but **without hardware control or live acquisition**. Installed on a second seat for post-processing work.

- 3D animations in superimposed channels.
- Full image analysis toolset (counting, colocalisation, intensity quantification, object metrics).
- Object tracking (requires tracking module).
- Deconvolution (requires deconvolution module).
- Macro programming and report generator.

### 14.3 General Analysis 3 (GA3) — Node-Based Analysis Engine (MQS43150 × 2)

A visual node-graph analysis environment. Replaces traditional step-by-step macro workflows with a drag-and-drop pipeline builder. Documentation: [GA3](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/app.ga3.html)

**Node categories available:**

| Category | Capabilities |
|---|---|
| Image Processing | Local contrast, filtering, background subtraction, noise reduction |
| Image Operations | Arithmetic, logical, and channel operations |
| ND Processing & Conversions | Max/min/sum projections, Z-stack operations, channel splitting/merging |
| Segmentation | Thresholding (intensity, adaptive, Otsu), spot detection, edge detection, NIS.ai-based segmentation |
| Binary Processing | Morphological operations (erode, dilate, open, close, fill holes), 3D binary processing |
| Binary Operations | AND, OR, XOR, subtract between binary layers |
| Measurement | Object features (area, perimeter, circularity, intensity, etc.), field features, custom features |
| Data Manipulation | Table operations, filtering, sorting, grouping |
| Results & Graphs | Result tables, scatter plots, histograms, heat maps |
| Input & Output | Load images, reference images, save results |
| NIS.ai | Enhance.ai, Convert.ai, Segment.ai, Segment Objects.ai nodes (see §14.7) |
| Tracking | Object tracking nodes for time-lapse analysis (see §14.8) |

**Built-in workflows:**

- ND Processing: batch process multi-dimensional datasets.
- Object Counting: automated cell/particle counting.
- Time Measurement: intensity over time in ROIs.
- Cell Measurement: segmentation-based per-cell quantification.
- Tracking: object tracking in time-lapse data.
- Assays on Well Plate: per-well quantification with plate-map layout.
- 3D Counting and Tracking: volumetric object detection and tracking.
- Deconvolution and AI Restoration: integrated deconvolution/Enhance.ai nodes.
- Denoise: noise reduction pipelines.
- **Python Integration**: execute custom Python scripts as GA3 nodes, enabling integration with external libraries (scikit-image, cellpose, stardist, etc.).

### 14.4 2D/3D Deconvolution (MQS42700)

Documentation: [Deconvolution](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/app.deconvolution.html)

- **Blind deconvolution**: Iterative Asymmetric Blind Deconvolution — PSF estimated iteratively; no prior PSF measurement needed.
- **Non-blind deconvolution**: Iterative Constrained Tikhonov–Müller algorithm — requires known PSF (calculated from microscope parameters or measured with sub-resolution beads).
- **Richardson-Lucy**: iterative likelihood-based method; recommended first choice for most confocal and NSPARC data.
- **Landweber**: iterative gradient-descent method; alternative to Richardson-Lucy.
- **MAP / MAP Automatic** (v7): maximum a posteriori deconvolution; CUDA-accelerated on the installed RTX 4000 Ada GPU.
- **Automatic**: auto-selects the best algorithm (typically Richardson-Lucy).
- Operates on individual 2D images or full Z-stacks.
- Z-stack deconvolution incorporates information from neighbouring planes.
- **Live Denoise & Deconvolution**: GPU-accelerated real-time processing on the live confocal feed.
- Also available as a GA3 node for pipeline integration.

### 14.5 JOBS and HCA — Automated Experiment Design

Documentation: [JOBS and HCA](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/ug.jobs.html)

The JOBS module is a visual-programming environment for designing automated acquisition and analysis experiments. It consists of three components:

**JOBS Editor** (MQS43130) — drag-and-drop experiment builder with the following task categories:

| Task Category | Available Tasks |
|---|---|
| Sample Definition | Well Plates (6 to 1536-well), Sample Holder (slides, dishes), custom geometries |
| Spatial | Stage XY Points (multipoint lists), Large Images (tiling grids), ROIs |
| Temporal | Time Series (interval, duration, count) |
| Z-Axis | Z-Stack (range, step, relative/absolute), focus surface |
| Focus | Autofocus, Focus Surface, PFS (engage/disengage/offset) |
| Acquisition | Capture (camera/confocal), Optical Configuration switching |
| Stimulation | Photobleaching, photoactivation via scan head |
| Logic | Conditions (if/then/else based on analysis results or expressions), Expressions (arithmetic/logical), Questions (user input prompts during run) |
| Analysis | GA3 Processing (run GA3 pipelines inline), inline measurement |
| System | Wait, email/SMS notifications (SMTP), metadata labelling |
| Device Control | Direct hardware commands (nosepiece, filter turret, laser, stage) |

