Microscopic microscopes - not yet for the masses

  • At the 2014 forsight institute conference Neil Sarkar held a talk called "microscopic microscopes for the masses"
    (There was a video online but I can't find it anymore. Was it removed?)
    Today I got an e-mail notifying me that they finally sell their microscopes.

    With a price point near $10k I doubt this will make it to the masses anytime soon though.
    I think that's about one order of magnitude to high.
    As I understand the point of the project was/is to pack/cram most of the functionality into/onto chip.
    With $250 I think the tips are bearable priced.
    I'm wondering though why the frame and rough positioning turned out to be so darn expensive.
    Maybe this cries for a DIY 3D-printable replication.
    AFAIK there are some people (on thngiverse) successfully doing high accuracy optical system mirror adjustments with just crude 3D printed parts.
    For drift reduction maybe adding temperature control (via a Peltier element) is needed but I don't see that pushing the price to near $10k
    So what makes the price?
    Software ?! Machining ?! The linear bearings ??
    Long range rough positioning vibration Piezo drives ?
    Maybe secondary amplifier electronics ?

  • I presume this is the firm's site (let me know if it isn't):

    Obviously the price is based on their expectations of number of units they hope to sell versus the cost they've so far spent in developing the microscope plus per-unit costs. And also what competitors are charging and getting away with - even though they claimed it was for the masses. I had not previously looked at the current market for AFMs but your post made me curious and I came across the following document which gives some estimates of its size and growth:


    According to this link that market analysis probably wasn't cheap to purchase either! I presume that issue is either out of date or was posted in violation of copyright (if the latter I would expect it to vanish eventually.) In any case, I believe the US$7900 price of the AFM is relatively inexpensive - compared to competitors.

    I know that the most common 3D printer plastics, PLA and ABS, are likely fine because their thermal expansion properties are similar to metals they would replace. This article has a reference to a 3D printed head that was compared to a plastic one (actually only partially plastic; I found a copy of the complete article here:(

    We have employed such a technique to manufacture an atomic force microscopy (AFM) head, and we compared its performance with a copy milled from aluminum. We tested both AFM heads for single molecule force spectroscopy applications and found little to no difference in the signal-to-noise ratio as well as in the thermal drift. The lower E modulus seems to be compensated by higher damping making this material well suited for low noise and low drift applications. Printing an AFM thus offers unparalleled freedom in the design and the rapid production of application-tailored custom instruments.

    But the cost to machine the AFM's head doesn't make it expensive, it is the content of the heads of the sellers that makes the AFM expensive.

  • Yes icspicorp.com is what I was referring to.
    Sorry, I forgot to add that link.

    So the price for the frame-component really is oriented on demand and likely far above potential (minimal) production cost.
    They probably reason that dropping the price (too sudden) won't raise the selling numbers enough to compensate.
    Just as it was with home 3D-printers a few years ago but this time their judgement might be more accurate.

    Even though the integrated SPM chip may have been more expensive in development than the frame-component it seems they want to amortize their development costs mostly with the frame-component. (except the chips - contrary to advertised - break regularly - which I don't really believe)

    I see kind of a dilemma here though:
    Assuming our reasoning that a much cheaper frame-component is possible is correct and someone pretty soon actually makes a ten times cheaper replacement
    with minimal development costs (even in small batches! e.g. kickstarter - legality questions ... ) then icspicorp's small volume frame sells might drop to a point where might not amortize their development costs anymore. And without demand climbing proportional to the dropped price a price battle won't be an option.
    If this goes so far that icspicorp stops selling their cheap integrated SPM chips then the price under-cutter has basically killed its own nourishing ground.

    Is a little less exaggerated scenario (just drops in business not full failure) too far fetched?
    Do they have thought about such a scenario?
    Is there some defense in place (being it intentional or unintentional) ? Proprietary software?
    Or - as I realize while writing all this:
    Is it possible they assume and want early price undercutters to take the risk of testing the market readiness from them?

    The outcome will most likely end up somewhere between the pessimistic and optimistic extreme.
    But I guess I'm just over-analyzing this.

  • Appended to this post is the content of an email I sent to their info@ address. If I get a response I'll post a summary of its content here. If they don't answer by the 8th I'll try to find another way to contact them.

    Since the chips are essentially consumables they should have priced them well above cost to manufacture. The patents on their web site all pertain to the technology in the chips and not to the stage. The assignee for each of the six patents they mention are Niladri Sarkar, Icspi Corp., and Zyvex Corp (4 of the 6.) This appears to be one of the results of a US DARPA funded APM project awarded to Zyvex back in 2008.

    I recently learned of your nGauge AFM and was hoping you would be willing to provide answers to two questions:

    Are there any plans or schedules to offer chips that allow scanning tunneling (STM) or other scanning modes in addition to the AFM mode chips?

