# Run with implicit solvent

**You will learn:** how to run a simulation where the solvent is modelled as a continuum dielectric (GBSA-OBC) rather than explicit water molecules.

**Prerequisites:**

- ACEMD installed.
- An AMBER topology (`.prmtop`) **built for implicit solvent** — no water box, no ions, and the right Generalised Born radii baked into the PRMTOP. A topology built for explicit solvent won't give correct dynamics here, even if you turn `implicitsolvent: true` on at runtime. With [HTMD](https://software.acellera.com/htmd/), {py:func}`~htmd.builder.amber.build` accepts an `igb`/implicit-solvent flag that produces the right PRMTOP; see HTMD's docs.
- A coordinate file for the solute (PDB or any moleculekit-readable format).

Implicit solvent is much cheaper per step than explicit-water simulations and is useful for small-molecule conformational sampling, peptide-folding studies, and quick equilibrations. The trade-off: no water-specific dynamics (hydrogen-bond reorganisation, water-mediated interactions) and inflated diffusion coefficients.

## Setup

```{code-block} yaml
:caption: input.yaml

structure: solute.prmtop
coordinates: solute.pdb
implicitsolvent: true
igb: 2
thermostat: true
run: 100ns
```

What's different from an explicit-water run:

- **`implicitsolvent: true`** — turns on Generalised Born implicit solvent.
- **`igb: 2`** — the implicit-solvent model. `2` is OBC1, the recommended default. Valid values: `1` (HCT), `2` (OBC1), `5` (OBC2), `7` (GBn), `8` (GBn2). See the [OpenMM AMBER implicit-solvent table](http://docs.openmm.org/latest/userguide/application/02_running_sims.html#amber-implicit-solvent).
- **No `boxsize`** — non-periodic system.
- **No `barostat`** — no box volume to regulate.
- **No `cutoff`** — ACEMD uses a 20 Å non-periodic cutoff internally for implicit solvent. The top-level `cutoff` option is ignored on this path.

## Run

```bash
acemd
```

## Gotchas

- **AMBER only.** ACEMD's implicit-solvent path is built around AMBER's Generalised Born radii. The flag isn't gated on topology type, so a CHARMM PSF or OpenMM-XML setup will *run* but produce wrong physics (no valid GB radii). Stick to a PRMTOP built for implicit solvent.
- **`barostat: true` + `implicitsolvent: true` errors out.** There's no periodic box, so the barostat setup raises at startup.
- **Inflated diffusion.** Diffusion coefficients are higher in implicit solvent because there's no viscous drag from explicit water. Useful for conformational sampling; not for kinetic / transport quantities.

## See also

- `implicitsolvent`, `igb` in [Input options](../reference/input-options.md).
- [OpenMM AMBER implicit-solvent table](http://docs.openmm.org/latest/userguide/application/02_running_sims.html#amber-implicit-solvent).