- Job Database: structured storage of all runs, results, and images.
- Labels and Metadata: per-well/per-position metadata (treatment labels, concentrations) — importable from spreadsheets.
- Templates: pre-built job definitions for common experiments.
- Advanced import: batch import of external image files into JOBS analysis.

**JOBS Wizard** — guided setup for routine experiments (well-plate selection, optical configuration, Z-stack, autofocus, labelling).

**JOBS Viewer** — browse and review results: thumbnail navigation, inline analysis display, Volume/Tile/Slice views, conditional workflow traceability, JOBS database retrieval.

**Workflow relevance for LLM-based experiment design:** The JOBS Editor's task graph maps to a directed acyclic graph (DAG). An LLM can propose JOBS workflows as ordered task sequences with branching conditions.

### 14.6 FRET Module

Documentation: [FRET](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/app.fret.html)

- FRET image capture and calibration.
- FRET methods: sensitised emission, acceptor photobleaching, and custom equations.
- FRET image and FRET view creation.
- Ca²⁺ ion concentration measurement (ratiometric calcium imaging).
- Ratio experiments with calibration curves.
- Custom equations for user-defined ratiometric or algebraic channel operations.

### 14.7 NIS.ai — AI Image Processing (MQS43200)

Documentation: [NIS.ai](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/nis.ai.html) · [NIS.ai Menu Reference](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/nis.ai.menu.html)

Based on U-Net deep learning architecture. Users train custom models on their own data or use pre-trained networks. NIS-Elements v7.01 includes several functions beyond the original Enhance/Convert/Segment set.

#### 14.7.1 Core NIS.ai Functions (trainable)

| Function | Input | Output | Use case |
|---|---|---|---|
| **Enhance.ai** | Noisy / low-SNR image | Denoised / enhanced image | Reduce exposure time, extend cell viability, improve confocal Z-stacks |
| **Convert.ai** | Image in one modality | Predicted image in another modality | Brightfield → DAPI + FITC prediction; DIC → fluorescence; reduce labelling |
| **Segment.ai** | Fluorescence or brightfield image | Binary segmentation mask | Semantic segmentation of complex morphologies, label-free samples |
| **Segment Objects.ai** | Image with touching/overlapping objects | Instance segmentation (individual object IDs) | Counting confluent cells, nuclei in clusters |
| **3D Enhance.ai** | 3D Z-stack | Denoised 3D volume | Z-stack restoration with 3D context |
| **3D Convert.ai** | 3D modality A | 3D modality B | Volumetric modality transfer |
| **3D Segment.ai** | 3D Z-stack | 3D binary volume | Volumetric semantic segmentation |
| **3D Segment Objects.ai** | 3D Z-stack with touching objects | 3D instance segmentation | 3D cell/organoid counting |

Training: M → N channel mapping; train on multiple files; supports RGB. Inference uses TensorRT (~40× vs CPU); training uses LibTorch (PyTorch C++ backend). All functions available as GA3 nodes.

#### 14.7.2 Pre-trained Denoise & De-blur Functions (no training needed)

| Function | Description | When to use | Notes |
|---|---|---|---|
| **Denoise.ai** | Deep-learning shot-noise removal using a CNN trained on thousands of confocal (resonant + galvano) and widefield images. Operates in real time on GPU. | Noisy single frames; reduce exposure/averaging without losing structure. Best for static scenes (moving objects may blur). | Requires spatially uncorrelated noise. NOT compatible with Nikon Qi2 sensor (not installed on HCF4). Can run live on the AX Pad or post-acquisition. |
| **ND Denoise.ai** (v7) | Z-stack and time-lapse specific denoising. Analyses up to 60 neighbouring frames (30 before + 30 after the current frame) for context-aware noise reduction. | Confocal Z-stacks and time-lapse datasets with ≥10 frames. Preferred over single-frame Denoise.ai for Z-stacks. | Two modes: **Correlated Noise** (removes AX resonant scanner line noise) and **Uncorrelated Noise** (standard point noise). "Automatically" mode selects by analysing the image. Edge frames may have slightly reduced denoising quality. |
| **Clarify.ai** (v7) | Neural-network removal of out-of-focus blur. Parameterless — no tuning required. Does not increase resolution and does not denoise (combine with Denoise.ai for noisy images). | Widefield images of thick samples; under-sampled images where deconvolution would introduce artefacts. Preferred over deconvolution for under-sampled data. | Requires the Deconvolution module (installed: MQS42700). Uses valid image metadata (modality, NA, magnification, calibration). For well-sampled images, use standard deconvolution instead. |

