membrane-autopsy

What this skill does

# Membrane Autopsy

Reference for deciding when to autopsy a membrane element, how to ship it, what each lab method reveals, and how to translate findings into operational action or a warranty claim.

Convention markers:
- `⚠️` — likely bug or incorrect assumption in calling code
- `💡` — enhancement opportunity (autopsy-driven lifecycle update, vendor warranty workflow integration)

---

## §1 — When to Autopsy (Decision Matrix)

### 1.1 Trigger conditions (any one suffices)

| Trigger | Threshold | Notes |
|---|---|---|
| Salt rejection drop (passage rise) | > 10 % vs baseline, persistent after 2 CIPs | Avista / AWC / Vipanan canonical trigger |
| Normalized ΔP (ΔPn) rise | > 15–25 % not recovered by CIP | DuPont 45-D01652: 15 % action; many operators use 20–25 % post-CIP |
| Normalized permeate flow (NPF) decline | > 10–25 % not recovered by CIP | Biofouling signature: ΔP impacted first, then flux, then rejection |
| Unexplained instrument disagreement | Persistent after sensor calibration | Rules out membrane vs instrumentation root cause |
| End-of-warranty diagnostic | Before warranty expiry | Required to file manufacturing-defect claim |
| Premature failure pattern | Multiple elements < expected life | Especially when concentrated in lead or tail positions |
| Suspect chlorine breakthrough | Cumulative ppm·h record exists OR ORP excursion | Fujiwara test confirms halogenation |
| Pre-replacement diagnostic | Before bulk replacement decision | Validates cleaning options before procurement |

### 1.2 Cost-benefit framing

Public vendor websites do not publish autopsy price lists — all require a quote. Indicative ranges from industry practice (planning indicators, not contractual):

| Scope | Indicative range (USD) | Use case |
|---|---|---|
| Basic: visual + dye + weight-loss | ~$500–1 000 | Quick screen, single element |
| Standard (Genesys "Tier 2-3"): + SEM/EDS + FTIR + cleaning study | ~$1 500–2 500 | Most common; supports warranty claim |
| Advanced (Tier 4-5): + 16S rRNA + cross-section + Fujiwara + ATP + LOI | ~$2 500–5 000 | Complex/persistent failures, expert-witness |
| Vendor warranty autopsy (DuPont, Hydranautics, Toray) | Free, if claim accepted | RMA process; vendor-affiliated lab |

Pays off when: multiple elements need replacement, warranty claim plausible, recurring failure pattern needs root-causing.

---

## §2 — Sampling Protocol

### 2.1 Which element to autopsy

| Position | What it reveals | When to pick |
|---|---|---|
| **Lead element of lead vessel** | Particulate / colloidal fouling, biofouling, oxidation damage, organics, oil/grease | Default; fouling-dominant problems; suspected chlorine breakthrough |
| **Tail element of last vessel** | Scaling (CaCO₃, CaSO₄, BaSO₄, SiO₂), compaction | Rejection drop suspected scale-driven; LSI/recovery review |
| **Lead + tail** | Complete fouling picture | Generous budget, systemic problem |
| **Failed element (via probing)** | Localized integrity loss | Probing identifies a specific outlier |

For budget-limited cases pick **lead** unless symptoms point to tail.

### 2.2 Element preparation post-removal

1. **Do not air-dry** — desiccation alters biofilm morphology and chemistry, masks NOM, kills microbial signal
2. **Wrap in heavy plastic** (double-bag); label with element serial, position, removal date
3. **Refrigerate at 4 °C** within 24 h; ship in insulated container with cold packs
4. **Do not freeze** — ice crystals damage PA and disrupt cross-section integrity
5. **Ship within 48–72 h**
6. **Photograph** all external surfaces under good light **before** shipment (telescoping, glue line, end-cap, brine seal, fiberglass shell). Photos are essential for warranty claims

⚠️ Do not preserve the entire element in SMBS if microbial culture from the main sample is required — SMBS suppresses what you want to grow. Keep a **separate witness coupon** in 0.5–1 % SMBS if biofilm is the suspect and shipping is slow.

