From ChromFAQ

How do autosamplers work?

Essentially, autosamplers are comprised of two components:

  • the injection valve
  • the sample loader.

Injection valves are generally some variant of a six-port, two-way valve. The fixed outer portion of the valve (called the "stator") has six ports connecting to the pump, the column, the sample inlet, the waste line, and to the sample loop. The loop is a section of tubing which will actually contain the sample to be injected. The rotating inner portion of the valve (called the "rotor") has three passages connection adjacent pairs of ports. The rotor can alternate between two positions, commonly referred to as the "load" and "inject" positions.

In the "load" position, the pump is connected to the column, while the sample inlet and waste lines are connected to opposite ends of the loop, allowing sample to be introduced into the loop.

When the valve is switched from the load to the inject positions, the pump is connected to the loop, allowing the flow to push the sample contained in the loop into the column. The sample inlet is connected to the waste line to allow flushing of the sample line.

In the simplest case, called "filled loop" injection, an excess of sample is pushed through the loop during loading, and the full loop volume is transferred to the column during injection, as shown in the figure on the right. This approach is not necessarily accurate (the loops are not made from precision-bore tubing, and actual loop-to-loop volume may vary considerably from the "nominal" volume) but does provides excellent repeatability (the exact volume of a given loop is always transferred the same way). It is also inflexible; in order to change injection volume, the loop must be replaced.
The more common approach is called "partial loop" or "partial fill" injection. This meters in a volume of sample which is smaller than the actual loop volume, as shown in the figure on the left. The volume of sample loaded is controlled by a syringe. Compared to filled loop injection, partial loop injection is typically more accurate (the syringes *are* made of precision-bore tubing) and more flexible (in principle, any volume up to the nominal loop volume can be loaded; in practice, it is usually safer to stay below about half the nominal loop volume), but not as precise.

Sample loaders are designed around one of three possible arrangements:

Push to fill autosamplers essentially automate a manual sample loading procedure. The syringe picks up the desired volume of sample from a vial in the tray, physically moves over to connect with the sample inlet port on the valve, and then pushes the sample into the loop. The advantage to this approach is minimum waste of sample. All the sample removed from the vial is transferred to the loop, except for a minimal remainder inside the valve. The disadvantage is maximum risk of sample carryover.

Pull to fill autosamplers. An alternative arrangement uses a syringe connected to what would normally be the waste port of the injector as shown in the figure to the right. Because this allows positive wash-out of the needle, it minimizes the potential for sample carryover. The disadvantage is a certain amount of wasted sample contained in the needle and the associated needle-to-valve tubing. In addition, because the sample is essentially pushed into the loop by atmospheric pressure, this type of autosampler can have problems with very viscous samples.

Integral-loop autosamplers are more complex, utilizing a movable needle which can dock with a pressure seal on the injector to become part of the sample loop. These samplers minimize both wasted sample (the entire sample is contained in the loop) and sample carryover (the entire flow path in contact with the sample is flushed by mobile phase) at the cost of more complex design and associated reliability problems (the needle must make a pressure-tight seal with the valve).