A 0.8 nm amorphous carbon coating grown on hard disk media matches the protective performance of films three times thicker, opening a path to far higher storage density.
(Nanowerk Spotlight) The explosive growth of artificial intelligence is creating a storage crisis. Training a single large language model can require petabytes of data, and every query to a chatbot, every generated image, every automated analysis adds to an accelerating flood that must be stored somewhere. Solid-state drives handle the fast-access needs of consumer devices, but data centers still rely overwhelmingly on magnetic hard disk drives for bulk storage because no other technology matches their cost per terabyte at scale. To keep pace without endlessly expanding the physical footprint and energy consumption of server farms, the industry needs to pack more data into every square inch of magnetic disk.
The obstacle is not the magnet. It is a carbon film just 2.5 nm thick that coats the surface of every hard disk, protecting the magnetic recording layer from corrosion and mechanical damage. The thinner this overcoat gets, the closer the read/write head can fly to the magnetic surface, and the more tightly data bits can be packed.
Theoretical calculations suggest that if the overcoat could be thinned below 1 nm on both the disk and the read head, areal densities approaching 10 Tb/in² would become feasible, roughly seven times today’s commercial standard. But the material the industry relies on, diamond-like carbon, cannot get there. Its sp³ bonding structure, the same atomic arrangement found in diamond, projects out of the film plane and prevents the formation of a continuous, defect-free coating at sub-nanometer thickness.
And the demands on this film are growing. Next-generation recording technologies such as heat-assisted magnetic recording (HAMR) use a focused laser to heat individual magnetic bits to around 750 K in less than 1 nanosecond, allowing data to be stored in high-coercivity iron-platinum alloys that resist accidental erasure. Diamond-like carbon graphitizes and cracks under these rapid thermal cycles.
Graphene has shown promise as an alternative in laboratory transfer experiments. But growing it directly on a hard disk’s complex, multi-material surface at a temperature low enough to preserve the magnetic layer has so far proved unsuccessful.
A study published in Advanced Materials (“Breaking the 2‐nm Barrier in Hard Disk Drives Using Monolayer Amorphous Carbon Overcoats”) offers a different solution. A team based primarily at the National University of Singapore grew monolayer amorphous carbon (MAC) directly onto commercial hard disk media, producing a uniform protective film just 0.8 nm thick across an entire 2.5-inch disk. Unlike graphene, MAC has no crystalline lattice. It is fully disordered, with sp²-type bonding, meaning its carbon atoms form a flat, planar network rather than the out-of-plane diamond-like configuration.
This structural difference allows the film to remain continuous at sub-nanometer thickness and gives it the thermal resilience that conventional overcoats lack.
Characteristics of HDDs with directly grown monolayer amorphous carbon overcoating. (a) Schematic illustration of the HAMR device zoomed at the head (including laser)/media interface. Magnetic bits are indicated by blue and red stripes with a protective carbon overcoat (COC) layer (black) and lubricant (gray) on top. (b) AFM surface topography image used for roughness analysis of as-grown MAC on hard disk (RMS ∼ 0.18 nm). (c) Phase contrast MFM image of a hard disk after MAC growth, clearly showing ordered magnetic bit lines. (d) Large view cross-sectional TEM image of as-grown MAC on HDD capped with an Au layer. (e) Histogram of the thickness distribution obtained from multiple TEM images. (f) Zoom-in cross-sectional TEM image of as-grown MAC showing a thickness of 0.8 nm. (g) Zoom-in cross-sectional TEM image of a commercially purchased hard disk with COC showing a thickness of 2.5 nm. (h) Ellipsometry mapping of MAC thickness across the entire 2.5-inch disk surface showing a thickness variation smaller than 0.2 nm. (Image: Reproduced from DOI:10.1002/adma.202519149, CC BY) (click on image to enlarge)
The researchers started with commercially available 5 TB Seagate hard disk platters. They removed the existing lubricant and carbon overcoat using argon-ion plasma cleaning inside a chemical vapor deposition (CVD) chamber, then grew MAC in the same chamber using methane as a carbon source. A copper foil placed beneath the platter served as a catalyst; control experiments without it produced incomplete films with poor performance. A 248 nm ultraviolet laser and a low-power plasma supplied the energy to decompose the methane and deposit carbon on the disk surface at a final substrate temperature of approximately 300 °C. The entire growth step took 10 minutes.
