2 amps Driver for a 100 picosec pulse

Hi all.
In order to ignite a laser I need a CURRENT driver of 2 amps for a pulsewidth of 100 pico-seconds

Doesn't sound feasible at first look, but you may want to give a complete specification anyway.
Load impedance? Rise- and falltime? Repetition rate? Jitter specification?

That's a measly 10GBPS. You might find some laser diode
drivers that claim to be up to the task (bandwidth-wise)
but that high a current is doubtful, and the trick would be
in getting all the current into the load and not the stray
capacitance, through the stray inductance.

You might have to parallel some number of drivers.

And the voltage rating is not likely to be any more than a
diode laser needs; you don't say what kind of laser.

Now, there are non-solid-state systems I've been around
that will do tens of kV and hundreds of amps in that timeframe.
But I don't think you want an arc-ignited, 30 foot long, 12
foot high tank of custom machined capacitor plates in sulfur
hexafluoride gas.

Diode lasers don't want to be "lit" every pulse; you idle them
just above lasing threshold current for "0" and bump it up as
little as you think you need, for "1". So maybe consider
a dual-fed arrangement with a choked DC feed for the lasing
baseline current, and a pulse driver for the "1" incremental
current pulse. That might relieve some of the difficulty.

Thanks dick_freebird

The laser is not for data. The pulses are generated every 1 microsecond.
Can you please specify the drivers you recomend for parallel driving.

Added after 5 minutes:

Thanks FvM

Here is the data you asked for:

Load impedance?
The laser Ron is about 2 OHM

The Rise & falltime:
About 30 pico-Seconds

Repetition rate: 1 to 10 MHz

Jitter specification?
The application is not sensitive to jitter


Thanks again

Maxim has some app notes which look like the output is
current-mode, hence gangable.

**broken link removed**

Though I suppose it's possible that they could squabble
over the exact division of current, it seems likely to work.

They indicate the norm is ~50-80mA, so that might be a lot of
ICs.

There are folks in the aerospace / defense industry doing very
high power slab lasers, like Northrop-Grumman. But I bet you
would not like the price tag, even if they would sell you any.

I think, all kinds of gas discharge switches can be excluded due to these device's low repetition rate and limited lifetime,
if they would achieve the rise time specification.

I've used silicone BJT in avalanche mode, for an about one order of magnitude slower specifcation, but there are faster transistors.

The moderate low voltage would allow to use fast unsaturated switches with state-of-the-art microwave transistors.