It's nice to see you discuss this topic! My first internship had me working for a startup that had designed a Lab-on-a-Chip that had the purpose to bring Lithium measurements for bipolar patients to the point of care. Bipolar disorder is treated by using Lithium, but it has a narrow bandwidth within which it is effective. Too little, and it doesn't have an effect. Too much, and it can be poisonous. Patients treated with Lithium are therefore typically confined to staying close enough to their hospital to regularly have their Lithium values checked. The LoC developed by the startup company I did the internship for brought this check to the patient. By just applying a drop of their blood and waiting a few minutes, the LoC would use the principle of electrophoresis to separate ions in the drop of blood and then determine the level of each type inside the blood, Lithium in particular. While the results were mostly for indicative purposes only, it did allow the patient a lot more freedom of movement. Only if the LoC results went 'out of bounds' could they consult their doctor to take immediate action, but the patient could be proactive about it and do so at home, or while on vacation (which the LoC allowed them do take!). It may be a small application without a big market behind it, but I do believe that the quality-of-life improvement this particular LoC offered is more than worth the investment in the techniques behind it.
I'm a strong believer in the "KISS" principle: keep it simple, stupid. The problem in today's academia is that researcher's main task is to get published, and get published in the most prestigious journal, if at all possible. As a result, they are too often compelled to do "hero" experiments, that are extremely complex, and require multi-million dollar labs full of the latest equipment. When it comes to transferring that to a commercial product that is low cost and easy to mass produce, they are completely clueless. They are just not trained to think about those things. Successful commercial products generally adhere to the KISS principle. The glucose test does one thing, and one thing only. I'm currently working on a (very) low-cost, optical fiber sensing technology. The couple of papers I've published have attracted next to zero attention, and I'm not surprised, and yet I can get 100 times better resolution than all other papers in the field, at a 1000x smaller cost, but talking about cost in an academic paper is a big taboo. You can talk about "potentially" low cost, but an actual low cost, high performance device won't get you published in Nature Photonics, because there is no "big science" in it. Making things simpler is just not considered "advanced" science. This, despite the fact that technologies such as the IC started off as very simple devices. In fact, it is their simplicity, compared with bulk circuits, that made them so successful. Complexity was added later, but only because the process was, at its core, very simple. If the technology is complex and difficult to start with, it will never make it commercially because you can never get to mass production and high yields. So I think microfluidics is suffering from that disease. I've followed the field for 20 years, and considered using the technology, but it is actually very complex. I will stick to my little fiber optic sensors: easy to make, easy to mass produce, and extremely low cost due to all the cost reduction brought about by the massive deployment of optical fibers in the past 40 years. But I can tell you, convincing investors about it is THE most difficult thing in the world. They'd rather go with a fancy AI app. One more.
this comment, a complaint about research impact and publication bias in a field that isn't even microfluidics, has an order of magnitude more likes than any other at present.
@@lucas2nded461 I'm not even unsympathetic to the basic point -- I do some work on democratization and access projects for frugal biotech -- but the commenter's comparison to photonics seems very hollow. Applied life sci and physical sci are subject to substantially different pressures. There's actually pretty strong momentum at present for papers on cheap and open biotech. Even if it's only begrudgingly acknowledged by the academic publishing cabal.
I tried doing some electrophoresis nanofluidics during my PhD and one of the biggest challenges was bonding my top substrate to my bottom substrate. I really needed two glass substrates to allow the electric field to penetrate the liquid, because a silicon substrate would just cause strong capacitive coupling. And bonding two glass substrates with a nanoscale cavity in between is really really really hard! First of all: both of your substrates have to be incredibly flat in order to allow cavities to form. Second: you need to deal with the wringing effect that also occurs in guage blocks. Third: you need to evacuate the air between your substrates during bonding, because you always need a thin substrate to allow you to use a 100x microscope and those substrates flex quite easily. Fourth: most literature on substrate bonding is a variation on anodic bonding, where an electric field assists in pulling the two substrates together, but it always requires some electrode nearby. In the end, the only viable candidate I found was sodium silicate bonding, the same technique as the one that was used in the experiments of a Nobel laureate (the experiments in question weren't the ones he had won the Nobel prize for though). As luck would have it, my professor managed to get some students of this Nobel laureate professor to give us a Zoom seminar on their work. When I asked about details and recipes on the silicate bonding, professor X got his knickers in a twist about me trying to get valuable IP from his students (said recipes had already just been published in an addendum to a PhD thesis, his student just sent me a link to it afterwards). 😂 It was both hilarious and sad that a Nobel laureate from a top-5 ranking university was trying to hold back published information from a first year PhD student - from a university that's over 50 places lower on the Shanghai ranking list! - for fear of his own lab losing the monopoly on successfully pulling off a technique. That was the day I learned that the system of "first author gets all the credit" impedes the entire process of reproduction and building upon other people's work. Other than the bonding, I'd say that a major impediment to microfluidics research is the fact that small scale production of chips in a research clean room facility is prohibitively expensive time wise. In order for a researcher to freely experiment with a device, they need to know that they can easily procure a new one if they contaminate or break the device they have on hand. As a PhD student, you have to design, produce, test, experiment with and clean each and every single device you make. And if you are to stand any chance of getting published in today's climate, you need positive results on a novel enough experiment. And if you want to meet your publishing quota, you need to do most of the work all by yourself. That's the real reason why micro- and nanofluidics are not taking off: it takes way too damn long to build enough devices to get to a workable experiment. I also concur: lab-on-a-chip is not happening anytime soon. The natural development path really is to have more and more chip-in-a-lab devices, which might one day be integrated into a lab-on-a-chip. But you need to prove and improve all the individual components before you can assemble them into a VLSI-like system.
