Why don't magnets stick to hard drive platters?
May 12, 2016 4:02 PM   Subscribe

Conventional hard drives are magnetic media, so why is it if you disassemble one and stick a magnet to a platter, you don't feel any magnetic attraction at all? I can understand feeling a very weak level of attraction, but it feels like sticking a magnet to a piece of plastic or wood.
posted by Bugbread to Technology (7 answers total) 1 user marked this as a favorite
The layer on the drive that holds the bits is necessarily very, very thin. That's because the "head" that reads the on/off has to get close enough to feel the magnetic field. The rest of the platter is deliberately non-magnetic.
posted by wnissen at 4:09 PM on May 12, 2016 [3 favorites]

Ah! Since the whole platter looks like an undifferentiated piece of metal, it hadn't occurred to me that there would be different layers and that the majority of the body would be non-reactive, but that totally makes sense!

Okay, quick followup question, since the main question was answered so swiftly: The whole surface appears to be made of the same metal. Is the whole non-magnetic body covered with a thin film of reactive material (including the outer and inner vertical edges, like the red part in this photo), or is it just the upper and lower faces?
posted by Bugbread at 4:28 PM on May 12, 2016

There is no use for the magnetic material on the edges, but it probably makes the manufacturing process easier to coat that too.
posted by Chocolate Pickle at 4:39 PM on May 12, 2016 [1 favorite]

To elaborate, a ferromagnetic platter would make it hard to read the magnetic field of the tiny "bit" of surface. With disks having on the order of 100 gigabits (10^9 bits) per square cm, each bit is on the order of nanometers in size. If you have an old floppy disk or VHS tape or similar magnetic medium, it won't stick to a magnet either, for the same reason. And they have bit densities that are way, way lower than a hard drive. The thinness of the coatings and the precision of the head and the motors that position it while the medium screams around thousands of times a second are out of the range of human comprehension, really. The miniaturization required to create a $100 hard drive that will last for five years is one of the great technical accomplishments of our civilization.

Your question is out of my realm of expertise, my guess is they probably coat the whole platter because it would be more difficult to mask off the parts that aren't supposed to be coated. They do these coatings with a process called "sputtering", which despite the name creates an unbelievably fine "mist" of metal particles that then deposit on the surface. It's not like painting a coating on where you control a spray head or something like that, the "mist" just settles into position very evenly over everything. On preview I see Chocolate Pickle is making the same guess I am.
posted by wnissen at 4:40 PM on May 12, 2016 [3 favorites]

Since you seem interested, here's some detail about the manufacturing process in general.

There are 3 main options for the basic platter substrate, i.e., the underlying rigid disk object that is ~1/2-1 mm thick: glass, ceramic, and aluminum. Last time I had any reason to pay attention (5 years ago), it was kind of arbitrary between ceramic and Al; the industry may have consolidated on one option since then.

Starting with a high-precision manufactured aluminum (for example) disk, the first step is to lay down an "underlayer" that acts as glue between the substrate and the magnetic layer, since the optimum composition for the magnetic layer may not be the optimum composition for sticking to the substrate. After that the magnetic layer itself is deposited, typically a cobalt layer because cobalt is weakly magnetic. And yes, given the high-volume manufacturing processes for these disks, they will definitely coat the entire platter including the edges rather than trying to mask anything.

Finally it is sealed by applying a thin "interface layer" or "overcoat" of amorphous hydrocarbon. I can't speak for the underlayer and the magnetic layer, but sputtering hasn't been state of the art for the interface layer for about a decade, maybe longer. At the scale we're talking about, sputtering deposits boulders when we're aiming for pebbles. Rather, the overcoat process uses plasma-assisted chemical vapor deposition which starts with a low-pressure hydrocarbon gas (like acetylene) which is raised to a high temperature and magnetized to "aim" it in a stream at the HD platter. This helps prevent corrosion of the magnetic layer because, while it's far too thin to prevent mechanical damage (e.g. from the read head itself "crashing"), it does keep out the oxygen.

Incidentally this final layer is the reason HD platters look "oily." The composition is somewhere between diamond and amorphous "carbon glass," so it's relatively clear. And it's so thin that different wavelengths of light are preferentially reflected through and refracted by the layer depending on the angle from which you're looking.
posted by Joey Buttafoucault at 6:17 PM on May 12, 2016 [18 favorites]

Thanks, y'all! Interesting and awesome!
posted by Bugbread at 3:02 AM on May 14, 2016

There would be other problems too if the disk itself was magnetic as in it would induce current in all the surrounding metal causing electronic noise and it would generate heat too.
It is a bit like audio tape where the tape itself is plastic and only a thin film on it is magnetic.
posted by boilermonster at 10:26 PM on May 14, 2016

« Older Cat had a muscle contraction attack   |   How to tell my supervisor that we need more... Newer »
This thread is closed to new comments.