#### 14.7.3 Pre-trained Detection Networks (Segment Objects.ai)

NIS-Elements v7.01 ships with pre-trained networks that work immediately without user training:

| Network file | Use case |
|---|---|
| `10x_nuclei_fl.oai` | Nuclei detection at 10× magnification (fluorescence) |
| `20x_nuclei_fl.oai` | Nuclei detection at 20× magnification (fluorescence) |
| `10x_nuclei_fl_apoptosis_cytotoxicity_all.oai` | Detection of all cells for apoptosis/cytotoxicity assays (10×) |
| `10x_nuclei_fl_apoptosis_cytotoxicity_dead.oai` | Detection of apoptotic and dead nuclei only (10×) |

#### 14.7.4 Segmentation Editor (v7)

A dedicated ground-truth annotation tool for preparing NIS.ai training data:

- Binary editor toolbar with drawing tools (red) and erasing tools (blue); Tab key toggles between modes.
- Multiple binary layers per image; drag-and-drop layer reordering.
- Tools: point, freehand, polygon, rectangle drawing; separate, fill holes, clean, smooth, invert operations.
- Adjustable pen size; custom tool presets (importable/exportable as `.msegpreset` files).
- Cutout area support: exclude regions from training by masking with a binary layer.

#### 14.7.5 NIS.ai Explorer & Training Infrastructure

- NIS.ai Explorer: catalogue of all trained networks (local and cluster); queue management for batch training.
- **Cluster training**: distribute training jobs to a network of workstations via HTCondor (requires Compute Cluster Support module — availability on HCF4 to be confirmed). Image files must be on shared storage.
- GPU requirement: NVIDIA RTX 4000 Ada (installed) — fully compatible. Not compatible with Quadro K/M/P or GeForce GTX or earlier.

### 14.8 Tracking

Documentation: [Tracking](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/howto.tracking.html)

- **ROI Tracking**: track intensity/position of user-defined ROIs over time.
- **Binary Tracking**: track segmented objects across time-lapse frames (frame-to-frame linking algorithm).
- **Tracking 3D**: volumetric object tracking in XYZT datasets.
- **Single Particle Tracking (SPT)**: sub-pixel localisation and trajectory analysis for diffusing molecules.
- **Intra-Nuclear SPT (I-SPT)**: specialised single-particle tracking within nuclei.
- **Real-Time Tracking Using XY Stage**: live object tracking with motorised stage following.
- Tracking features: velocity, acceleration, displacement, heading, mean squared displacement (MSD), diffusion coefficients.
- Visualisation: track overlays, colour-coded trajectories, kymographs.
- Also available as GA3 nodes for pipeline integration.

### 14.9 Smart Experiment

Documentation: [Smart Experiment](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/app.se.html)

- Simplified assay-based experiment runner — users select a pre-configured assay and run it without exposure to the full NIS-Elements interface.
- Per-assay settings (exposure, channels, Z-stack, tiling).
- Results browsing with built-in gallery.
- User management with role-based access (admin / operator / viewer).
- Designed for multi-user core facilities where non-expert users run standardised protocols.

### 14.10 Slide Scanning

Documentation: [Slide Scanning](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/app.slides.html)

- Automated whole-slide scanning for brightfield and fluorescence.
- AI-driven region-of-interest detection and label reading (OCR).
- Four-step workflow: Slide Selection → Edit Regions → Region Settings → Final Inspection.
- Smart Live tool: side-by-side comparison of scanned image with live sample for quality control.
- Gallery: cascading tree structure for organising scan results.
- Multi-user support with user/group management.
- Compatible with Nikon slide holders and motorised stage.