People here have been bonding microfluidic chips together comprising glass and composite substrates with nanoscale channels for years using HSQ ebeam resist. Many of these siloxanes are optically transparent and can be used to seamlessly bond most substrates together very well so its unfortunate to hear that they even think this is anything special in terms of IP or knowhow. Much of the problems I see with scalability of research chips for experimentation is poor planning in my opinion and lack of sufficient computational resources or related to simulation students will not have the ability or foresight to design their simulations in such as way that allows for parametric iteration of key feature geometries when referencing rebuild models and constraint specific meshing densities. With access to a decent cluster and suitable software package you can run parametric optimizations of vast feature libraries so long as the key geometric aspects defining the parametric of the model are adequately parametrized and are referenced to a performance simulation output adequately. Even with a simple algorithm you can build something that varies the permutation set in the direction of optimized performance until a target is met. When it comes to experimentation with real research models, good planning is the best approach. Knowing exactly what scope of feature geometry permutations you want to test, and generating a mask comprising those libraries of designs from the start will save tons of time when done at wafer scale. Often students are too short on time and energy and just want to jump into fab without much planning or mask design and end up taking far longer than necessary. I have seen projects that take years to complete and never go anywhere get completed in 6 months with proper planning and permutation evaluation strategy. Even highly divergent cases, you can often sample a broad permutation space structure library using primitive sub elements of a design that can be stitched together during lithography (assuming you have a basic R&D stepper or direct write tool) into a broad combinatorial permutation set to be evaluated. I have seen such methodologies implemented in a practical directed optimization research study in which a large parametric design space was optimized down to a small number of highly efficient prototypes with record breaking Dean Flow separation capabilities as just one example. Its unfortunate how different the level of capability and knowhow can be in various labs across the planet. People hold back a lot but there are usually things that are going on light years ahead of what you might be aware of.
@@Exascale the IP I was talking about concerns recipes for doing sodium silicate bonding. Sodium silicate bonding is incredibly finicky, because sodium silicate isn't shelf stable, so it really saves you a lot of time if you can start from an exact recipe, rather than the laughably vague methods and materials section in a paper. The recipes were published as an appendix in a PhD thesis. It was so fucking petty. Concerning the planning and methodology, in my case, there was just an incredible lack of supporting infrastructure. My supervisors had no clue about planning and they viewed it as superfluous. In terms of fab know how: we had no fab technician on staff, so all important know how had to be transferred from PhD student to PhD student and there was a lot of know how being spilled on every handover. Plus: we were using machines and bays from other research groups, so we'd always be last in line for timeslots. Concerning the parametric design in simulation: absolutely, this will speed up your work a lot. Under the condition that you have your very first proof of concept run of the actual fabbing done.
@@LReBe7 Ahh yes the flatness issue with bonding. Yeah what a nightmare that can be if your starting wafers are not near perfectly flat. You can get away with about 10um or better TTV in a 150mm CZ-Si wafer depending on your cavity density and layout but for glass you have to be within 5um TTV for a reliable bond. Even then without 1um TTV you are still going to often lose some die. Sounds like you had a nightmare of a graduate experience which is not uncommon. Often advisors get into a tenured position without understanding the full scope of what they are managing within their own labs. This can create a disconnect in expectations as well as respect for the needs required to get things done. Most labs are underfunded and most students and post docs are just trying to get through their project asap so they can graduate and move on. The deep thought and genuine curiosity has been replaced long ago with the anxiety and stress of staying afloat, graduating, or getting that coveted recommendation letter. There is so much corner cutting and embellished data. Countless SI's that are nowhere near reproducible. Countless images of the one perfect square um area on a 1sq.cm die of garbage. As everything gets more economically constrained its only getting worse. Even at top schools its getting very bad. Best of luck and hopefully you have made it to a place that is not full of fakers/bullshyterz. These days more than half the students I see coming out of universities dont know a thing about much yet have excellent GPA's and LinkedIn profiles that make them appear like the next up and coming nobel laureate. Society is crumbling as a result and every year that goes by there are fewer and fewer talented folks that can execute.
