That Changed Everything
When Artemis I splashed down in 2022, the inspection of its heat shield revealed something no one had predicted — and triggered a two-year investigation that fundamentally changed how Orion returns to Earth.
On December 11, 2022, NASA’s Orion spacecraft splashed down in the Pacific Ocean after a flawless 25.5-day journey around the Moon on the uncrewed Artemis I mission. The capsule was intact. The mission was a success. Then engineers removed the heat shield for inspection — and found something they hadn’t seen before, hadn’t predicted in their models, and didn’t fully understand.
What they found would take two years and 121 individual thermal tests to explain. It would delay the first crewed Artemis mission by over a year. It would trigger public controversy, calls from former astronauts to ground the crew, and a sweeping redesign of how Orion reenters Earth’s atmosphere. And it would change the fundamental reentry trajectory for every crewed Artemis mission that follows.
This is the full story: what happened to the heat shield, why it happened, how NASA investigated it, and what they changed to protect the four astronauts flying on Artemis II.
Meet Avcoat: The Material Between Astronauts and 5,000-Degree Plasma
When Orion reenters Earth’s atmosphere after a trip to the Moon, it hits the top of the atmosphere at roughly 25,000 miles per hour — about 7 miles every second. At that speed, the friction between the spacecraft and the air generates a plasma sheath around the capsule with temperatures that can exceed 5,000°F (2,760°C). For reference, that’s hotter than the surface of the Sun at its photosphere. No metal known to engineering can survive that temperature.
The solution, developed for Apollo and revived for Orion, is a material called Avcoat — an ablative heat shield. “Ablative” means it’s designed to burn away, and that’s precisely the point. As Avcoat chars and erodes, it carries heat away from the spacecraft, leaving the crew capsule beneath it cool. Think of it as a controlled, deliberate sacrifice: the shield destroys itself so the crew doesn’t.
Imagine blowing on a hot coal. The outer surface glows and crumbles — but the coal’s core stays intact because the heat is being carried away with the material that burns off. Avcoat works the same way. As it chars, it produces gases that carry thermal energy away from the spacecraft. The char layer that forms is itself a reasonable insulator. For this to work properly, the gases produced by that charring process need to be able to escape outward through the material. That outward venting is what the investigation eventually found was missing.
For the Orion spacecraft, Lockheed Martin manufactured Avcoat as roughly 186 pre-machined blocks (about the size of thick tiles), bonded onto a carbon fiber carrier skin. This was a change from the Apollo era, when workers filled over 300,000 individual honeycomb cells by hand — a process that took months. The block approach was faster and more modern. But it also subtly changed how the material behaved — and those changes turned out to matter enormously.
The Discovery: Chunks Where There Should Have Been Smooth Erosion
After Artemis I splashed down, the heat shield was removed and transported to the Operations and Checkout Building at Kennedy Space Center for inspection. What engineers found was not smooth, gradual ablation as designed. Instead, they found more than 100 locations across the heat shield where the charred outer layer had cracked, broken apart, and ejected in chunks. Deep gouges and holes punctured blocks that should have eroded evenly. The material had not performed as the models predicted.
1. Heat shield spalling: Chunks and divots in the Avcoat leave voids in the shield. If a void is large enough, unprotected capsule structure underneath is exposed to plasma — potentially leading to burnthrough.
2. Bolt erosion: Four large separation bolts embedded in the heat shield are supposed to withstand reentry. Three of the four had their thermal barriers melted through on Artemis I, due to a flaw in NASA’s heating model. Melted bolts can expose the capsule to hot gas ingestion behind the shield.
3. Debris impact risk: When spalling ejects pieces of heat shield into a 25,000 mph hypersonic airstream, those fragments can strike the top of the capsule, potentially damaging the parachute compartment.
NASA’s initial public communications were carefully measured — a Lockheed representative noted there was “a healthy margin remaining of that virgin Avcoat.” But in May 2024, the NASA Office of Inspector General released photographs of the heat shield, and the extent of the damage became unmistakably clear. The problem wasn’t gradual erosion. It was structural fracture in dozens of places.
“The Avcoat material is not designed to come out in chunks. It is supposed to char and flake off smoothly, maintaining the overall contours of the heat shield.”
— NASA OIG analysis, 2024The Root Cause: Trapped Gas, Internal Pressure, and an Unpermeable Shield
After two years, 121 individual thermal tests at facilities across the country — NASA Ames, NASA Langley, CUBRC in New York, the University of Kentucky, Lawrence Berkeley National Lab — engineers had a complete answer. The failure had two interlocking causes, and understanding them requires understanding the skip reentry maneuver that Artemis I used.
The skip reentry (also called skip guidance entry) is a technique in which the spacecraft dips into the upper atmosphere to slow down, uses its aerodynamic lift to briefly bounce back out into near-vacuum, and then reenters for a final descent to the ocean. The maneuver was designed to give Orion more flexibility in choosing a landing site — by extending the range it could fly after the reentry point. Think of a stone skipping across a pond: controlled bounces, controlled energy dissipation.
Here is the physics of what went wrong. When Avcoat chars during reentry, a chemical process called pyrolysis occurs — the resin in the material decomposes under heat and produces gases. In normal operation, those gases escape outward through the porous char layer and help cool the surface. This only works if the char layer is permeable — porous enough for the gas to pass through.
The skip reentry created a critical problem: during the “skip out” phase, when Orion was briefly back in near-vacuum between atmospheric dips, heating rates dropped sharply. This slowed char formation — but it didn’t stop gas production. The result was a gap: gases were forming inside the shield faster than the char layer was becoming porous enough to release them. Pressure built up inside the Avcoat. When the spacecraft re-entered the atmosphere for the second dip, that trapped, pressurized gas had nowhere to go — and it cracked and blew chunks out of the heat shield.
