A single scaled pulsar created an 80-nanosecond pulse with 440 gigawatts of peak output power at 1.1 million volts
The Fusion Report has covered the architecture of Pacific Fusion’s pulser-based inertial fusion system, which utilizes an array of impedance-matched Marx generators (IMGs) to produce a high-current, high-voltage beam directed towards the fusion target. A full power system will utilize 156 pulsars, each one three times the length of the one shown above. But prior to building a full-scale system, Lawrence Livermore National Laboratories (LLNL) and Pacific Fusion decided to build a one-third scale pulser module and test its output capabilities. This article is about the test of the 1/3rd scale pulser, and the results of those tests.
Pacific Fusion and LLNL’s Pulser Concept
For those of you who are not familiar with Pacific Fusion’s pulser concept, it is a quintessential inertial fusion approach, but one that does not use lasers. Like the Sandia Z-Machine, it utilizes massive electrical currents to generate extreme pressures on a fuel target, resulting in fusion. The currents are produced by cascading arrays of Impedance-matched Marx Generators (IMGs), with the arrays forming 156 “pulsers”.
Each pulser consists of 32 circular stages, with each stage having 10 “bricks”; a brick consists of two matched capacitors and a high-speed gas discharge switch. The stages produce a travelling electromagnetic wave in the electromagnetic media in the center of the stages that make up each pulser. The pulsers are configured on three rings by 52 rows around the fusion chamber, where the 156 beams converge to implode the target.
Like all inertial fusion energy (IFE) system, Pacific Fusion’s machine is pulsed rather than continuous in operation. Like most IFE systems, it has a nominal pulse repetition rate of roughly 1 Hz (one pulse per second), which means that the life of each “brick” is measured by the number of “shots” that it can sustain before being replaced. With a nominal shot rate of 1 Hz, that means that a capacitor array with a lifetime of one year would have to sustain roughly 31.5 million shots before failure. Testing whether a scaled-down pulser could operate for 3,000 consecutive shots (equivalent to 1 hour of continuous operation) was the purpose of this test.
The Press Release: Lawrence Livermore-Pacific Fusion collaboration breaks new record to advance America’s high-gain fusion capability
LLNL pulsed-power prototype surpasses 3,000 shots through collaboration with Pacific Fusion, demonstrating the maturity of a key U.S. capability and the value of government-industry partnership.
FREMONT, Calif. (July 16th) — Pacific Fusion today announced that a pulsed-power prototype designed and built at Lawrence Livermore National Laboratory has surpassed 3,000 shots under a Cooperative Research and Development Agreement with Pacific Fusion, marking a key milestone in the development of high-gain fusion and a practical example of government-industry partnership at work.
The prototype, called Sirius, is an impedance-matched Marx generator (IMG), which is a new pulsed-power architecture designed to deliver short, powerful electrical pulses efficiently and repeatedly. It is the same core technology Pacific Fusion is now scaling for its own fusion system. The 3,000-shot campaign shows how technology first developed at a national laboratory can be validated through repeated testing and scaled quickly by industry to support national needs.
That matters because the United States urgently needs new sources of affordable, reliable electricity to power surging demand for AI and electrification, as well as new high-yield experimental capabilities for stockpile stewardship. Pacific Fusion is building rapidly to support both, developing a modular pulser platform to produce abundant, affordable energy and advance the high-energy-density science essential to national security.
“This is a concrete example of what government-industry collaboration can deliver when both sides are committed,” said Keith LeChien, Co-founder and Chief Technology Officer of Pacific Fusion. “We are taking innovation from the lab and moving quickly to build and scale systems that serve America’s needs. The result is a credible shot at high-yield, high-gain fusion this decade — and a path to turning U.S. scientific leadership into a critical capability that will keep the U.S. ahead of China.”
Pulsed power works by taking electricity from the wall plug, storing it briefly, and releasing it in an enormous burst — like a lightning bolt — in about 100 nanoseconds. For fusion energy, that burst can create the extreme conditions needed to compress fuel and release energy. The same pulser technology can also create extreme states of matter for high-energy-density science experiments, with applications ranging from materials testing, radioisotope production and national security.
The IMG was co-invented by LeChien and LLNL researcher Bill Stygar as a more direct way to deliver a fast, efficient electrical pulse. Conventional pulsed-power machines, called Marx generators, stack voltages. By contrast, the IMG stacks waves using a pulser by charging capacitors in parallel and then discharging them in a single step, through carefully timed stages on a common transmission line. The pulse is then transmitted directly to the target. In the Sirius campaign, each pulse of the four-stage prototype delivered 60 gigawatts to a resistive load in a 100-nanosecond pulse, with 95% energy efficiency.
Researchers at LLNL say the appeal of the IMG approach is its relative simplicity. “Conventional pulsed-power machines often require several stages of pulse compression, which can add complexity, maintenance demands and safety considerations,” said Kumar Raman, LLNL project manager in a LLNL blog.
The 3000-shot campaign supported through the Pacific Fusion CRADA focused on component lifetime and reliability – real experimental data that’s proven invaluable to understanding how components behave over many repeated shots.
It’s data that Pacific Fusion has used to design, build and scale its own pulsed-power technology at extraordinary speed. In June, the company announced the completion and validation of a pulsed-power prototype that expanded the Sirius platform by roughly 11x – delivering ~440 GW of peak output power and ~1.1 MV peak voltage in 80 nanoseconds – the highest-power, single-step pulsed-power driver ever demonstrated. Pacific Fusion is now working to demonstrate a system that’s roughly 40x the size of Sirius.
Pacific Fusion has raised more than $1 billion in private capital and is working to achieve net facility gain by 2030. Later this summer, Pacific Fusion will break ground on the world’s largest high-gain, high-yield facility called the Demonstration System in Albuquerque, New Mexico. The system is designed to produce fusion bursts exceeding 100 megajoules – or more energy output than the total stored energy used to drive the reaction, making it the first and only net facility gain fusion facility in the world.
“China is moving aggressively to build the next generation of fusion infrastructure, pouring billions into new facilities,” LeChien said. “The U.S. invented many of the breakthroughs that made this moment possible here and abroad, but to win we have to build. The government needs a significant coordinated investment in energy and national security fusion infrastructure, and months matter. The stakes could not be higher.”
Conclusion: A Great Demonstration, But Still a First Step
This was clearly an important first step towards a working production system. However, a production pulser would need to be able to operate for at least a year (~31.5M shots), roughly 10,000 times the number demonstrated here. Additionally, Pacific Fusion’s targets consists of frozen deuterium and tritium (D-T) fuel, and while D-T fusion is energetically favorable (it requires the lowest temperature of any Earth-bound fusion reaction), it has several drawbacks, including the need to be able to sustain continuous high-energy neutron damage, and to breed tritium. Many believe that these challenges are surmountable given time (all IFE systems have these challenges); now the real question is how much time is needed to do so…