### 14.11 Macro Programming & Python Integration

Documentation: [Macro Functions](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/fr.html) · [Command Reference](https://www.nisoftware.net/NikonSaleApplication/Help/Docs-AR/eng_ar/Command_Reference.html)

**NIS-Elements Macro Language** — a built-in scripting language for automating any NIS-Elements operation:

| Macro Category | Scope |
|---|---|
| Acquire | Camera properties, capture, experiment, large image, multichannel, ROI, scan area, stimulation, optical configuration, triggered acquisition, AVI |
| Device | Confocal AX control, laser output, filter, nosepiece, condenser, PFS, XY/XYZ/Z stage, piezo, zoom, correction collar, water immersion dispenser, incubator, shutter, TTL I/O, TCSPC (FLIM: PicoQuant, Becker & Hickl) |
| Analysis | Object count, colocalization, pixel classifier, object classifier, intensity profile, time measurement, ROI analysis, GPA |
| Binary | Threshold, morphology (erode/dilate/open/close), advanced morphology, 3D binary, layer operations |
| Image | Adjust, channels, convert, deconvolution, NIS.ai, preprocess, registration, scale, sequences |
| ND | Experiment, kymograph, multichannel, multipoint, projections, stimulation, time-lapse, Z-series |
| Measure | Manual, automated, field, object, histogram, custom features, export |
| Database | Fields, filters, records, tables, ODBC |
| JOBS/HCA | Job execution, HCA analysis |
| System | File operations, wait, timer, streaming, interpreter |
| Advanced API | Dialogs, math, memory, paths, strings, windows, COM port, sockets |

- Macro recorder for capturing UI actions as script.
- Macro editor with syntax highlighting.
- **Python integration**: GA3 nodes can execute Python scripts, enabling access to external libraries (NumPy, scikit-image, cellpose, stardist, pandas, etc.) within NIS-Elements pipelines.

### 14.12 Developer Resources & Open-Source Tools

NIS-Elements is developed by [Laboratory Imaging s.r.o.](https://www.laboratory-imaging.com/) (LIM, Prague). LIM maintains a public GitHub organisation with open-source tools, example recipes, and an API for programmatic control:

**GitHub: [github.com/Laboratory-Imaging](https://github.com/Laboratory-Imaging)**

| Repository | Description | Language | Relevance to HCF4 |
|---|---|---|---|
| [JOBS-examples](https://github.com/Laboratory-Imaging/JOBS-examples) | Example JOBS recipes for automated acquisition workflows | HTML | Ready-to-import JOBS templates; study these to design custom screening/time-lapse/tiling workflows |
| [GA3-examples](https://github.com/Laboratory-Imaging/GA3-examples) | Example GA3 recipes for complex image analysis pipelines | HTML | Ready-to-import GA3 analysis pipelines (counting, segmentation, tracking) |
| [limnd2](https://github.com/Laboratory-Imaging/limnd2) | ND2 file reader and writer — Python SDK | Python (MIT) | Read/write ND2 files outside NIS-Elements; integrate with Python analysis pipelines (Fiji, napari, scikit-image) |
| [documents](https://github.com/Laboratory-Imaging/documents) | Acquisition guides for NIS-Elements | CSS/HTML | Step-by-step acquisition guides for confocal, tiling, time-lapse, and multiphoton workflows |
| [scheduler-api](https://github.com/Laboratory-Imaging/scheduler-api) | HTTP API for running and managing automated NIS-Elements tasks (JOBS, GA3) | — | Programmatic remote execution of JOBS and GA3 tasks; enables integration with LIMS, scheduling systems, or custom scripts |
| [aria-sdk](https://github.com/Laboratory-Imaging/aria-sdk) | Fluigent Aria SDK fork adapted for NIS-Elements embedded Python | Python | Microfluidics integration — control Fluigent Aria pumps from within NIS-Elements JOBS/macros |

**ND2 Viewer** — a browser-based viewer for ND2 files: [laboratory-imaging.github.io/Nd2Viewer](https://laboratory-imaging.github.io/Nd2Viewer/index.html)

**Nikon e-Learning** — [register for the Nikon Customer Learning Center](https://www.microscope.healthcare.nikon.com/resources/e-learning) for access to video tutorials and courses, including the [full JOBS A-Z curriculum](https://training.nikoninstruments.com/).