I worked in years for microfluidics (under dr Ramsey and Henly- check the lit they’re big guys) I was an undergrad at the time so no papers for me. I don’t really think this is more than a dead end frankly. I tried so hard to solve a few problems. There are some surprising issues that make this much harder than you think. Though it’s fascinating. I’m sure some of the early IC guys lost faith too. So take my opinion with a grain of salt. I worked on 2 phase separation - the first separated on hydrophobic proteins vs drag then injected into another CE separation where it separated based on charge vs drag. The big problem is the mixing problem and the non-orthogonal nature of those two methods. And it is surely not “self powered” We ran 10kV through it. In other words a crap ton. Small currents. But you require the voltage to create the electro-osmotic flow. The glass (borosilicate) becomes all negative charged because that’s what borosilicate does. Then the water being a dipole lines up the positive part to interact with the walls, all the positive outside water molecules run to ground . Because they are all the outside water in the “channel” they drag the rest of the water with mass flow. So the channels have to be small. We separated proteins from single cells. Normally if you do genetics on a biopsy you are analyzing the genetics of ALL the cells which are all doing different things. Doing a single cell let’s you see what one cell is doing. We also did DNA sequencing, by passing dna through a small pore and reading the resistance as it passes through one at a time. DNA has a negative charge so you can do this. But clogging. Dear god clogging . 95% of my time was Declogging my chips. It took days to make a batch of chips By hand buy using old photolithography methods. If one clogged, which happened all the f-Ing time it could take days and days and days to unclog. Maybe it was dust. Maybe some of the salt came out of solution in your buffer solution, you would stare under a microscope for hours and hours and pass solvents through the channel to try and dissolve the clog and it never freaking worked
I’d be happy to have my “paper” (honors thesis) that explains things at a much lower lever since I started from fundamentals and worked My way up. Mostly because I had a lot of pages to fill, wanted to show I understood the science, and the post doc assigned me a problem he couldn’t solve which was sorta a jerk move. So my “paper” had no “nove” findings and useless for anything other than the honors on My degree. Anything I tried he’d condescendingly say I tried that. So I would scour research looking for ideas and I’m convinced the problem isn’t solvable on the purest sense. Smaller gate channels might help but then you run into lithography errors cutting such narrow channels and even more clogging. If only it didn’t take weeks to make more chips.
is it possible to use a normal inkjet printhead (the resolution is already in the 10s of micrometers) and etching to produce micro fluidic devices? similar to how pcbs are made?
The biggest issue with microfluidics in the diagnostics (DX) space, excluding the limited cases where they are already used, is that they don't really do anything BETTER than current technologies. Most current DX systems already handle essentially all of the sample prep, sample control, and specific tets run. The operator just checks the tube of blood, scans the barcode, and inserts it into the system. The system handles the rest, right down to uploading the results. Microfluidics emphasizes the concept of multiplexing tests, but there is little to no value in running unnecessary tests. An assay that runs 100 different tests when only two are needed is only worth those two tests; the other 98 are just wasted costs. Most modern DX systems will multiplex a handful (up to a dozen or so) of tests together which would logically be run together (e.g.multiplexing IgG, IgM, and Antigen tests for a particular infection). There are only so many tests that can be logically grouped together. Lastly, for some DX tests, the usable concentration of analytes is so low that you can run into issues with minimum sample volume. This is especially true if a quantitative or semi-quantitative result is desired, rather than a simple binary result.
That reminds me of how most of the successful commercial development in the parallel computing space so far has been in single instruction multiple data (SIMD) processors, rather than multiple instruction single data (MISD). The former performs the same operation on a hundred samples simultaneously, while the latter performs a hundred different operations on the same sample. Most of the time, the former is more useful.
"Better" comes in a lot of ways. The lab on chip is dirt cheap compared to big machines. They also require no specially trained people to go around fixing and calibrating. No one going around cleaning and sterilizing. Then there's the drastic reduction in electricity needed to run tests. And let's keep in mind that if you have a box of 500 loc's, you can have 500 people being tested at once.
Not only that, but all the money they raised itself didn't go for startups that could have gone right. They took also time from that industry to move forward
Absolutely. A relative of mine works at an (actual) blood test company and they are very interested in solving this problem. But since Theranos, it's become incredibly hard to get funding for it, even if you try to solve it legitimately and have the track record to back up your claim of knowing how to develop and provide blood tests. Kinda sad because the technology has a lot of potential.