### 2.3 Pre-autopsy data package (ship with the element)

The lab — and, if applicable, the vendor warranty desk — both require:

- Element identifier (serial number, model, lot, manufacture date)
- Position in train (rack / vessel / position), hours on-stream
- Feed water source + quality history (TDS, SDI, T, pH, alkalinity, hardness, silica, Fe/Mn, oils, chlorine, ORP)
- Baseline NPF / NSP / ΔPn and recent normalized values
- CIP history (dates, chemistry, pH, T, contact time, recovery achieved)
- Cumulative free chlorine exposure if tracked (ppm·h)
- Abnormal event log (pressure shocks, dosing failures, dechlor failures, feed upsets)
- ORP / free-Cl₂ continuous log around the suspected exposure event
- Photographs taken at removal

⚠️ Without this data package the lab can identify the *what* but not the *why*, and warranty submissions are routinely rejected for missing pretreatment logs.

---

## §3 — Lab Method Suite

For each: what it reveals, indicative cost, limitations.

### 3.1 Visual inspection
- **External**: telescoping, fiberglass cracks, glue-line cracks, abrasion, brine-seal damage, biofilm rim on lead face, surface color (brown = Fe, white/tan = scale, dark green-brown = biofilm/NOM)
- **Internal** (after dissection): leaf-by-leaf deposit distribution, spacer compaction, permeate-spacer imprint on PA (compaction signature)
- **Cost**: included in any autopsy package
- **Limitation**: qualitative

### 3.2 Weight-loss / Loss-on-Ignition (LOI)
- **Reveals**: total foulant mass and organic vs inorganic split. Wet weight baseline → post-CIP weight (mass removed); coupon ashed at 550 °C (organic burns off, inorganic residue remains)
- **Cost**: ~$100–200 add-on
- **Limitation**: bulk number; doesn't identify foulant species

### 3.3 Cleaning effectiveness test
- **Reveals**: whether field CIP can recover the element. Bench cell applies sequential chemistries (high-pH detergent → low-pH acid → non-oxidizing biocide → reductant if needed) and re-measures flux + rejection after each step
- **Decision rule**:
  - Recovery ≥ **90 %** → CIP-treatable in field, deploy the recipe
  - **70–90 %** → marginal; consider one more aggressive CIP, then replace
  - **< 70 %** → replace; no CIP benefit
- **Cost**: ~$300–600
- **Limitation**: bench cell ≠ field hydraulics; treat as upper bound

### 3.4 Salt-rejection retest
- **Reveals**: membrane chemistry change. Standard conditions per vendor (typically 2 000 ppm NaCl, 225 psi, 15 % recovery, 25 °C for BWRO; 32 000 ppm NaCl, 800 psi for SWRO)
- **Interpretation**: rejection > 5 % below vendor spec → confirmed chemical/structural damage (oxidation, hydrolysis, compaction) → replacement
- **Cost**: ~$200–400
- **Limitation**: does not identify cause, only confirms loss

### 3.5 Dye test (Rhodamine WT or Red Dye #40 per ASTM D6908)
- **Reveals**: integrity defect location and type. ASTM D6908 lists Rhodamine WT and Red Dye #40 (both MW ~500) as soluble-dye tracers for RO/NF

| Stain pattern | Defect |
|---|---|
| Diffuse, full-membrane | Chemical attack (oxidation, hydrolysis) — confirm with Fujiwara |
| Localized point | Pinhole — hydraulic shock, particulate puncture |
| Linear / straight line | Glue-line crack — manufacturing or handling |
| Edge concentration | O-ring or brine-seal bypass |

- Methylene blue is the cheaper alternative
- **Cost**: ~$150–300 — cheapest discriminating test
- **Limitation**: confirms breach but not chemistry of attack

### 3.6 SEM / EDS
- **SEM**: high-resolution surface morphology — foulant thickness, crystal habit (rhombohedral CaCO₃, needle CaSO₄, tabular BaSO₄), biofilm 3D structure
- **EDS**: elemental composition, semi-quantitative. Map: Ca, Mg, Ba, Sr, Si, Fe, Mn, Al, S, P, C, N, Na, Cl

Cheat sheet:

| EDS dominant elements | Foulant |
|---|---|
| Ca + C (+ Mg) | CaCO₃ scaling |
| Ca + S | CaSO₄ |
| Ba + S | BaSO₄ (very low solubility — usually irreversible) |
| Sr + S | SrSO₄ |
| Si dominant | Silica (amorphous or quartz) |
| Fe (± Mn) | Iron/manganese oxide |
| Al + Si | Aluminosilica