**Coming soon from LIM:** ND2 file format specifications (will enable third-party tools to read/write ND2 natively).

### 14.13 GPU-Accelerated Functions

The installed NVIDIA RTX 4000 Ada GPU enables GPU acceleration for many NIS-Elements functions. Functions fall into two tiers:

**GPU functions (any GPU manufacturer):** Volume View, Surface View, 2D/3D Deconvolution (Automatic, Richardson-Lucy, Landweber, Blind), Live Denoise & Deconvolution, Extended Depth of Focus (EDF), HDR, GA3 3D Binary Processing (Colorize, Clean, Close, Close Holes, Dilate, Erode, Fill Holes, Open, Smooth).

**GPU + CUDA functions (NVIDIA only — available on HCF4):** Distortion Correction.ai (stitching, scanning wizard), Denoise.ai (confocal live), Auto Laser Off, Find Mode, AutoSignal.ai (uses TensorRT), MAP/MAP Automatic deconvolution (2D and 3D), all NIS.ai inference functions (Enhance.ai, Convert.ai, Segment.ai, Segment Objects.ai — 2D and 3D — via TensorRT, giving ~40× faster inference vs CPU), all NIS.ai training functions (via LibTorch — PyTorch's C++ backend), Cells.ai Criterion, Realtime Denoise.ai and Clarify.ai for Hamamatsu cameras.

Without a compatible NVIDIA GPU, CUDA functions fall back to CPU (except real-time live-feed functions, which are disabled). GPU drivers should be kept up to date.

---

## 15. Operational Procedures & Constraints

This section contains practical operating rules and parameters that affect experiment design. It is derived from Nikon training documents for HCF4.

### 15.1 Power-On and Power-Off Sequences

**Power ON (confocal/multiphoton):**
1. Turn on PC.
2. Turn on laser bed (power button on rear, then turn key).
3. Turn on AX controller — wait for indicator lights to become solid green.
4. Turn on Ti2-E microscope main body.
5. When Windows has loaded, start NIS-Elements.

**Power OFF:**
1. Close NIS-Elements.
2. Shut down PC.
3. Turn off Ti2-E microscope main body; wait for front LEDs to stop flashing, then turn off Ti2 controller.
4. Turn off AX controller (front button).
5. Turn the key on laser bed, then turn off power button on rear.

**Power ON (widefield only — no confocal):**
1. Turn on cameras (rear power button; wait for initialising LED to stop flashing).
2. Turn on D-LEDI (control pad power button).
3. Turn on Ti2 controller, then Ti2 main body (do not leave a long gap between them).
4. Turn on PC last.
5. Start NIS-Elements.

### 15.2 AX Pad — Key Acquisition Controls

| Control | Function | Notes |
|---|---|---|
| Scanner | Resonant / Galvano | Resonant: fixed pixel dwell time; Galvano: adjustable dwell time (1.9–83 µs range) |
| Averaging | 1, 2, 4, 8, 16 | Can be switched to Integration in Device Setup |
| Channel Series | Fastest / Balanced / Minimal Crosstalk | See §15.5 |
| AutoSignal.ai | Automatic laser power + gain | Sets optimal signal for your sample type; uses TensorRT on GPU |
| Denoise.ai | Live AI denoising | Removes shot noise from resonant images; quality indicator must be amber or green |
| Auto Laser Off | Laser blanking | Turns laser off when no FOV/Z change detected in live mode; prevents unnecessary bleaching |
| Pinhole | 1.0 AU / 0.7 AU / custom | 1.0 AU = optimal confocal; 0.7 AU = enhanced resolution; adjustable per channel |

### 15.3 NSPARC Imaging Protocol

⚠️ **Alignment must be performed before every important NSPARC acquisition.**

**Alignment procedure:**
- Use a sample with defined structure and good signal filling the FOV (avoid too much black background).
- Sample must be precisely in focus.
- Recommended settings: zoom 6, 561 nm laser.
- NSPARC alignment does NOT work with the 405 nm laser / blue channel.
- Auto-alignment works in both CF and SR mode; CF mode sometimes has a higher success rate.
- If signal is too low: increase laser power without excessive bleaching; open virtual pinhole; zoom in/out.
- When the alignment indicator turns green, acquisition is safe.