    Under what conditions, if any, is documentation on the electrical, mechanical, and signal interface to the AFM chips available so others might develop their own equivalents of the nGauge stage?

    Thank you for your time!


    Jim Logajan

  • I received a reply from the president, Dr. Neil Sarkar. I'm going to go out on a limb and assume he would have no objection to his response being published verbatim, so here it is:

    Hi Jim,

    Thank you for your interest in the nGauge AFM system, I'm happy to answer your questions (apologies for the delay).

    We have a conductive path to the tip in all of our AFM chips, but we do not yet support electrical modes. Our team has produced scanning microwave microscopes, which are on the product roadmap but no hard dates that we can disclose yet.

    We would be happy to share the electrical, mechanical and signal interface with customers that would like to build their own coarse approach mechanisms. One customer has built a wafer-scale multi-probe AFM tool, while another is interested in roll-to-roll metrology. Please feel free to tell us about the application you have in mind if you'd like some feedback.


    I think that is as good a response as can be expected. The price of the chips makes the idea of a hobbyist coarse approach mechanism quite appealing. I'm definitely going to look into it.

  • I just realized that since the last time I've checked there was a lot of movement in the DIY SPM space.
    Specifically beside DIY STMs there's also stuff about DIY AFMs online now.

    The first two nearest things (near in therms of google search metric) that sprang into my face where:
    1) https://openafm.com/
    2) http://www.stromlinet-nano.com/ (~$3000 closed source)
    There's an image of DNA origami made with this microscope: (JUMP TO 0:28)

    At least for DIY AFMs the imaging of DNA structures seems to be on the very limit.
    Maybe this would work a lot better with DIY STMs.
    I think so because AFAIK it is way harder to get down to atomic resolution with AFMs than with STMs due to:
    * the lack of the exponential current decay benefit
    * the tips being fabricated from some material like silicon nitride with more rounded tips then the metallic etched wire tips (true?/false?)

    So contrary to what I thought there's no electrical pre-amplication on the icspicorp chip yet.
    Doing the ultra low noise amplification is a critical more expensive part of the system that I thought could be mostly avoided now.
    So on the current chips there is only stuff for excitation of vibrations for AFM tapping mode ??
    Does that mean for usage in STM mode the current chips are no better than a cut piece of wire?

    "Please feel free to tell us about the application you have in mind ..."
    Hmmmm .... A cheap stage targeting the DIY-geek-tinkerer / hobby-science / citizen-science market?

    Given the not exorbitant or at least not plainly obvious amount of application cases: Would there be sufficient demand?
    Selling it just as "toy" where the usefulness factor is not as important would not require a 10x price drop but rather a 100x price drop - so that's not viable. ...
    Personally I'd like to know about the feasibility of imaging (and manipulation) of DNA structures (no full atomic resolution needed) with cheap solutions.

    * AFAIK a lot of the the (professional) really cool atomic resolution and manipulation stuff was done with high stiffness tuning forks -- non contact tapping-STM I think.
    * "scanning microwave microscopy" -- first time I hear about that -- I guess I need to extend my knowledge in that direction.

    PS: If you really wanna try something and need something 3D modeled for 3D printing I can help there.
    I did quite a lot of fully parametric 3D modelling in OpenSCAD in the past as you can see here:

  • I did reply to Dr. Neil Sarkar saying I was interested in any information they could share. I indicated a general interest in scanning probe nanolithography since that seemed the most relevant use of their product given its resolution and what I would really attempt with such a device. I have not yet heard anything from them.

    I was not aware of the two efforts you found. The ISCPI AFM is about 2.6 times more expensive, but has about 40 times greater resolution, than the Stromlinet Nano AFM. I could not find any mention of cost for tips for the Stromlinet Nano AFM. The ISCPI AFM comes with 4 replacement chip tips.

    The OpenAFM project doesn't seem to have progressed since November 2015.

    I have found a fair number of articles on DIY STM projects. Most accomplished with under US$1000. Atomic resolution attained with some of them.

    Your 3D modeling skills are quite impressive. I have a Robo 3D R1+ but I followed a project that modified the bed after I managed to chip the glass on the original bed. I now have build area of 22 cm wide by 32 cm deep by 20 cm high. The new bed heater can get quite hot very fast. For design I started with the free DesignSpark Mechanical but I really wanted to draw some of the lines and planes of my projects with equations. I also wanted to be able to import and export several file types that the free programs didn't always support, so I finally ended paying as much for the Rhino 3D CAD system as I did for the printer (if I include the cost to upgrade the bed - otherwise the software cost more!) It was definitely a bit of a learning curve after DesignSpark but the free evaluation period for Rhino 3D is 90 days, giving me plenty of time to make sure it would work for me.

    I do have some ideas for an STM I'd like to try but have to work around my day job.

Participate now!

Don’t have an account yet? Register yourself now and be a part of our community!