The Soviets at one time had the largest fleet of submarines in the world. In many respects, advanced Soviet submarines were superior to their US cold war counterparts. Several Soviet submarine designs could dive to a greater depth than a standard US MK 48 torpedo!
US submarines (from the arrival of Thresher) had held the advantage of stealth, and superior sonar equipment. Soviet submarines, in contrast, held the advantage of survivability (due to double-hull construction and huge amounts of reserve bouyancy), weapon-carrying capacity, greater depth and top speed.
The US experimented with a variety of reactor/propulsion designs, but only one used a liquid-metal cooled reactor:
USS Seawolf (SSN-575) was the second US submarine (after USS Nautilus), and the only US submarine to have a liquid metal cooled reactor. The reactor was cooled using liquid sodium, which of course would be problematic for the crew if it ever leaked. Seawolf also had steam superheaters, for added efficiency. These were also problematic, and thus were seldom in service. Because liquid metal is much more efficient than water at removing core heat, the propulsion plant was only 40% the size of Nautilus'. Seawolf was eventually converted to a more typical S2W pressurized water reactor (PWR) with a saturated steam plant. PWR and saturated steam plants in US submarine design continues to this day.
The Soviets' emphasis on submarine speed, depth and power of course led to more propulsion designs that used liquid metal cooled reactors. Soviet reactors of this type used a Lead-Bismuth coolant that was far less hazardous than liquid sodium, at least from a fire hazard standpoint. From a power-weight (and size) standpoint, the liquid metal cooled reactor is far superior to a light water cooled reactor. From a safety standpoint, not so great.
Recall that liquid metal cooled reactors are Fast neutron reactors, or sometimes intermediate speed reactors. All liquid metal cooled reactors have a positive void coefficient of reactivity. That means that if the coolant inadvertantly boils in the core, reactor power will increase. Which will boil more metal, and increase power even more. This happens rapidly, and core damage (meltdown!) is fairly common with this type of reactor.
So with that background, lets talk about the Soviet submarine K-27, or Projekt 645.
The Soviet's first class of nuclear attack submarines was called the November class. They used dual 70 Megawatt PWR reactors for propulsion. 13 of these were built before technology allowed creation of superior designs. Even so, they were superior to the USS Nautilus, in speed, depth, and stealth. One could also argue that Nautilus was really an experiment to prove that nuclear propulsion could work on a submarine, rather than a true nuclear attack submarine, however, and not be wrong.
Profile of a November-Class Submarine:
Back to K-27. This was a unique single-ship design by the Soviets, just as Seawolf was for the US Navy. K-27 was a November-Class submarine with a unique power plant. Rather than two 70 Megawatt PWRs, the Soviets used two VT-1 liquid metal cooled reactors, with an output of 73 MW. The advantage of smaller footprint and weight of the metal-cooled reactors allowed more weapons to be carried.
She was laid down on June 1958 and launched in April 1962. She was commissioned October 1963 after full-scale builders sea trials and official tests. She performed well (although with heavy maintenance for the new metal-cooled reactors) until a reactor accident in the port (left) reactor happened in May 1968.
The ship was making a full speed submerged run, when a reactor automatic control rod withdrew itself. Boiling occured, and reactor power plummeted from 83% to 7% in about 90 seconds, as the core melted. Unfortunately for the crew, poor decisions made after the initial accident would cost many of them their lives.
The main purpose of cladding U-235 in a reactor with Zircaloy or Stainless steel is to keep the highly radioactive freshly split atoms from getting into the coolant and spreading. When the fuel assemblies melt down, these radioactive atoms mix in the coolant, and get outside the heavily shielded reactor vessel.
Unknown to the crew, the captain had the radiation alarms disabled. Radioactive gases were released from the fuel, which the crew were exposed to. Another captain might have surfaced the ship and ventilated it with the massive air blowers all submarines are equipped with. The ship limped home on the starboard reactor and was laid up for several years. Five sailors who worked in the propulsion plant died within a week of the accident, while 30 more died between 1968 and 2003. Quite a high death rate for a crew of young, healthy men.
K-27 was brought into shipyard, and the starboard reactor coolant was kept liquid by steam piped in at the shipyard while the radioactivity in the port side reactor died down. In 1973 the decision was made that repairing or replacing the reactor in the aging ship was not worthwhile, and the ship was decomissioned in February 1979.
Her disposal was... interesting. Rather than remove the melted down mess that remained of the port side reactor, the Soviets decided to fill her reactor compartment with a solidifying agent. Next they towed her, not out to sea, but very close to land. In 1982 they sunk her in just 100 ft of water.
She didn't want to sink, however, so they ended up having to ram her.
K-27 refusing to be scuttled:
There is now a great deal of urgency in re-floating K-27 and removing her radioactive coolant system and fuel. This is an environmental hazard that will eventually become a serious problem, and quite close to shore. Interestingly there is equipment available to de-fuel this unique ship that was used on many other liquid-metal cooled ships at the end of the cold war. However, this now-unused de-fueling equipment will not remain in optimum condition forever, so the race is on...