I'm not disagreeing with you, but I had never heard of Lab on a Chip - even while Theranos was advertising - until it blew up into a huge scandal that every news report covered. So, while it did some damage in the short term, it may have actually helped the industry long-term by bringing awareness of the technology to the general public. When I started looking into Theranos to see what the story was about, it was pretty apparent in most coverage that I read, in spite of the fraud by Holmes, that the science behind the technology was in fact valid and being pursued by some of the world's premiere research institutions. Now I try to keep tabs on the industry, excited for the next real breakthrough.
I think the tech-knowledgeable investors are still going at this field. There have been quite a few impressive breakthroughs in specific areas of testing since this affair. Nothing as ridiculously grandiose as Holmes' BS. Theranos duped a lot of outsiders who's looking to make $$$. Notice she either failed or didn't go after health-technology focused investors.
@@JustJustSid If they had anyone that was truly educated on the science of the issue regarding blood testing and analyte detection then they would not be interested in solving this problem because it is virtually impossible for most screens of interest. The fact that there is interest in my opinion would be a red flag that they are not managed properly by scientific first principles. The reason such technology does not exist has nothing to do with sensitivity. We have had single molecule sensitive assays for decades now. The issue is the statistics of the problem within the blood mixture, concentrations, and probabilities. For most analytes of interest you might have 1 molecule in a 10mL blood sample. You might draw 10mL and detect 1 molecule or detect 2 molecules of your analyte or detect zero with 3 sequential blood draws! So you see the pipedream of detecting any disease in a drop of blood is beyond folly. It is an absolute absurdity. No scientist worth his credibility should ever even be entertaining such research goals. If you see it you know they are either not scientists, poorly trained scientists that have no brain, total idiots, or charlatans. The fact that Theranos got as far as it did is a testament to how much dumb money there is out there being thrown at people with letters after their name that came from anything remotely related to Stanford. What I have learned about people from stanford is that they are really good at marketing and bullshyting. They are professional bullshyters with no fundamental ability to do anything useful in reality. They will extract your money and charge huge salaries all while giving TED talks posted on LinkedIn on how smart they are.
There's another lab-on-a-chip concept that involves more biological instead of physical/chemical interaction. Growing tiny tissue clusters and then subjecting them to specific treatment (medicines, bacteria, chemicals, etc) produces an observable output that can help accelerate and unblock several new therapies that are either too expensive or risky to trial on human subjects.
you mentioned flow cytometry briefly but it's a really interesting field, my mother worked as researcher for one of the most prominent experts in that field for over 20 years. flow cytometry is most useful because it can quickly and reliably count cells and other objects when they have a low concentration in a fluid that makes other methods impractical.
We have that iStat device at my hospital. It is useful but I wouldn't call it "lab quality," it's mainly helpful for giving us an approximate answer very quickly e.g. when we have someone come in who might be in diabetic ketoacidosis and we want to do a blood chemistry. We still pretty much always send blood to the regular lab for definitive results.
I'm uncomfortable with manufacturer's claims of "Lab Quality". They muddy the water on these issue and it is interesting Abbot's i-Stat product webpage plaster a warning to me "BASED ON YOUR CURRENT LOCATION, THE CONTENT ON THIS PAGE MAY NOT BE RELEVANT FOR YOUR COUNTRY." Which probably meant Health Canada called their bullshit and restricted their approved labeling for medical device. I'm reasonably confident that their equipment can do a good job of accurate measurement ...but what is the quality of what they are measuring? I am not confident that a few drops of whole blood can be reliably reflect the actual intravenous status of patient. A pointless test given the sample limitation IMO, hope they put more money into speeding up and reducing the cost of actual lab testing.
Currently working on organ on chip model, and the chip in the lab is such a true thing, the amount of equipment you require to develop one, its better to optimize conventional methods but LoC holds great potential for diagnosis in the medical field. Great video by the way
Hey this is what I am working on at my job! We're a start up called Lightcast Discovery and I work there as a software developer, the potential is amazing. Keep an eye on this field around early next year, we'll be making some very big announcements.
Just for some anecdotal historical perspective, the potential of LOC was amazing when I was a software developer in lab automation... 23 years ago 🙂 Best wishes for the success of your endeavours!
@@JonWallis123 Everything is impossible until it isn't anymore :) I wish I could say more but due to the NDA, I cannot, but what I will say is that what we are doing was absolutely not possible 23 years ago, which is what makes this so exciting. If the potential was amazing then, just think of what we can do with it now! But we are building off the hard work of people like you, so thank you for the encouragement. I remain hopeful that this succeeds, not neccessarily for the companies profits but for all the good it will bring.
I've been aware of MEMs, photonics, and (non micro) fluidics, for decades. This is a great intro, to microfluidics, for a general audience. Like most things, its more evolutionary, than revolutionary. Hopefully, Theranos did derail things, too much. Still a better investment than crypto.
This is a great and succinct summary of the technology and industry! I studied this for years and it’s as good as many introduction chapters of phd thesis I’ve read!
