Editor's note: The opinions expressed below are solely those of the author and not his employer.
On February 21, the U.S. Navy used a tactical SM-3 missile fired from a navy Aegis ship near Hawaii to destroy the disabled spy-satellite USA-193. The satellite contained a fuel tank filled with 454 kilograms of frozen toxic hydrazine fuel. U.S. government agencies maintained that if the satellite had been allowed to reenter the atmosphere, its fuel tank would have "survived intact," landing on the ground and venting its load of hydrazine, possibly killing or harming someone. At a briefing a week prior to the shootdown, the hydrazine was described as being "similar to chlorine or to ammonia in that when you inhale it, it affects the tissue in your lungs. It has the burning sensation. If you stay very close to it and inhale a lot of it, it could in fact be deadly."
USA-193 failed immediately after its December 2006 launch. While in orbit, it remained powerless and incommunicado. Though it was immediately clear post-launch that it would eventually reenter the atmosphere, the government didn't consider intercepting it until a year later. The public wasn't informed of the alleged danger from the hydrazine tank until February 14--a week before the interception. The operation cost taxpayers more than $100 million.
Washington didn't release a technical threat analysis or firm numbers of the risk of death or injury prior to the shootdown, feeding public suspicion that other motives inspired the test. Popular speculation included: The interception was carried out to destroy any--even minute--classified components on board the satellite and to render the debris field in a determined location over the ocean; it was a reply to China's January 2007 antisatellite (ASAT) test; it was designed to test the inherent ASAT capability of the SM-3 missile defense interceptor; or it was a way to prevent more high-resolution pictures of the classified satellite from being taken. Or, of course, all of the above.
Absent any firm technical analysis from the government, it was impossible for the general public--or even experts--to properly judge any of these speculations. According to a government official intimately involved with the risk studies, much of the reentry threat analysis inexplicably continues to remain undocumented and secret.
In the interest of informed public debate regarding the technical merit and alleged necessity of the interception, I filed a Freedom of Information Act (FOIA) request with NASA headquarters on March 25, asking for the release of any documents regarding the pre-interception reentry threat analysis of USA-193. On July 18, the request was redirected from NASA headquarters to Johnson Space Flight Center, and a few days thereafter some of the requested studies were released to me. The study relevant to the survivability of the hydrazine tank can be found here.
The roughly spherical tank contained 454 kilograms of hydrazine. It was 3.56 millimeters thick and about 0.52 meters in radius. Contrary to earlier anecdotal statements, it wasn't at full capacity: The volume of the tank was about 0.579 cubic meters of which the hydrazine occupied 0.442 cubic meters; therefore, the tank was only around 76 percent full.
As mentioned in the study--conducted by NASA and the Engineering and Science Contract Group (ESCG), a NASA consortium contractor: "Two bounding heat transfer modeling techniques were considered. The first envisioned a solid ball of frozen [hydrazine] in the center of the tank with a vacuum (following depressurization of the tank at the time of vehicle breakup) between it and the tank inner wall. In such a case, the principal subsequent heat transfer mode into the [hydrazine] would be via radiation. Alternatively, the [hydrazine] could be envisioned as a uniform shell, i.e., material layer, affixed to the complete inner wall of the tank with a dead space in the center. This internal layer of [hydrazine] would have a thickness of 0.197 meters. Both cases were analyzed."
While both cases may have been analyzed, the study reported only on the latter case--an idealization that cannot be physically supported. (The first case would have led to complete ablation of the tank and puzzlingly wasn't further mentioned in the study.) Here's why: The frozen hydrazine acts as a heat sink, and in reality, this "heat sink" is missing over a large fraction of the tank, making the tank susceptible to ablative burn-through in any region not touching the hydrazine. Although the hydrazine may melt and be repositioned during reentry, it cannot cover 100 percent of the tank given that the tank was only about three-quarters full.
