25

What is the physical weight of one gigabyte of memory/storage? Lets say this is on a hard disk.

What is the weight associated with the atoms that are actually storing the data on the disk? How has this value changed as densities of disks have increased?

Breakthrough
  • 34,497
Callum
  • 1,606
  • 22
    The answer is 7x. I'll leave determining the value of x as an exercise for the reader. – Pesto Jul 23 '09 at 13:54
  • 5
    Simple: 1 GB weighs 1024 MB – splattne Jul 23 '09 at 14:07
  • This question is not meant to be serious, is it? – txwikinger Jul 23 '09 at 14:11
  • It's serious in the sense that zeros/ones on a disk are all data. It is interesting to consider the 'active' amount of substrate on the disk. – Callum Jul 23 '09 at 14:19
  • 6
    Callum: I can't tell yet whether you are just acting dumb or actually have no idea what you are talking about. – TheTXI Jul 23 '09 at 14:30
  • 3
    @TheTXI: I really don't think it is an act. – Geoffrey Chetwood Jul 23 '09 at 14:32
  • Obviously the disk won't weigh more with a gigabyte on it than without. Could the question could be rephrased: how much does the physical platter space required to store a gigabyte of data weigh? – jamuraa Jul 23 '09 at 14:49
  • 3
    jamuraa: That is even worse, seeing as how the space required to store a gigabyte is constantly shrinking...which is how we managed to go from storing 1GB to 1.5TB without having to constantly build bigger pc cases and drive enclosures. – TheTXI Jul 23 '09 at 14:54
  • TheTXI: which makes it a perfect SU question, because the answer can be updated to match the current technology. As gigabytes per square inch increases with new technology, a gigabyte on platter "weighs less". – jamuraa Jul 23 '09 at 15:00
  • @jamuraa: ...Wow. Just wow. – Geoffrey Chetwood Jul 23 '09 at 15:10
  • 1
    jamurraa: If you feel like posting an explanation on the size (area) required to take up one GB, feel free. And feel free to come back here twice a year and edit your post to make sure it is accurate. – TheTXI Jul 23 '09 at 15:13
  • 1
    @TheTXI - I'm actually researching now. It's difficult to determine though, because I can't find the thickness of an average hard drive platter (which also will tend to be smaller). Thinking of giving up as it may be trade secret or something and I'm not willing to tear apart a reasonable new hard drive to measure a platter with calipers.

    @Rich B - Hard drive companies are ALWAYS bragging about how they have increased their data density. Therefore gigabytes weigh less as time goes on.

     – jamuraa Jul 23 '09 at 15:19
  • Its a valid question, silly, but still valid. – Pyrolistical Jul 23 '09 at 16:25
  • Give the guy a break - he's only 10 years old. – pelms Jul 23 '09 at 23:42
  • See also http://stackoverflow.com/questions/758753/does-data-weigh-anything on SO. – mmyers Jul 29 '09 at 18:41
  • This has been done before...... http://social.answers.microsoft.com/Forums/en-US/vistahardware/thread/720108ee-0a9c-4090-b62d-bbd5cb1a7605 – zildjohn01 Aug 06 '09 at 22:31
  • 3
    It depend on the data. 0's look really fat and heavy compared to 1's. So if your data has a lot of 0's, it weighs more. :-) – kenj0418 Mar 01 '10 at 22:29
  • 1
    I would say 42. – Loïc Wolff Jul 23 '09 at 13:56

7 Answers7

77

Hard drive density is measured in bits per square inch, the highest of which are currently (5/2013) 750 gigabits per square inch. This means that a gigabyte of data will take up about 6.88 millimeter2. The weight of an area of a platter consists of the substrate (usually glass and ceramic) and the magnetic layer which actually holds the magnetic grains storing the data. The magnetic layer is usually made of a mostly cobalt alloy of 10-20 nm thickness. Assuming 10nm thickness to make the math easier, This gives us about 6.88 * 1013 nm3 of magnetic layer material for one gigabyte.

Given the density of cobalt, this means that we can approximate the weight at 0.612471 micrograms.

I'm not sure how much the substrate weighs, but it's almost certainly more than that.

2012 Update: This is all about drives that are shipping now - there is a lot of buzz about Seagate getting to 1 terabit per square inch recently, but that is a tech demo and not shipping quite yet.

2013 Update: It looks like the areal density of Hard Drive platters is stagnating, according to an interesting IBM report on the subject of Areal density. TDK says that they can approach the 1.5Tbits/inch2 mark, but they won't show up in the market until 2014. The Seagate tech touted in last year is supposed to show up in 2014 as well. Next year should be exciting for the weight of gigabytes.

2022 Update: After a jump high in 2016, it seems like the available HDD Areal density has fallen, and stalled again around 1.1TBit/in2. (HDD Areal Density Growth Slows as Capacity Increases). Things might heat up if HAMR, MAMR, or EAMR ever actually get something into the market (with promises of up to 2.6TBit/in2). We may be in for a wait, and they require new substrate, which may change how much a GB weighs too. Stay tuned.

Previously on "How much does a gigabyte weigh on a hard disk?"

  • 2009: Areal Density 400 Gbit/in2 = 1.1518 micrograms (ref)
  • 2010: Areal Density 541.4 GBit/in2 = 0.84817 micrograms (ref)
  • 2011: Areal Density 625 GBit/in2 = 0.734966 micrograms (ref)
  • 2012: Areal Density 744 GBit/in2 = 0.617411 micrograms (ref)
  • 2016: Areal Density 1.3 TBit/in2 = 0.35 micrograms (ref)
  • 2022: Areal Density 1.1 TBit/in2 = 0.45 migrograms
jamuraa
  • 3,776
31

The data kept on a disk does not increase the weight of the disk. The only weight differences in disks would be in the overall size of the disk (example: regular HDDs are larger than laptop HDDs in terms of size and typically mass, and larger sized disks can have more platters to hold data than older ones) itself and in the materials used to make the disk.

