DIY Laser Lithography: Micron resolution

hey folks today i'm really excited to share this contraption here which is a diy direct right laser lithography machine photolithography is simply a process to create a pattern onto a photosensitive resist a photoresist and most people are probably familiar with pcb development and etching and it's the same idea just shrunk down to a very small scale so that you can fabricate transistors or mems or microfluidics anything at kind of the micron level this is a form of maskless lithography meaning that it directly writes the pattern onto the substrate rather than using an intermediate mask that you first generate and then project through i've been indirectly thinking about and working on a project like this for four or so years it's actually kind of the reason i started this channel way back in the day although obviously i got a little sidetracked along the way so i'm really excited to share this and go through the optics about how it all works it looks complicated but the optical train is actually pretty simple once you break it down into pieces but before we get too far i wanted to show you an example of what this machine can do this is a test grid that's checking different exposure times and focal depth and you can see lines being written into the photoresist these lines are between one and three microns wide depending on the settings and while it's very far from perfect there's a lot of work to be done here it gives you an idea but the capabilities of a machine like this essentially cobbled together with aluminum extrusions and 3d printed parts so let me explain kind of the big pieces first and then we'll turn it on and look at how the laser works its way through the system so it's made up in three chunks there's this optical system here in the back there's the galvo and electronics here in the middle and then there's this front section which has an xyz stage and a microscope objective and this is where your substrate goes so the whole idea is that you put something you want to expose down here and get this laser to bounce all the way through these optics and come out and expose your photo resist let's start with the back of the machine so this is kind of an optical breadboard which holds some of the optics used to clean up the laser beam itself so the laser is just a cheap four or five nanometer uv diode laser and because it's a diode it has a pretty poor beam quality it's very elliptical in shape so we want to clean that up because the spot that comes out of this laser is directly used to expose the resist so if it's elliptical it'll make kind of an elliptical path through the photoresist there's a couple ways to do this but the easiest is to just put it through a small pinhole so this is a 20 micron pinhole on this little metal disc here and that will essentially turn the ellipse into a nice circle now there's haze in the air to help see the laser diode a little bit better but even with the haze you can't see it once it comes out of the pinhole just because the intensity is reduced so much we can put a piece of paper in the way and you can see that it is indeed coming out but it's too hard to see so i'm going to remove the pinhole from the system just so we can see everything work a little bit better the laser diode focuses onto the pinhole and then it begins to diverge as it comes out and we want to re-collimate the beam so this lens takes the diverging laser beam and collimates it back to essentially a parallel straight laser the beam then continues on and hits this dichroic mirror the dichroic is designed so that it reflects ultraviolet wavelength and that continues onwards down into the optical train but anything that's above a certain wavelength will be passed through and you can see there's a little bit that does make it through it's not a perfect reflector but about 90 of it goes this way and not as much gets out the other side the back section of this optical breadboard is not currently in use it was originally designed to be a focus detection mechanism so you can see i've got a little red laser diode here on kind of a kinematic mount that shoots through the dichroic and onto the rest of the system and so that's why a dichroic is used here to reflect the ultraviolet and pass the red laser and the point of this was to co-align the red laser diode with the ultraviolet laser and then send it down the optical system bring it back it would hit this which is actually a beam splitter on kind of a kinematic mount made with magnets hits this beam splitter bounces to this lens and is captured by a camera over here to detect focus it ended up being too difficult to get all this aligned and i didn't want to mess up the alignment of the ultraviolet laser so this back section is not currently in use and i'll talk about how to fix this a little bit later there's a better way to do this up closer to the kind of front of the optics so on the side here we can see the dichroic filter is over on this side and it's bouncing the light from the optical breadboard over to these two mirrors you can see it's walking its way over the objective is to get it into this small hole over here and so to do that we need to use a configuration of these two mirrors angled at kind of opposite 45 degree angles the laser beam is coming out at this height but we need it down at this height and so this configuration which is known as walking the beam is a common way to change the orientation and height of a laser in an optical setup so it bounces over hits this 45 degree and shoots straight down and then from this mirror it hits the other 45 degree mirror and bounces in to this little cavity right there from the front of the machine we can start to see what's going on inside this cavity there are two small mirrors that are attached to two very lightweight servos and this forms an xy galvo scanner and as the two mirrors rotate and move it changes the position of the laser beam coming out of the galvo so we can see the laser bounces into the cavity from here hits two mirrors in kind of opposing configurations just like these two mirrors here and then is redirected out of the front you'll note that the beam quality is pretty garbage coming out of it and remember we removed the pinhole so that this is more easily seen with the pinhole in place this is a considerably better it's also a little bit out of alignment just because i've been shifting things around and everything wiggles a little bit so the alignment here is not perfect i'll show some footage of what it looks like when the pinhole is in place and you can see that the beam is considerably better let's shift it around a little bit more and see what's going on with these lenses in front and i've had to turn the lights down even more so that we can visualize the beam better and just because it's getting increasingly harder to see up here but in front of the galvo scanner