Lab on a chip already exists. A number of POC devices uses LOC technology but often it's fairly basic stuff. A big issue with LOC is typically the price of the device. Commercial lab analysis is all about price and just reducing the amount of chemicals necessary isn't saving enough money.
Great video about an interesting topic, as usual :) At first glance, deterministic lateral displacement can be easily explained by fluid dynamics, or just make an experiment yourself: On a stormy day, try to shield yourself from the wind by standing behind an advertising pillar - or any large, perpendicular, cylindrical object at your disposal. You will find, that the wind speed behind that object is in fact larger than the wind speed blowing at its front. Due to Bernoullis law, an increase in speed of a fluid results in a decrease in static pressure. An object subjected to different levels of pressure experiences a force towards the lower pressure zone (think of buyoyanci). The magnitude of that force is determined by the size of the area affected, therefore large particles experience a stronger force than smaller ones (assuming spheric particles, this factors in with radius squared) But the inertia of the particles counteracts this force. The inertia factors in with radius to the third power, therefore the resulting acceleration of smaller particles towards the zone of lower pressure is larger than that of larger particles, assuming their density being similar.
You know there's a crisis brewing in science, when a random RUclips comment can explain in simple terms what top scientists in high paying positions can't ;)
Well done review. I worked on the product development that used a few technologies you described in the video with some success. But "thanks" to Theranos it was impossible to get funding for further development. Over all very accurate review.
Corning developed microfluidic devices in glass as early as the mid-1960s for use in pneumatic logic circuits. They sold them into the late 1980s if I'm not mistaken.
Hi. I have been very impressed with the quality of the videos that you have been producing. I am very keen to learn what are the questions that you pose to yourself when you research an entirely foreign and new topic? You would try to dig into these topics to what extent? I would be much obliged if you can make a video to discuss the methodology you adopt in researching a topic. Thank you very much.
Thanks, this is a good review of the field. As a researcher in this field, I can say that the technology to miniaturize regular tests is there but the big question is economics. This is especially important when i-Stat already offers a lot of these advantages. Instead of making lab-on-a-chip, microfluidics will evolve to process biological samples which are erstwhile hard to process or result in poor yield by conventional bench-top technology.
We are growing organoid in the chips and the major breakthrough was when the 3D printers got sufficient resolution to be able to print the molds (200um features). Otherwise there is absolutely zero tasks for microfluidics which can’t be done on the multiwell plates.
Before I started the video, I wanted to comment immediately about Elizabeth Holmes. 0:44 seconds in, and a picture of Holmes shows up. I cannot get Elizabeth Holmes out of my head whenever microfluidics is discussed.
Some of the principles you mentioned have entire books written on them that are very thick and very dry. Great job on zeroing in on the important stuff and bringing the whole story together in such a short video 👍
really liked the general overview of the whole topic, and while you touched on digital microfluidic biochips, it's a whole interesting concept in its own, it addresses a lot of problems in traditional microfluidic biochips, but also have its own set of shortcomings, a lot of good work is being done on it now.
I am in awe of your overall knowledge of all science and technology. Me on the other hand, can say that tech and I are not the best of friends, however I enjoy when I can feel I comprehend sophisticated stuff so I keep my interest high. That is why I have been very interested in "truth seeking for dummies" which at its core among multiple controversial conspiracy theories, tries to analyze things at the basic level applying mainly facts, experimentation and most of all common sense. Based on that premises, in the past few years I still have never seen real photographs of the earth that can be genuinely proven taken from space. Further more the photographic evidence provided by nasa shows we have never landed on the moon nor it can prove the legitimacy of most of the space missions going on in the past 6 decades, and when you link the dots and follow the money and political trail, all indicates deception and shady institutions running the show of making people believe myths, stories and fake narratives. What I am implying is that sometimes sophisticated science at all levels serves as a diversion or distraction to keep the people dumb and their behavior automatized and predictable so they can remain asleep. What about you, other than being good at all science, how far are you willing to go to dispute or debunk the establishment if you were certain you have been lied to? Would you even dispute common knowledge if you knew its false or if there was a great degree of reasonable doubt in the way the narrative has been structured or taught. What seems obvious and undisputable because we are such an advanced and capable species therefore the premise is to never dispute the established science of centuries simply because we are standing on the shoulders of giants past and present, could be a wrong assumption. Anyway, let me remind you that gravity is still a theory, the laws of thermodynamics are always broken to explain how our world functions and for some reason blue origin and virgin galactic can only fly sporadically to aprox 60 miles altitude to claim they take few passengers after 25 years of research and development. By the way those supposed few individuals only claim to see darkness up there. Should we investigate or be satisfied with what doesn't make sense? Well, before we continue to understand the complexities of our advanced civilization don't you think that some introspection is required to answer some really dumb questions, and to demand explanation for millions of people waking up around the world that understand that a lot is hidden from them. Finally, if you know things but keep them hidden away for personal convenience or you haven't put enough effort to investigate things you know dong make sense, does that make you a fake. ? If there even less than 1% of doubt then that is too high of a percentage to just go along with the flow.