**Oversampling XY:**
- The blue square in the scan area window indicates the diffraction-limited pixel size. Zoom in until inside the blue square to break the diffraction limit.
- Recommended pixel size: **≤ 50 nm** with 60× or 100× high NA objectives.
- Pixel size should be **2.3× smaller** than the expected resolution after reassignment and deconvolution.
- Pixel density adjustable: 512×512, 1024×1024, 2048×2048.

**Photon budget (brightness):**
- Intensity in the super-resolved region must stay **below the pile-up limit** of the NSPARC detector. Exceeding this makes reassignment imperfect and data non-quantifiable.

| Scanner | Pile-up limit formula / guideline | Recommended photons |
|---|---|---|
| Galvano | 2091 × pixel dwell time (µs) × number of integrations | 300–400 photons; minimum 50 for reassignment |
| Resonant 1K×1K | ~15–20 photons without integration | Use integration to reach ≥ 50 |
| Resonant 512×512 | ~30–40 photons without integration | Use integration to reach ≥ 50 |
| Resonant 256×256 | ~60–80 photons without integration | Use integration to reach ≥ 50 |

- Galvano mode: saturation indicator shows pile-up pixels in false colour.
- Resonant mode: saturation indicator does NOT show pile-up.
- Virtual pinhole can be opened to collect more photons (reduces XY and Z resolution).

**Oversampling Z:**
- Z-step size: **120–130 nm** for 60× and 100× objectives (2.3× smaller than expected axial resolution).
- Required for proper deconvolution.

**Reassignment:** Use SR NSPARC mode.

**Deconvolution settings for NSPARC data:**
- Apply 3D deconvolution.
- Modality: **NSPARC SR Mode**.
- Pinhole size: **1 AU** (if virtual pinhole was 1 AU during acquisition).
- Noise level: **Clear** or **Low Noise** (test different levels if artefacts appear).
- Iterations: **< 10** with Automatic Stopping.
- Recommended algorithms: **Richardson-Lucy** (first choice), Automatic, or Blind.
- Weak out-of-focus signal may cause deconvolution artefacts — limit Z-stack range or use single-band emission filters to reduce crosstalk.

### 15.4 Confocal Experiment Setup — Preset Types

Three preset categories are available in the Experiment Settings window:

| Category | Use case | Channels | Flexibility |
|---|---|---|---|
| Multi-Channel | Simple fluorescence experiments | Up to 4 dyes + TD | Dye-based wizard auto-selects lasers and detectors |
| Advanced | Complex experiments (>4 channels, custom laser-detector pairings) | Unlimited (with multiple passes) | Full manual control of every laser, detector, and emission window |
| Spectral | Lambda scanning with DUX-VB tunable detectors | Tunable | Sequence or Cascade excitation modes |

The Experiment Setup uses the NIS-Elements fluorophore database to auto-assign laser lines, dichroic positions, and DUX-VB emission bandwidths. No Optical Configurations are required for the AX.

### 15.5 Channel Series & Multiple Pass Modes

**Channel Series** (linewise switching — within a single pass):

| Mode | Behaviour | Speed impact |
|---|---|---|
| **Fastest** | All channels scanned simultaneously; each line scanned 1×. | No speed penalty |
| **Balanced** | Channels grouped in pairs with sufficient spectral gap. | Frame rate reduced by number of line-groups |
| **Minimal Crosstalk** | Each channel scanned separately (linewise switching from longest to shortest emission). | Frame rate reduced by number of channels |

**Multiple Passes** (framewise switching — when >4 channels or detector reuse needed):

| Mode | Behaviour |
|---|---|
| **Simultaneous** | All passes acquired at once (if hardware permits) |
| **Fast Sequential** | Channels grouped into fewest passes possible; GaAsP detectors reused with different tunable emission ranges between passes |
| **Quality Sequential** | Each channel gets its own pass, using the same detector for consistent calibration; maximum crosstalk rejection |

Channel Series indicated by Arabic numbers in the spectral graph; Passes indicated by Roman numerals.

### 15.6 Stimulation and FRAP

⚠️ **On HCF4, FRAP and photostimulation are only possible as Sequential Stimulation in Galvano Scanning Mode.** The AX R does not have a hybrid scanner. Simultaneous stimulation (using the resonant scanner) would require an OMS scanner attached to LAPP ports with dedicated stimulation lasers — this is not installed on HCF4.