is a 50 millimeter plano convex lens this is placed so that its focal point coincides with the center of rotation of the galvo meaning that this is 50 millimeters away from the galvo and then up here we have another lens of plano convex which is a hundred millimeters in focal length so these two are spaced 150 millimeters apart and this forms a scan and tube lens configuration i'll talk about this scan and tube lens configuration in a minute because this is really important but let's finish the overview first the beam comes out of the galvo goes through these two lenses and is redirected by this 45 degree mirror again on a kinematic mount which shoots the laser beam straight down into this microscope objective which is mounted into a little diy kinematic mount so that you can adjust kind of the tip and tilt of the objective itself that then emerges from the microscope hits your substrate which is mounted onto this xyz stage so you can adjust where everything is and then the holding is kind of just hanging out the end of this extrusion a little wobbly but it works so let's take a closer look at this front optical section to simplify things let's ignore the 45 degree mirror to start and just put the microscope objective in line with the laser galvo we can imagine that the microscope objective has a pivot point at its back focal plane and this is the location where light needs to enter it to be focused down to a point the microscope objective will convert angles of light into linear motion what i mean by that is as light enters the microscope objective the angle at which it's entering this pivot point will be converted into a linear deviation of the spot unfortunately the laser galva works in a pretty similar manner that we tell it to draw a line and it will rotate the mirrors and shoot the laser light off into all kinds of different directions so we need to couple the galvo to the objective using two lenses in a relay setup the first lens is placed so that its focal distance coincides with the center of rotation of the galvo this will take the various angles that the galvo emits and straighten them out converting essentially the angle into a deviation in x or y so it's converting from angle space to kind of cartesian space but we need to get from cartesian back to angle space as centered on the pivot of the microscope objective and that's what the second lens is used for and this will take the light and refocus it down onto the pivot point of the microscope objective so you can see that as the laser galvo sweeps its angle from top to bottom this is converted into a linear displacement from top to bottom between the two lenses and then as we emerge from the last lens we're converted back into a change of angle which coincides with the back focal plane of the microscope objective one final thing to note is that this relay system operates as a very mild beam expander as well because the focal lengths are different between the two lenses 50 millimeters and 100 millimeters there's a two to one expansion of any beam that's being relayed through it so a five millimeter beam coming in will leave as a 10 millimeter beam this is helpful because microscope objectives give the best resolution when their entrance pupil is completely illuminated the entrance people on these objectives are about six or seven millimeters and the beam that comes out of the laser is only about one millimeter in diameter maybe two millimeters so giving any kind of beam expansion helps improve the resolution by filling up that entrance people a bit more it's still too small i really need a larger expansion ratio here but given the constraints of the system and where all the components were this is about the best i could do cool let's jump in and take a look at some of the results so my first batch of tests were done using a commercial easily accessible photoresist for pcb development daytech er71 and it is a negative photoresist i unfortunately didn't have great results with it and i don't know if it's due to the age of the resist mine is a couple years old or if it's just the resolution limits of this particular formulation when a negative photoresist is exposed the portion that is exposed to the ultraviolet light will cross-link more heavily and basically becomes insoluble in the developer so what we should be seeing are lines of photoresist that are left over and then bare substrate where all the rest of the parts that that expose gets washed away and we unfortunately don't really see that the parts that are exposed do stick around but you can kind of see a lot of pitting and etching through the part that should be insoluble and then the portions that aren't exposed and are supposed to be soluble on the edges don't really seem to get developed fully and you get all these little globs sticking around so i tried a bunch of variations and i really just couldn't get this to work fortunately a couple years ago i snagged a bottle of commercial photo resist off ebay microposit s1818 and it's a positive resist that's designed for this type of resolution and application and immediately i saw much much better results this is a positive photoresist meaning it's the opposite of the daytech so anything that's exposed becomes more soluble in the developer and will wash away so you can see on this test pattern that all of the parts that have been exposed are trenches in the photoresist that have been etched away to varying degrees this test was done with a 20x objective and we can see a few different issues here most prevalently is that it did not fully develop or etch through the photoresist so the center of the sunburst is actually fully etched down to the substrate surface but everywhere else is uh not fully etched but if we zoom in on the part that is exposed we can see that this line is about seven microns in width which is a pretty good for this system if we look at this sample this is done on a 40x objective and you can see tons of wobbling on these lines so if you look at the center number 13 for example the lines right above it are supposed to be straight and these are not problems with optical aberration what you're seeing is just the galvos being unable to maintain a straight enough path to keep a straight line so this is just purely an issue with the hardware or maybe the controller of the galbos this particular test was looking at exposure times so 0 through 15 have different amounts of exposure of the laser light number 15 has the best exposure where more of the substrate is etched away if you look at the other samples they're not fully etched through i also ran focal distance tests with this particular sample the 40x objective has a 0.5 millimeter working height and the focal depth is super shallow so if you're not exactly on 0.5 millimeters the spot will grow very quickly and you can see that just a little above or a little below the right focal distance we don't get any