Really, really hi-res printing

August 13, 2012 10:09 AM

The highest possible resolution images — about 100,000 dots per inch — have been achieved, and in full-colour, with a printing method that uses tiny pillars a few tens of nanometres tall. The method, described today in Nature Nanotechnology, could be used to print tiny watermarks or secret messages for security purposes, and to make high-density data-storage discs.

The pretty lady in the picture may look familiar.

posted by Egg Shen (22 comments total)

So, how many angels can be printed on the head of a pin?
posted by jfuller at 10:18 AM on August 13, 2012 [2 favorites]

A head of a pin has a diameter of roughly 2mm, or an area of pi*(d/2)^2 = 3.14E-6 m^2. Assuming we use this picture as a reference point for 'angel', it has an area of 50um x 50 um = 2.5E-9 m^2. Thus, we have 3.14E-6 / 2.5E-9 = ~1,250 'angels' printed on the head of a pin.

Of course, YMMV given different values for 'pin' and 'angel'.
posted by Arandia at 10:28 AM on August 13, 2012 [4 favorites]

1,250 'angels' printed on the head of a pin

Enough to store every Playboy centerfold ever.

Though going blind would become a genuine risk.
posted by Egg Shen at 10:35 AM on August 13, 2012 [2 favorites]

Isn't the highest possible resolution going to depend on the wavelength of the colors? So an image with a cool palette (shorter wavelengths) could be printed smaller than an image with hot colors?
posted by qxntpqbbbqxl at 10:44 AM on August 13, 2012 [1 favorite]

The appropriate question is how many Zippies can be printed on the head of a pin.
posted by pracowity at 10:45 AM on August 13, 2012 [1 favorite]

They should try to break the record for darkest black. I want somethinjg so dark the light just falls into it.
posted by humanfont at 10:49 AM on August 13, 2012 [2 favorites]

humanfont, that record has already been achieved.
posted by dances_with_sneetches at 10:56 AM on August 13, 2012 [1 favorite]

humanfont, that record has already been achieved.

The odd thing about this comment is that, although the link needs to be there, there is no need for anyone to click it to find out where it goes.
posted by howfar at 11:03 AM on August 13, 2012

Good news: The printer is only $79.99 at Staples.
Bad news: The cartridges only last 4 or 5 prints, and the replacements are $1,495.99. At Staples.
posted by Thorzdad at 11:17 AM on August 13, 2012 [5 favorites]

For more science less Metal, the method for darkest black also involves nanotubes.

I suppose if you wanted full-colour pictures, you would be limited by the longest wavelength in the visible spectrum.
posted by RobotHero at 11:34 AM on August 13, 2012

humanfont: "They should try to break the record for darkest black. I want somethinjg so dark the light just falls into it."

None. None more black.
posted by symbioid at 11:38 AM on August 13, 2012 [2 favorites]

Call me pedantic.

1250 is an overestimate by ~50 'angels', because I neglected to include the images that get cut through by the edge of the pin.

The circumference of the pin = 2*pi*r = 6.28E-3 m. When bisected by this edge, an 'angel' will take between 50um and 70um of this circumference, depending on the angle at which the angel is cut (70um = 50um * root(2), when cut at 45 degrees). Call this the 'cut length'. Assume that at these scales, the curvature of the cut length is negligible and can be modeled as a straight line. The cut length per angel will therefore be a sinusoid between 50um and 70.8um, for four periods (50 at 0 degrees, 70 at 45, then back to 50 at 90 degrees, etc). We could integrate, but with these values it's the equivalent as taking the average -- each 'cut angel' will take 60.4 um of the circumference, and we therefore will have 6.28E-3 m / 60.4E-6 m = 104 'cut angels'.

However, the area calculation includes these. Assuming that the cut area for each is ~1/2 of an 'angel', that means that the previous calculation is only off by 52.

Therefore, we will have ~1200 'full' 'angels' and ~100 'partial' 'angels' on the head of a pin. Of course, if we were to adjust where we started the tiling, we could probably squeeze out one or two more 'full' 'angels'...

...ok, you can call me really pedantic.
posted by Arandia at 11:53 AM on August 13, 2012 [3 favorites]

Of course the test image is Lena Söderberg.
posted by migurski at 11:54 AM on August 13, 2012

Not that anyone would want to be pedantic about angels on the head of a pin.
posted by dances_with_sneetches at 12:06 PM on August 13, 2012

Too bad they're still using serial write EBeam Lithography. Doesn't scale, not commercializable.

There's some developing technology for parallel write EBeam Lithography, which uses a technological cousin to the micromirror array in a DLP projector for electrons. With that sort of equipment, then this might be more workable.
posted by Chekhovian at 12:08 PM on August 13, 2012

Isn't the highest possible resolution going to depend on the wavelength of the colors? So an image with a cool palette (shorter wavelengths) could be printed smaller than an image with hot colors?

Yeah, so if I'm getting this right, this is the maximum possible resolution capable of reflecting light with a wavelength up to 500 nm, which means it can reflect BIV but gets blurry for ROYG. If you had a very low contrast image using just colors at the violet end of the spectrum, you could have a sharp image down to 200 nm pixels, which would get you up to 125,000 DPI. But it would be purple text on a purple background, and that might not be a winner design-wise.
posted by jhc at 1:08 PM on August 13, 2012

What is this, a printer for ants?!?

Sorry, I've been spending a lot of time on imgur lately
posted by hincandenza at 1:41 PM on August 13, 2012

1250 is an overestimate by ~50 'angels', because I neglected to include the images that get cut through by the edge of the pin.

But angels have holographic properties in that each fractional angel contains all the information of a full angel. Angels are funny that way. Anyway, your first estimate is theologically acceptable.

To someone.
posted by GenjiandProust at 1:58 PM on August 13, 2012

Chekhovian: Experts have told me parallel e-beaming is almost impossible, but that was some time ago. Do you have a reference or a good overview article about it?
posted by springload at 4:20 PM on August 13, 2012

Demonstration of lithography patterns using reflective e-beam direct write.
posted by Chekhovian at 5:01 PM on August 13, 2012

Thanks!
posted by springload at 4:51 AM on August 15, 2012

And the data storage discs (at least the consumer versions) will still only last 4-5 years max. Long term storage of digital data is doomed unless you have google-like resources.
posted by InsertNiftyNameHere at 1:07 AM on August 17, 2012