**Stimulation settings:**
- Configured in the AX Stimulation dockable window.
- Per-configuration parameters: laser line, laser power (%), pixel dwell time (µs).
- Unlimited stimulation configurations can be stored (no 3-config restriction on AX).
- ROIs are drawn in Live/Frozen window and assigned to stimulation configurations via right-click → "Assign to AX".
- PMT gain behaviour during stimulation: **"Keep Gain"** recommended for sequential stimulation.

**FRAP workflow:**
1. Acquisition phase (pre-bleach time series).
2. Bleaching/Stimulation phase(s) — one or more, each assigned a stimulation configuration.
3. Acquisition phase (post-bleach recovery time series).

**FRAP analysis:**
- Time Measurement with Background ROI (label "B") and Reference ROI (label "R") for photobleaching correction.
- FRAP button calculates half-maximal recovery time automatically.

### 15.7 Large-Image Stitching — Practical Rules

**Two stitching approaches:**
1. **Multipoint → post-acquisition stitch** (ND Acquisition or JOBS → stitch later via ND Processing, GA3, or JOBS right-click). Preserves raw unstitched tiles; allows testing different stitching methods.
2. **Large Image on-the-fly** (ND Acquisition or JOBS Large Image commands). Simpler but raw tiles may not be preserved.

**Stitching methods:**

| Method | Behaviour | Best for |
|---|---|---|
| **Optimal Path** | Finds least-difference contour in overlap region | Most samples (try first) |
| **Blending** | Averages overlapping pixel intensities | Homogeneous signal with little background |

**Image Registration (Precise Stitching):** Refines XY tile alignment beyond stage coordinates. Adds processing time but improves results. In JOBS, select the channel with the best contrast for registration.

**Overlap:** Default 10–15 %. Smaller overlap = less precise registration but fewer bleaching artefacts in overlap zone. 5 % acceptable if structures are well defined.

**3D stitching constraints:**
- Registration is done on ONE plane and applied to the full Z-stack.
- GA3/JOBS stitching uses EDF of the Z-stack for the registration map.
- ND Acquisition Large Image creates the registration map from the first Z-plane — ensure this plane has sufficient signal.
- For large 3D Z-stacks: acquire Z first, then XY (faster; stitching done post-acquisition).
- For 3D time-lapse: acquire XY first, then Z (reduces movement artefacts; only for slow dynamics).

**System preparation for stitching:**
- Warm up the AX controller for **≥ 3 hours** before stitching experiments.
- Clean the objective lens and coverslip (100 % ethanol or petroleum ether).
- Level the sample holder using PFS (tolerance ±3 µm).
- Re-calibrate objective after any exchange or cleaning.
- Move PFS dichroic out of the optical path (reduces shading).
- Remove DIC prism if not needed.
- Close condenser shutter to block LED autofluorescence; dim room lights.

**Resonant scanner stitching:** Acquire raw (un-denoised) images; apply Denoise.ai **after** stitching, not before.

**Immersion objectives:** May cause image distortion depending on immersion medium viscosity. Slow down XY stage speed. Water-dipping objectives may cause "wobbling" at final stage position — add a 1 s macro wait command before each tile acquisition.

**If stitching artefacts persist:** Zoom in slightly (1.4×) to use only the centre of the scan field.

**Shading correction:**
- Automatic (default in JOBS/GA3): computes correction image per channel. Fluorescence uses MinIP + filtering; brightfield uses MaxIP + filtering; DIC/PhC uses average + filtering.
- AX Shading Correction: pre-computed, hardware-specific, dedicated to certain objective lenses. Configurable in Device Setup → Scan Area.
- Manual: capture ≥ 10 reference images without sample at different positions.

---

## 16. Capability Summary

This system supports the following imaging modalities and workflows:

| Modality | Key Components |
|---|---|
| **Widefield fluorescence** | D-LEDI (4-channel LED), quad-band filter cube, ORCA-Flash4.0 v3 |
| **DIC / Phase contrast** | DIC sliders (20×/40×/60×), analyser cube, intelligent polariser, LWD condenser |
| **Laser-scanning confocal** | AX R scan head (galvano up to 8192²), LUA-S6 six-line VIS laser, DUX-VB4 4-ch descanned detector, GaAsP/MA PMTs |
| **Resonant confocal** | Resonant mirror (up to 2048²), AD-R2K conversion unit |
| **Multiphoton (2P/3P)** | Insight X3+ dual laser (680–1300 nm + 1045 nm fixed), NDD with GaAsP PMTs (up to 4 ch), diascopic detector |
| **Super-resolution (ISM)** | NSPARC unit (100 nm lateral / 300 nm axial), works with galvano and resonant scanners |
| **FLIM / FRET / FCS** | PicoQuant 2-detector system (SymPhoTime 64, NovaFLIM), NIS-A FLIM module, AUX laser port |
| **SHG (Second Harmonic)** | Diascopic detector (AX-DUT-MP), IR laser excitation |
| **Live-cell imaging** | PFS v4 (MP-optimised), BOLD LINE 3 cage incubator, water immersion dispenser, piezo Z stage (500 µm) |
| **Tiling / large-area** | Motorised encoded XY stage (100 nm step), NIS-Elements stitching |
| **AI-assisted analysis** | NIS.Ai (Enhance.ai, Convert.ai, Segment.ai), AI-powered scan-head parameters |

---

## 17. Excitation / Detection Quick-Reference

### 17.1 Available Laser Lines

| Source | Lines (nm) |
|---|---|
| LUA-S6 (VIS) | 405, 445, 488, 514, 561, 640 |
| Insight X3+ Main | 680–1300 (tuneable) |
| Insight X3+ Dual | 1045 (fixed) |

### 17.2 Widefield LED Channels

| Channel | Wavelength (nm) |
|---|---|
| 1 | 385 |
| 2 | 475 |
| 3 | 550 |
| 4 | 621 |

### 17.3 Detector Spectral Coverage

| Detector | Range | Type |
|---|---|---|
| DUX-VB4 (descanned) | 400–920 nm (400–650 with IR / 400–750 with VIS) | Tunable (5 nm resolution, 1 nm steps) + filter wheels; detection width 10–320 nm; GaAsP/IRGA/GaAs/MA selectable |
| NDD Epi (multiphoton) | 400–750 nm (1300 system; filter-defined) | GaAsP PMTs; up to 4 ch |
| DUT-MP (diascopic) | Up to 920 nm | Transmitted; SHG-capable |
| NSPARC (super-res) | Filter-defined (see §5) | ISM detector |
| PicoQuant FLIM | Filter-defined | Up to 8 ch; 100 ps – µs lifetime range |
| ORCA-Flash4.0 v3 | Camera-dependent | sCMOS widefield |
| Basler MED acA2440-75uc | Camera-dependent | Colour; widefield/navigation |

---

## 18. Key Specifications at a Glance

| Parameter | Value |
|---|---|
| Z-drive resolution (closed loop) | 20 nm |
| Z-drive resolution (open loop) | 10 nm |
| XY stage travel range | 114 × 73 mm (±57 × ±36.5 mm) |
| XY stage minimal step | 100 nm |
| XY stage repeatability | ±500 nm |
| XY stage speed | 25 mm/s |
| Piezo Z range | 500 µm |
| Confocal max pixel size (galvano) | 8192 × 8192 |
| Confocal max pixel size (resonant) | 2048 × 2048 |
| Galvano max frame rate | 10 fps (512 × 512); 240 fps (512 × 16 band scan) |
| Resonant max frame rate (2K) | 7.5 fps (2048 × 2048); 30 fps (2048 × 512); 720 fps (2048 × 16 band scan) |
| Pinhole range | 6–153 µm |
| Zoom range | 1–1000× continuous |
| Simultaneous channels | 5 (including diascopic) |
| DUX-VB4 wavelength resolution | 5 nm; tunable in 1 nm steps |
| DUX-VB4 detection width | 10–320 nm (continuously variable) |
| NSPARC lateral resolution | ~100 nm |
| NSPARC axial resolution | ~300 nm |
| Scan-head FOV | 22 mm |
| MP laser tuning range | 680–1300 nm |
| MP laser fixed output | 1045 nm |
| FLIM temporal resolution | 100 ps |
| PFS version | v4 (MP-optimised) |
| Objective NA (max, oil) | 1.42 (60× Lambda D Oil) |
| Objective NA (max, water) | 1.15 (40×C WI) |
| Objective NA (max, silicone) | 1.25 (40×C Sil) |

---

*Document generated for LLM-assisted workflow design. All pricing, personal identifiers, and contact information have been removed.*