Thanks Asianometry, great Video. As a scientist working in the field, I like to make you aware that there a many microfludic based test for many diseases these days. For HIV here is a list from WHO: extranet.who.int/pqweb/sites/default/files/documents/211220_prequalified_IVD_product_list.pdf . Nearly all modern automated devices for medical diagnostic are using internally microfludics. However the periphery is usually makro sized, so that you end up with a bench-top device an not small chip. However I totally agree today we have a "Chip-on-the-Lab" more than a "Lab-on-the-Chip".
Thank you Asianometry, for introducing this technological concept of microfluidics to me and perhaps some others seeing this well produced short video for the first time. This topic is new to me.
Nice video, I know the video have a limited time, but maybe discussing drop based micro fludics was worth noting as it has the potential of being conuntitative analysis for the samples.
An established market, but lots of potential here. Would be great to see more of these LoC's in the household to take away some of the "market" of medical professionals.
We have a NanoOne 3D printer to custom make chips in our lab, the integration to the rest of the processes is super simple since the new chips have pumps and sensors integrated and all connectors are printed instead of glued on ! Just like MEMS packaging it solves the integration for us.
Just a suggestion, but I'd LOVE to see a video on josephson junctions and the josephson voltage standard. I think something that's so obscure, but so incredibly fundamental that it defines something most of us work with on a daily basis warrants a video. =) But I'm sure you've already got a huge list of topics you want to cover.
Hello! Thank you for these videos. I find them to be very calming yet very intriguing. I often have problems staying focused on any type of content, but these videos are very interesting and keeps me glued. The topics you bring up are great and you always seem to be well versed yet down to earth, really makes me curious to learn more about the topics. Keep it up :)
The number of papers is far higher than 2000, that figure was year on year. It's in the tens of thousands. Otherwise nice video - I'm a biomedical engineer working in this field.
Microfluidics was used for high throughput RT-PCR testing for SARS-CoV-2/COVID-19. I think it worked well in this case because there was a need for large batches of a single test. Far from a lab on a chip, but there have been some practical applications.
@Asianometry i have a question and i would appreciate your answer. Do you research about the subjects you make videos(sepetially the ones on tech stuff) or is it part your field of study, and if it is the latter and if you dont mind whats ur field of study/research/work, because u seem well rounded in many topics(espetially microships, but also material sciences ,optics, pharma, tech, ai.. and even economics) but you also able to post frequently which i suppose leave you with little time to possibly research ur videos ?
Possibly, it would be worthwhile to have one chip to perform all of the tests that are now performed on different chips. A single fab is less expensive than several even when it fabricates a more complex chip.
For classic organic synthesis this method gives high yields. Compare it to a bulk method...every molecule have to shake hands inside a big hall, but if you restrict them to a hall-way, it gets easier.. aka higher yields
It think that a lot of people including you look on this topic from the wrong perspective. If there are chemical tasks to be done then the (main) functionality does not lay in the chip itself but of course the chemicals used and this is the area where research has to go into and not the chip itself that is just performing supportive and/or protecting tasks. This is the case because as you rightfully point out a lot of tasks in both detection and especially in preperation just cant be provided with a chip and need a lab anyway. The only way to work around this is developing analytical chemicals that perform these tasks since they are the only part here that can be scaled down with no differences/ loss in functionality. But since investors, politicans etc. just see the word "chip" and ignore the "lab" part ... theres yet much time needed I guess.
You don't need to add the "but first" when shouting out things like patreon, it scares viewers into thinking there's gonna be a long ad (Love the vids)
Framing the patreon + substack as ad reads makes it easier to transition into other plugs if he gets sponsors later on, and shows prospective sponsors where their ad would appear in videos. It's not really for you, it's for Jon and them.
There's a bigger reason these don't work out so well - after you've placed all these super sensitive structures and devices on a wafer....how do you dice the darn thing without applying chemicals or water that'll break or clog the device? This is where Stealth Dicing holds a big edge on other methods - without the need for chemicals like ablation lasers or plasma, nor the water jet and vibration of a blade saw, the NIR-Laser's use of creating internal cracks within the silicon has been an important step in helping that industry. I worked with such "lab on a chip" concept, but it had problems - the NIR can't function through metals, and since most traditional schematic to layout designers placed PCM's and other test structures along the dicing lane with the assumption they'd just use a blade saw, their miracle on a chip turned into a nightmare, and the most awkward five minute teams meeting meltdown I've ever seen, as their senior engineer just screamed **** repeatedly and started crying... There would be no "remasking" - this was supposed to be it, and they'd blown all their money on prototypes that hadn't foreseen the backend issues, and now they had the world's most sophisticated lab-on-a-chip, but no one in the world could properly(and cleanly) singulate the darn thing.
