Eightfold Rebound After an SVE Pilot: Redesign or Repeat the Mistake?

The pattern - rapid early mass removal, a sharp tail, and an eight-fold rebound (18 → 150 ppmv) after two weeks of shutdown - is the signature of diffusion-limited mass transfer, not of a nearly depleted source.

Mechanistically:

• SVE sweeps only the advective pore space. In silty clay with a permeability many orders of magnitude below the sand lens, vacuum-induced flow is negligible; TCE moves through the clay only by diffusion, on a timescale of months to years over tens of centimetres.
• The pilot effectively stripped the sand lens, which supplied the 240 ppmv. Off-gas fell to 18 ppmv once the advective pores were flushed and the clay matrix became the controlling supply.
• The rebound magnitude says most of the mass accessible to that well interval was still present and re-equilibrated into the swept pores during the two-week shutdown.
• A lens concentration of 55 mg/kg in a sand is high enough to raise the possibility of residual DNAPL or a persistent sorbed source, which would extend the tail indefinitely.

Implication for design: more vacuum and more blower capacity will not shorten the tail, because the rate is set by the diffusion path length. The design levers are smaller spacing, elevated temperature, or physical removal of the lens - not a bigger blower.

Spacing is the central flaw. The measured radius of influence was 6 m at 20 inches w.c. A defensible layout makes effective influence overlap, i.e., spacing roughly 1.5 × ROI ≈ 9 m - and only after a multi-well interference test confirms that ROI in the clay matrix. A single-well ROI in a heterogeneous silt overstates the radius at which gas actually moves. Ten-metre spacing gives a 5 m radius, inside the 6 m contour, so there is no overlap where it matters: the clay between wells is never swept.

Revised concept:

1. Delineate the lens footprint (CPT/MIP plus step-out borings). Treat the lens as the source area, not the whole 0.6 ha.
2. Excavate the lens where depth and the future building footprint allow; removing the sand lens removes the resupply reservoir that drove the rebound.
3. Over the remaining source area, shorten the diffusion path: 6–8 m spacing, pulsed operation (for example 3 days on / 3 days off, or 8 h on / 4 h off) to let concentration re-equilibrate into swept pores between cycles, and consider air injection or dual-phase extraction to bring the 5.5–6 m capillary fringe into the treated zone.
4. If six months is a hard constraint, thermal treatment (electrical resistance heating to 90–100 °C) is the only technology that reliably accelerates diffusion from silty clay at 4–5.5 m depth. Budget it explicitly rather than adding wells.

Prediction: 18 wells at 10 m over six months will reproduce the pilot's tail and fail a rebound test. Make a 2–4 week shutdown-and-rebound test the design acceptance criterion, not an afterthought.

Closure on one soil-gas sample per well is not defensible. Soil gas varies by orders of magnitude over a few metres in heterogeneous silt, and a single post-shutdown sample cannot distinguish a swept pore from an equilibrating matrix. A defensible package would include: soil concentrations compared with risk-derived screening levels over a defined grid with duplicates and tracers; soil gas as one line of evidence, not the sole criterion; a documented rebound test after at least 2–4 weeks of shutdown; and a statistical basis (for example a 95% UCL over the source area) rather than single points.

Also missing: the vapour intrusion pathway. If residual TCE remains near a future occupied building, closure must address it - a sub-slab depressurisation system as an engineering control plus a deed restriction is a legitimate outcome, but it is not a no-further-action letter.

On the lender's demand: it is a request for risk transfer, not a technical requirement. Advise the client to:

• Fix the price for a defined scope (lens removal plus one operating season of SVE) and hold the diffusion-limited clay tail in a separate unit-rate or contingency line.
• Build decision gates into the schedule (day 60 rebound test, day 120 confirmation sampling) so both parties know the exit criteria.
• State clearly that the regulator issues the no-further-action determination; no consultant can guarantee it contractually.
• Offer the lender financial assurance instead - an escrow or an insurance-backed remediation warranty - which addresses the credit risk without promising an outcome the hydrogeology will not support in 12 months.