Even with this oversimplification that biases toward tank survival, the study's authors still find that four out of the titanium tank's five concentric finite-element nodes would melt off since the reentry temperature exceeds titanium's melting point. Thus, even as modeled, out of an initial 3.56 millimeters tank thickness, only 0.7 millimeters remains, making it highly unlikely that the tank would survive the high g-forces and dynamic pressure of reentry. Had the study considered a higher-fidelity model with more than five nodes--say twenty, for example--it's almost certain that nineteen nodes would have ablated away even in the hydrazine contact region; the only node that will survive is the node in intimate contact with the hydrazine heat-sink. Of course, in reality, wherever hydrazine doesn't make intimate contact with the tank surface, the tank will burn-through completely.
Further, the released study treats the frozen hydrazine as a single finite-element layer--another oversimplification. This means that enough heat must first be absorbed to raise the temperature of the entire mass of hydrazine and then enough heat must be absorbed to melt the entire mass. In other words, until the entire hydrazine block melts, none whatsoever melts! Even the authors admit that's a far-fetched scenario: "It is extremely likely that the [hydrazine] in contact with the [titanium] wall will melt away in layers."
Next, the study assumes that the tank and its contents are spherically symmetric, moving in a tumbling and spinning motion, and undergoing a simple one-dimensional radial heat-transfer. The tumbling motion serves to more uniformly distribute the heat and would be perpetuated if the frozen hydrazine were symmetrically distributed within the approximately spherical tank. But since the tank is just 76 percent full, this isn't necessarily the case.
The center of mass of the tank-plus-hydrazine is offset from the geometrical center of the tank. Due to the asymmetric mass distribution in the tank's interior and the symmetric shape of the tank seen by the incoming airflow, a restoring torque is applied when the center of mass moves off of the line of sight through the geometric center of the sphere--similar to how an initially spinning shuttlecock will settle to point in the direction of the wind. This restoring torque will tend to stabilize any initial tumble that the tank may have had prior to reentry. Eventually, the tank will likely settle into an oscillating motion, with a frequency depending on the tank's moment of inertia and the restoring (aerodynamic) force, with the hydrazine-loaded end pointing in the direction of motion.
Since any initial tumble will tend to stabilize during reentry, the heat-load seen by this leading edge of the tank would significantly exceed that modeled in the study and result in ablative burn-through there. (If the tank isn't tumbling, then the heat isn't distributed over the whole sphere but rather, it's concentrated at the leading edge.) A proper--at least two-dimensional--heat-transfer code would be needed to model this--something the study doesn't do.
Although there were some indications that the ailing satellite may not have been stable before the interception, reports from experienced independent observers indicate that any initial tumble the satellite may have had wasn't excessive. For example, John Locker, a noted amateur satellite watcher, informed me, "Most, if not all, time lapse images showed no sign of flaring, flashing, or indeed a tumbling spacecraft."
Since the tank will separate from its fuel inlet plumbing at approximately 78 kilometers altitude, there will also be a hole where the tank's fuel intake line originally connected. Hydrazine melt in this region would likely evacuate due to high g-forces and result in a further reduction of the heat-sink material closest to the leading edge of the tank. Again, this wasn't modeled in the study.
So whether tumbling or not, there will be complete ablation of at least a portion of the tank: If the tank happens to continue its assumed initial tumbling motion, then the tank region not in contact with the frozen hydrazine will burn-through; if the tank becomes stabilized instead, then the leading edge will burn-through due to ablation.
The fidelity of the unrealistic NASA/ESCG model was further compromised by assuming an unreasonable value for the hydrazine's initial temperature--something the authors admit: "The initial temperature of the tank and the [hydrazine] was dictated [underlining mine] to be 214 [Kelvin] for this study, although the actual temperature would very likely be higher." Their choice of words is troubling. Who would "dictate" a temperature parameter in a presumably independent scientific study? And why?
A U.S. official familiar with the study has indicated that higher fidelity models than those in the released study were also used by government agencies to model the tank reentry. But since they haven't yet been made public, it's difficult to meaningfully comment on them. Therefore, although it should be presumed that the public-health reason for the interception was legitimate and made in good faith, the official study released so far certainly doesn't support the contention that the tank would have survived intact to the ground. In fact, despite its optimistic oversimplifications, the released study indicates that the tank would certainly have demised high up in the atmosphere.