Innovators: 'Been there done that'; study & manage your surface tension of the fluid and the "Surface Energy" of the device and you'll do well, cheers !
I've wondered, with microfluidics and the way fluid channels can be used for controlled chemical reactions with fluidic logic gates, would it be possible to use the technology for recreational or "fun" purposes?
Might be expensive but spamming sensors and chips distributed to up the scale of required data sets to monitor the entire human body (or anyliving entity)? Having a larger compute with AI overstructure can organize these sub structures like a broodmother with a swarm of sensors/SoC's in a body?
15:00 there is no patient-oriented reason to minimize the blood you need to such a small amount. Once the patient has been stabbed then really you can spare a decent amount of blood. Repeatedly. It’s the stabbing that people don’t like.
If somebody can do even a 10th of what Theranos promised, the blood test market will expand way beyond its current $20 billion. Right now even basic blood work requires an expensive doctor's appointment. It's something most people do only when they are sick, or maybe once a year for a physical if they have good insurance. If it can be simplified to where each test can be done with unskilled labor for less than $10, a lot more people will be doing blood tests on a monthly basis. And it doesn't even have to be some creepy fully automated vending machine like Theranos was promising - a Walgreens lab tech will be just fine. Basically, if you can cut anybody making over $20 an hour out of the loop, it will reduce test cost 10x.
Apparently because the economics of chip fabrication - It's cheaper to produce ten million chips that does a hundred tests than it is to produce thousands to a million chips for each of a hundred different tests. The first chip of a new product is insanely expensive, but if you produce enough the price drops through the floor.
Which do you have in mind? lab-on-chip tech isn't the primary blocker for adoption of rapid (molecular) field diagnostic kits. lyophilization, primer/probe multiplexing, heat source miniaturization, and overall price competition with lateral flow ag strip style tests are.
Yeah! the cross section picture of the strip at 5:17 remind me of the home Covid test strip I used last year. I was amazed that it can confirm my fever was positive within 15 minutes. During the pandemic, I have to walk in serpentine line in a parking lot, got poked way up my nose then wait 3 days to get the result by email. This video make me aware that microfluidics technology has touched my life.
Covid or ELISA tests are established technologies similar to pregnancy tests. They are easy and cheap to make. Antigen testIng where tests require positive and negative are easy. Tests where you need a concentration (eg. Sodium level) is much more complicated
i-Stat only measure lactate, pH, bicarbonate, base excess, oxygen saturation, the partial pressures of oxygen and carbon dioxide and total carbon dioxide. Not so useful to determine the existing of specific bactery or viruses. I work in a company that currently develop a small device that can be used to detect those, and we are looking for clinical trial right now. Perhaps we will launch this device in next 2 or 3 years. What people think that not yet reachable for the moment, but the acceleration of R&D world makes this happen quicker than we thought.
TTL to CMOS.... This study is more like Hydrolic to Bio metrics. A Fluid has partical contaminatin. Electricity has Noise. Fluid activity response is compromised by relitive conditions. Overly complicated.
Capillary electrophoresis is fairly established in bio labs but I don't know if you could put that on a chip. The chip would have to be very long I guess.
It's nice to see you discuss this topic! My first internship had me working for a startup that had designed a Lab-on-a-Chip that had the purpose to bring Lithium measurements for bipolar patients to the point of care. Bipolar disorder is treated by using Lithium, but it has a narrow bandwidth within which it is effective. Too little, and it doesn't have an effect. Too much, and it can be poisonous. Patients treated with Lithium are therefore typically confined to staying close enough to their hospital to regularly have their Lithium values checked.
The LoC developed by the startup company I did the internship for brought this check to the patient. By just applying a drop of their blood and waiting a few minutes, the LoC would use the principle of electrophoresis to separate ions in the drop of blood and then determine the level of each type inside the blood, Lithium in particular. While the results were mostly for indicative purposes only, it did allow the patient a lot more freedom of movement. Only if the LoC results went 'out of bounds' could they consult their doctor to take immediate action, but the patient could be proactive about it and do so at home, or while on vacation (which the LoC allowed them do take!).
It may be a small application without a big market behind it, but I do believe that the quality-of-life improvement this particular LoC offered is more than worth the investment in the techniques behind it.