Cross-sectional transmission electron microscopy confirmed the MAC layer was continuous and uniform at 0.8 nm, compared to 2.5 nm for the factory-applied coating. Ellipsometry mapping across the full disk showed thickness variation of less than 0.2 nm. Atomic force microscopy measured a surface roughness of 0.18 nm after growth, virtually identical to the 0.17 nm roughness of the original commercial disk. Magnetic force microscopy confirmed that the bit patterns on the disk survived the entire process intact.
The overcoat’s primary job is to block corrosion of the chemically reactive transition-metal alloys, including cobalt, iron, and chromium, in the data storage layer. Using standard three-electrode electrochemical measurements in a 0.1 M sodium chloride solution, the team found that MAC reduced the corrosion current density to approximately 3.8 nA/cm², yielding a corrosion protection efficiency of 82.7%.
That figure matches the protection offered by the much thicker 2.5 nm commercial overcoat and surpasses the performance of four-layer transferred graphene films tested in earlier work.
Thermal stability posed the more demanding test. The researchers subjected both MAC and conventional overcoats to an 800 nm femtosecond pulsed laser that mimicked HAMR’s rapid heating cycles. A 15-second exposure at an 80 MHz repetition rate simulated the cumulative laser dose expected over a five-year hard disk lifetime.
Raman spectroscopy, a technique that detects changes in carbon bonding structure, revealed clear graphitization in the conventional diamond-like carbon: the D-to-G peak ratio increased and the G band shifted to higher wavenumbers, both markers of structural breakdown. MAC showed no detectable spectral change across the full range of laser powers tested, up to 12 mW.
The friction results raise a further possibility. Hard disks currently require a lubricant layer between head and media to cushion unavoidable contact events, but this lubricant takes up space in the head-media gap and degrades under HAMR’s thermal stress. Using lateral force microscopy with a silicon dioxide ball tip to simulate head-disk contact, the team measured a friction coefficient of about 0.41 between MAC and bare silicon dioxide, too high for practical use. But when both surfaces carried a MAC coating, the coefficient dropped to approximately 0.27, comparable to a conventional overcoat paired with lubricant, and it held stable over 15,000 repeated cycles.
This suggests that MAC-coated heads and disks could potentially operate without any lubricant, further shrinking the head-media spacing.
From a manufacturing standpoint, the 300 °C growth temperature falls within the thermal budget already used in hard disk production. The process is compatible with double-sided disk coating, and the team demonstrated it on 2.5-inch substrates with potential scalability to at least 4-inch disks. Growth times, currently on the order of minutes, sit within one order of magnitude of industrial targets.
Limitations remain. The friction tests operated at the nanoscale; macroscopic wear testing under realistic drive conditions has not yet taken place. Long-term durability over billions of read/write cycles in a commercial product still requires demonstration. The authors also note that while MAC approaches the thermal robustness of graphene, full equivalence has not been established.
The core achievement is the simultaneous delivery of sub-nanometer thickness, corrosion protection matching films three times thicker, structural stability under HAMR-like thermal stress, and low friction without lubricant. No previous carbon overcoat technology has combined all four. If these results hold at manufacturing scale, MAC could remove the primary physical constraint limiting hard disk storage density, with potential applications extending to other technologies that require ultra-thin protective coatings, from nanoscale sensors to photonic devices.
For authors and communications departmentsclick to open
Lay summary
Prefilled posts
ORCID information
Barbaros Ӧzyilmaz (National University of Singapore)
, 0000-0001-7665-4578 corresponding author
Nanowerk Newsletter
Get our Nanotechnology Spotlight updates to your inbox!
Thank you!
You have successfully joined our subscriber list.
Become a Spotlight guest author! Join our large and growing group of guest contributors. Have you just published a scientific paper or have other exciting developments to share with the nanotechnology community? Here is how to publish on nanowerk.com.
