The UNIversal Linear ACcelerator (UNILAC) can accelerate all ion species from protons to uranium to 16% of the velocity of light, i.e., 11.4 MeV/u. Charged ions are provided by three different ion source terminals. Low intensities of highly charged ions are provided by an Electron Cyclotron Resonance (ECR) source. The Penning Ionization Gauge (PIG) source delivers low intensities of low to intermediate charged ions and is the oldest type of source still being operated. High intensity beams of low charged ions are provided by the third terminal which can be equipped with a MEtal Vapor VAcuum Arc (MEVVA) source, MUlti-Cusp Ion Source (MUCIS), or Cold or HOt Reflex Discharge  ion Source (CHORDIS).

After electrostatic extraction from the source the dc-beam is bunched and pre-accelerated along a Radio-Frequency-Quadrupole (RFQ). The bunched beam is accelerated along Inter-Digital (IH) cavities to 5% of the velocity of light, i.e., 1.4 MeV/u. To enhance the efficiency of further acceleration of low-charged beams from the High-Current-Injector (HSI), the ions are passed through a gaseous medium to strip off the outer electrons thus augmenting the ion charge state. This step is not required for highly charged beams delivered by the Hoch-Ladungs-Injector (HLI).

Acceleration to the final energy is done along the post-stripper section comprising five Alvarez-type cavities. Afterwards the beams may be passed through the Transfer Channel (TK) towards the synchrotron SIS18 for further acceleration or to various experimantal branches. A unique feature of the UNILAC is that some of the Alvarez-cavities can be operated without acceleration, hence also energies as 3.6, 4.8, 5.9, and 8.6 MeV/u can be provided quickly. Finally, a section comprising several single gap resonators (ERs) allows for setting any arbitrary ion energy between 3.4 and 11.6 MeV/u of which the exact maximum and minimum value depends on the ions mass to charge ratio.

The UNILAC can be operated at a repetition rate of 50 Hz. The beam parameters as ion species, energy, intensity, and final beam target area can be switched between two subsequent beam pulses. Doing so, many different experiments requesting different types of beam can be served quasi-simultaneously.