On October 28, the Antares rocket, an unmanned launch vehicle carrying non-critical supplies to the International Space Station (ISS), crashed seconds after lift-off at the NASA Mid-Atlantic Regional Spaceport on Wallops Island, Virginia. This was the fifth Antares launch since April 2013, and its first failed launch. It was carrying its heaviest payload yet of 2,290 kg for ISS crew, including 30 small satellites, consumables and spare parts. It is estimated that Orbital Sciences, the company that owns and manages Antares, incurred a loss of $200 million because of this launch failure.
It is unclear what caused the failure. From video footage taken near the launchpad, the first stage of the rocket is seen firing as the rocket lifts off. Within the next 10 seconds, however, sparks fly out of the lower end, a column of fire engulfs the rocket and it crashes almost perpendicularly to the ground. Like the video embedded above (taken from an airborne Cessna 177) shows, there are two distinct explosions. The first one is due to the first stage exploding and the second coincides with the rocket's crash. At a press conference, it was announced that Orbital Sciences, NASA and FAA will be conducting an investigation, in which time Orbital Sciences will be 'locked out' of revealing specific details of the failure.
In the meantime, a tweet by astronomer Pamela Gay indicates the launch was deliberately aborted after lift-off (after visual confirmation of some issue).
According to a factsheet available on the Orbital Sciences website, Antares has been designed to achieve a launch reliability of 95% or higher. It'd be meaningless to assess Antares's reliability after five launches (just as meaningless as reconsidering the future of commercial spaceflight a few hours after one failure). That chart below shows how the most reliable rockets are ones that have been services for many decades. The takeaway is not that they were in service because of high reliability but the other way round: that engineers were able to perfect each design only after multiple tests and launches.
Like the chart below shows, the more reliable rockets (darker dots) are also the ones that have been in services for longer (size).
The history of spaceflight is dotted with tens of failures, some more catastrophic than the rest. Failures - including during planning, manufacturing or testing - are not uncommon because spaceflight is an inherently difficult endeavor [scroll to page 42] due to its scale and complexity. In fact, every successful launch ought to be widely celebrated because failure comes way easier, and the success is singular testimony to how many hurdles have been overcome. Even if decade-long space programs have managed to institutionalize legacy processes, every new launch is still a potential disaster. Of 6,854 launches in the history of spaceflight (1957-2013), 549 have failed. That's a big number. And while commercial spaceflight can be tracked back to the 1960s, it's become more regular and visible only in the last four years.
The October 28 Antares launch was the third in a series of eight that Orbital Sciences has to execute per an ISS-resupply agreement with NASA.According to the factsheet, Antares is a two-stage rocket with an optional third stage. The first stage, comprising two Aerojet Rocketdyne AJ26-62 engines, is powered by oxygen and kerosene. The first of two explosions visible in videos is the kerosene burning up, with the remainder oxygen dissolving into the atmosphere. The second stage is the real bad news: it consists of one ATK Castor 30B motor powered by solid fuel, which has much higher energy density and is toxic, too. Its explosion is likely to have caused significant damage to the Wallops launchpad, which is the only facility that Antares can be launched from (similarly, SpaceX launches from the Kennedy Space Center, Florida).
On October 16, Orbital Sciences announced to shareholders that it had upgraded its second stage engine (although it didn't name the new engine). However, if preliminary reports are to be believed, then the first stage Aerojet A26-62 engines are the problem. These engines are derived from the Soviet-era NK-33 engines that were used on the N-1F launcher family in the 1970s. The N-1 was abandoned after all four of its launches failed. In fact, an A26 also failed during a test in May 2014, and before that in June 2011. However, this is only speculation; let's wait for the investigation to conclude before we come to conclusions because NASA also tested many of the A26s and had approved them for use.