type of proper development it's just kind of a blob of photoresist that gets indiscriminately etched we can zoom into the corner of this four and take a look at the actual etching profile of the photoresist under afm to get a better sense of like what's the sidewall steepness look like and that can help us assess if we're under or over developing or under over exposing and we can see in the afm that there's a few places where it's etched fully down to the substrate but in most places it didn't quite get to the substrate and the side walls are not as straight or sharp as they should be if you look at the data sheet for this photoresist it shows very nice you know perfectly 90 degree sidewalls whereas in comparison my side walls are very sloped and rounded some other afms that i've taken on different regions or different tests we can see a better sidewall so this is just a process variability issue that needs to be sorted out we can also see that the galvo struggles with repeatability if we try to trace the same path multiple times the laser should coincide with the prior exposures but we can see here that it very much doesn't there are multiple lines being exposed this effectively limits us to single passes and altering the speed rather than altering the number of passes with the 40x objective it looks like we're getting between one and three micron line width kind of depends on which lines you look at how much it's exposed first under exposed or overexposed but roughly in the range of one to three microns and i think probably two to three is a realistic expectation of this machine once everything gets dialed in now it's functional and i can get pretty good resolution when all the stars align but it's far from reliable and there's a couple reasons for that the galvo scanner has a lot of problems so a scavenge from an old 3d printer and it has a lot of baggage associated with that it accepts a very limited and kind of crude subset of g-code it doesn't accept feed rates for example so you have to set a global speed for the laser exposure the galvo itself turns out to be just a cheap 30k points per second kind of laser light show galvo it's nothing special they just repackage it into this unit even though it does actually have a very nice custom machined aluminum fixture for the galvo and the lens i didn't end up using any of that and so it's not important to me so the galva really should probably get replaced with either a better galvo that has finer control or just a diy you know galvo that i control myself rather than interfacing with a kind of kludgy 3d printer all these plastic parts are problematic they were convenient from a rapid prototyping perspective i could easily print things up get them into different arrangements to figure out what kind of optical train i wanted what worked all the different distances and just kind of get a sense for how everything fit together quickly and easily but plastic is not ideal for a precise optical alignment the thermal coefficient of the plastic is a lot different than all the metal parts so any temperature change will see the plastic growing and shrinking at a very different rate than the metal and more importantly all the components like the kinematic mount out here or the kinematic mounts on the sides that are holding weights kind of at a distance these plastic parts will creep and flow over time and you'll slowly lose the alignment and what that means in practice is that every few days you need to tweak the alignment of all the different components to get it back perfect down the microscope objective which is not a big deal for a test machine like this but something that you would want to just roll up turn it on and use not a good situation so at least some of these plastic parts should be replaced the overall layout is just bad like you might be wondering why it's set up this way where it's off the table and there's like an optical bench in the back and then this little stage up front just kind of hanging out there this was an evolutionary design it wasn't supposed to look quite like this things got added and removed and changed around so originally this whole optical back part was supposed to be up here as kind of a top thing looking straight down or off to the side going right into the galvo that was too hard to get working so i end up just mounting it to this old fixture plate from an old milling machine and this region out here in the front there's a reason for that which is this this is a really nice xyz wafer stepping stage that i got off ebay a couple years ago and it's just been sitting around waiting for this project to get finished it has sub micron movements and it was originally used to step a wafer under some type of inspection device in a semiconductor fab so the original idea for this device which is why it's set off the ground is so that the wafer stepper can go underneath the microscope objective and allow me to stitch together multiple fields of view as i mentioned briefly before the back section of this optical table is supposed to be used for focus detection honestly it'd be much better if it was closer to the objective there should be a beam splitter right before the objective which is where the laser diode and focus detection happens rather than way back here because it would make alignment just considerably easier and then finally there's just a lot of like minor misalignment problems none of the lenses or mirrors are perfectly on axis with the beam meaning that by time the laser spot gets to the very end it's got a fair amount of aberrations that's picked up there's some astigmatism it's skewed a little bit like it's not perfect and you can see that in a lot of the photo resist trials where especially when it's a little bit out of focus you get some weird development on the edges of the line and that's from a lot of the kind of aberrations that are picked up along the way make sure you subscribe so you can see future iterations of this machine or projects done with a machine like this i have some interesting ideas about making kind of mems devices to show off how like an electrostatic actuator works or a thermal actuator and i think they'd be doable with the kinds of resolutions that i'm getting off a machine like this if you'd like to support the channel financially i do have a patreon where i post behind the scenes of machines like this or things i'm working on in the shop and i do also have an archive of all the sem and afm micrographs that i take and there's a lot of ones that i don't show in the channel because there's not time for it but if you want any of them for like wallpapers feel free to join up download them print them do whatever you want i think that's all i got for you today thanks for watching i'll see you next time you

Channel: Breaking Taps

Views: 170,134

Keywords:

Id: XJ9w7DjI_Bc

Length: 21min 48sec (1308 seconds)

Published: Sat Mar 19 2022