Data is stored by switching the magnetic polarity on the disk, not by adding or subtracting something from the actual substance. A full disk will have the same mass and will therefore weigh the same (assuming you don't move the disk to a location where gravity is stronger or weaker, such as the moon).

Switching the polarity of a hard disk is like turning a magnet around so that the north and south poles are switched. It is not analagous to creating an ion (removing or adding electrons of an atom to give it a positive or negative charage). That could theoretically adjust the mass of the disk, but for all intents and purposes electrons do -not- have mass (so infinitesimally small that it almost appears so at least), so you are back to square one again if the disk did somehow operate in this manner, which it does not.

TheTXI
  • 3,599
  • 8
    Good answer to a rubbish question. Reminds me of a DBA I once worked with who suggested that we install the software on the servers after we moved them so they wouldn't be too heavy. He was being serious. Put me off DBAs for about 6 years. – Jon Hopkins Jul 23 '09 at 14:26
  • I have troubles believing someone actually upvoted the chosen answer, despite the presence of yours. Makes me rethink my desire to help people ho post questions. If you cannot tell the right from the wrong when both are written in plain, clear English, what is the point? I feel like howling at the moon. – MariusMatutiae Jun 24 '15 at 17:55

  • for all intents and purposes electrons do -not- have mass

    They have very small mass but you are multiplying it by a very large number.

     – StenSoft Jul 17 '15 at 14:15
  • @MariusMatutiae: Because the accepted answer is correct. The question asks "What is the weight associated with the atoms that are actually storing the data on the disk?" There are atoms that are actually storing the data on disk, and they have non-zero mass. That fact is not diminished by the fact that they have the same mass (within experimental accuracy) before data is written. You could not have written the data, were these atoms not present. The added weight is zero, the weight of the storage medium is not (in the presence of a gravitational field). – Ben Voigt Jul 08 '19 at 03:14
15

On the disk an individual bit weighs nothing, it's just a change in magnetic polarity; see TheTXI's answer for a more elaborate explanation of this.

In RAM, however, bits are comprised of electrons (or lack thereof) and they do have a mass which is about 9.10938215 × 10−31 kg. So for a GiB of memory, assuming equal distribution for zero and one bits, we get around

4294967296 n × 9.10938215 × 10−31 kg

4294967296 would be the number of one bits in memory (assumed to be 50 %) and n would be the number of electrons that are on average in one bit. I have found one source1 that specified this number at around 105.

So we can give an estimate of how much mass 1 GiB (or 1 GB) of memory would have:

1 GiB, half filled with ones ≈ 3.91 × 10−16 kg = 391 femtograms

1 GiB, completely filled with ones ≈ 7.82 × 10-16 kg = 782 femtograms

1 GB, half filled with ones ≈ 3.64 × 10−16 kg = 364 femtograms

1 GB, completely filled with ones ≈ 7.29 × 10−16 kg = 729 femtograms

So in general you can assume that weight to be pretty unnoticeable (or, with hard disks to be downright nonexistant).

1 These lecture slides, but they are in German.

Community
  • 1
Joey
  • 40,452
  • 3
    Trivia to add: On researching the numbers for this I stumbled over a fun calculation which estimated the mass of traffic on the internet at a single point in time to be roughly equivalent to the mass of a grain of sand. – Joey Jul 23 '09 at 14:52
8

Not-serious answer

It depends on what font size your text is saved in. 24-point font is very heavy, whereas 8-point is quite light. Bold text also increases the weight, and you should avoid saving lots of text in italics, because all the characters lean to the right, which changes the way the disk spins.

Serious answer

Data has no mass.

Graeme Perrow
  • 389
  • That sounds like something out of Dilbert. Actually, now that I think about it, I'm almost sure I've seen something about bold text weighing more. – mmyers Jul 29 '09 at 18:40
6

The correct answer is 0. He didn't ask how much hard drive does it take to store 1 GB, he asked how much 1 GB weights on a hard disk. As we know it uses magnetic storage and not an electrical charge (which would weigh something), then correct answer is 0.

Pyrolistical
  • 1,170
3

It depends on the data.

Yes, hard drives store data by flipping poles in magnetic domains on the disk--at first glance this means nothing is added or subtracted and thus no weight.

However, that's not the whole picture. The orientation of those domains matter. There is less total field energy when the domains are 1010101010 than when they are 11111111 or 00000000. I'm sure everyone is familiar with e=mc^2. Putting energy into the domains DOES mean mass, albeit an incredibly small amount of it.

My physics isn't up to even trying to estimate the mass but I'm sure it's beyond anything the most sensitive scale could possibly measure.

Loren Pechtel
  • 2,624
1

Depends on where you're doing the weighing. One of the answers immediately jumps into discussing femtograms, which are not a measure of weight, but instead measure mass.

On the moon things weigh less, on Jupiter they weigh more. In space they weigh nothing.

So, the answer is ... depends.

Trey Jackson
  • 4,021
  • 1
    Yes, because "I weigh 750 newtons". Technically, you are right, but in popular usage "weight" is almost universally meant to refer to mass. – Thomas Jan 02 '13 at 07:24