Think of water freezing inside a rock crack. The water (like the gas) can’t escape outward, so when it expands, it fractures the rock from the inside. The Avcoat didn’t fail because it got too hot. It failed because it didn’t get hot enough during the skip-out phase — not enough to form a properly porous char layer — but still produced gases that had nowhere to go.
There was an additional complicating factor: the Artemis I heat shield was built using a slightly different Avcoat formulation than Apollo’s, due to modern environmental regulations that banned certain chemicals used in the original material. The block construction method — replacing 300,000 hand-filled cells with 186 pre-machined tiles — also changed the material’s permeability characteristics in ways that weren’t fully captured by pre-flight testing. And critically, the pre-flight tests had been run at higher heating rates than Orion actually experienced in flight — conditions under which the char formed properly and gas vented normally. The real flight’s milder skip-out heating was the scenario NASA had never tested for.
Two Years, 121 Tests, and a Decision No One Fully Agreed With
The investigation was one of the most extensive in NASA’s post-Columbia era. Engineers took approximately 200 Avcoat samples from the Artemis I shield at Marshall Space Flight Center, performed non-destructive evaluation to look inside the material, and ran a battery of tests at facilities across the country.
“What they’ve done with all this learning is they put it in this category of us being able to understand the risk and make a risk trade.”
— Jeremy Hansen, CSA Astronaut, Artemis II Mission SpecialistThe Fix: Change the Flight, Not the Shield — and the Debate It Started
NASA’s decision was blunt: the Orion capsule for Artemis II was already fully assembled and mated to its service module. Rebuilding a new heat shield would take years and hundreds of millions of dollars. Instead, the agency concluded that by eliminating the skip reentry in favor of a steeper, single-pass direct reentry, the heating would be continuous rather than cyclical — and the Avcoat would always form a properly porous char layer without the dangerous cooling gap that caused the fractures.
The trade-off was real: a direct, steeper entry reduces the range Orion can fly to a splashdown point, limiting the choice of landing sites. For Artemis II, this meant targeting a specific zone in the Pacific Ocean off San Diego, with less flexibility to adjust based on weather conditions. The spacecraft would also experience higher peak heating — briefly hotter than Artemis I’s skip entry — but for a shorter, uninterrupted duration.
1. Reentry trajectory: Replaced the skip guidance entry with a steeper, direct single-pass reentry — no bounce back out of the atmosphere.
2. Trajectory correction burns: Three small return burns during the homeward journey ensured precise targeting of the Pacific reentry corridor with the new steeper profile.
3. Heat shield block design (for Artemis III onward): NASA identified that adjusting the density and porosity of the Avcoat blocks improves permeability. The Artemis III heat shield is being manufactured to a new, more permeable specification at Michoud Assembly Facility in New Orleans.
4. Ground testing infrastructure: New arc jet capabilities at NASA Ames can now accurately reproduce the specific moderate-heating conditions that caused the problem — conditions that previous testing could not replicate.
Not everyone was satisfied. Retired shuttle astronaut and heat shield expert Charles Camarda — former Director of Engineering at Johnson Space Center — became the most prominent public critic, arguing publicly that NASA was exhibiting the same motivated reasoning that led to the Challenger and Columbia disasters: building models to reach a predetermined conclusion rather than grounding decisions in physical evidence. He called it “flying based on vibes.”
The Artemis II crew itself, told everything about the investigation, publicly expressed confidence in the engineering team’s work. Victor Glover: “I think the crew is comfortable because of that team, and we know how many people are looking at this.” Commander Wiseman praised the independent review board, which included former astronauts and was led by the same engineer who guided the shuttle program back to flight after Columbia. But in a striking moment of honesty, Hansen put it most plainly: “There’s always going to be a risk there, but it is in family with all the other risks we’re going to take.”
“The process was extensive. We gave the team the time needed to investigate every possible cause, and they worked tirelessly to ensure we understood the phenomenon and the necessary steps to mitigate this issue for future missions.”
— Howard Hu, Orion Program Manager, NASA Johnson Space CenterWhat the Heat Shield Story Actually Tells Us
The Artemis I heat shield incident is easy to read as a NASA failure. In a narrow sense, it was: a material behaved unexpectedly in a way pre-flight testing failed to predict, and it took two years to understand why. But in a broader sense, it is precisely what a test flight is for. Artemis I was uncrewed because NASA knows that novel systems in novel environments will produce surprises, and surprises should be encountered without crew on board when possible.
The failure to predict the skip-out heating behavior came from a genuinely subtle interaction between material permeability, a specific entry trajectory, and heating rates that fell outside the tested range. The block-construction Avcoat behaved differently from the Apollo honeycomb in ways that even experienced engineers didn’t fully appreciate. The pre-flight test infrastructure couldn’t reproduce the exact conditions of lunar-return skip entry until new arc jet capabilities were installed after the fact.
NASA’s response — change the trajectory, redesign the shield for Artemis III, improve the test infrastructure — is methodical and defensible engineering. Whether the margin of safety it provides for Artemis II was sufficient is a judgment call that reasonable engineers disagreed about. The crew flew anyway, with clear eyes about the risk. Orion Integrity came home safe. The heat shield survived. And the data from Artemis II’s reentry will inform every Orion heat shield decision for the next decade.
The full lesson of Artemis I is not that the heat shield failed — it’s that uncrewed test flights exist to produce exactly this kind of knowledge before the stakes are human lives. In that sense, the mission did its job perfectly.
Comments are closed.