I'm a strong believer in the "KISS" principle: keep it simple, stupid. The problem in today's academia is that researcher's main task is to get published, and get published in the most prestigious journal, if at all possible. As a result, they are too often compelled to do "hero" experiments, that are extremely complex, and require multi-million dollar labs full of the latest equipment. When it comes to transferring that to a commercial product that is low cost and easy to mass produce, they are completely clueless. They are just not trained to think about those things. Successful commercial products generally adhere to the KISS principle. The glucose test does one thing, and one thing only. I'm currently working on a (very) low-cost, optical fiber sensing technology. The couple of papers I've published have attracted next to zero attention, and I'm not surprised, and yet I can get 100 times better resolution than all other papers in the field, at a 1000x smaller cost, but talking about cost in an academic paper is a big taboo. You can talk about "potentially" low cost, but an actual low cost, high performance device won't get you published in Nature Photonics, because there is no "big science" in it. Making things simpler is just not considered "advanced" science. This, despite the fact that technologies such as the IC started off as very simple devices. In fact, it is their simplicity, compared with bulk circuits, that made them so successful. Complexity was added later, but only because the process was, at its core, very simple. If the technology is complex and difficult to start with, it will never make it commercially because you can never get to mass production and high yields. So I think microfluidics is suffering from that disease. I've followed the field for 20 years, and considered using the technology, but it is actually very complex. I will stick to my little fiber optic sensors: easy to make, easy to mass produce, and extremely low cost due to all the cost reduction brought about by the massive deployment of optical fibers in the past 40 years. But I can tell you, convincing investors about it is THE most difficult thing in the world. They'd rather go with a fancy AI app. One more.
Sounds like you should hire Elizabeth to do your marketing and investor relations. I hear she’s not that busy these days.
I’d love a link to the papers!
this comment, a complaint about research impact and publication bias in a field that isn't even microfluidics, has an order of magnitude more likes than any other at present.
@@jeremycahill4662 right, just saying things people like to hear
@@lucas2nded461 I'm not even unsympathetic to the basic point -- I do some work on democratization and access projects for frugal biotech -- but the commenter's comparison to photonics seems very hollow. Applied life sci and physical sci are subject to substantially different pressures. There's actually pretty strong momentum at present for papers on cheap and open biotech. Even if it's only begrudgingly acknowledged by the academic publishing cabal.
check out companies like illumina, oxford nanopore, and 10x genomics. microfluidics is a huge component of genomic sequencing
Don't listen to people with cartoon girls avatars. They don't know what they are talking about.
@@interests3279 Well, in this particular case, they do. Microfluidics is essential to modern DNA sequencing.
I tried doing some electrophoresis nanofluidics during my PhD and one of the biggest challenges was bonding my top substrate to my bottom substrate. I really needed two glass substrates to allow the electric field to penetrate the liquid, because a silicon substrate would just cause strong capacitive coupling.
And bonding two glass substrates with a nanoscale cavity in between is really really really hard! First of all: both of your substrates have to be incredibly flat in order to allow cavities to form. Second: you need to deal with the wringing effect that also occurs in guage blocks. Third: you need to evacuate the air between your substrates during bonding, because you always need a thin substrate to allow you to use a 100x microscope and those substrates flex quite easily. Fourth: most literature on substrate bonding is a variation on anodic bonding, where an electric field assists in pulling the two substrates together, but it always requires some electrode nearby.
In the end, the only viable candidate I found was sodium silicate bonding, the same technique as the one that was used in the experiments of a Nobel laureate (the experiments in question weren't the ones he had won the Nobel prize for though). As luck would have it, my professor managed to get some students of this Nobel laureate professor to give us a Zoom seminar on their work. When I asked about details and recipes on the silicate bonding, professor X got his knickers in a twist about me trying to get valuable IP from his students (said recipes had already just been published in an addendum to a PhD thesis, his student just sent me a link to it afterwards). 😂 It was both hilarious and sad that a Nobel laureate from a top-5 ranking university was trying to hold back published information from a first year PhD student - from a university that's over 50 places lower on the Shanghai ranking list! - for fear of his own lab losing the monopoly on successfully pulling off a technique. That was the day I learned that the system of "first author gets all the credit" impedes the entire process of reproduction and building upon other people's work.
Other than the bonding, I'd say that a major impediment to microfluidics research is the fact that small scale production of chips in a research clean room facility is prohibitively expensive time wise. In order for a researcher to freely experiment with a device, they need to know that they can easily procure a new one if they contaminate or break the device they have on hand. As a PhD student, you have to design, produce, test, experiment with and clean each and every single device you make. And if you are to stand any chance of getting published in today's climate, you need positive results on a novel enough experiment. And if you want to meet your publishing quota, you need to do most of the work all by yourself. That's the real reason why micro- and nanofluidics are not taking off: it takes way too damn long to build enough devices to get to a workable experiment.
I also concur: lab-on-a-chip is not happening anytime soon. The natural development path really is to have more and more chip-in-a-lab devices, which might one day be integrated into a lab-on-a-chip. But you need to prove and improve all the individual components before you can assemble them into a VLSI-like system.
People here have been bonding microfluidic chips together comprising glass and composite substrates with nanoscale channels for years using HSQ ebeam resist. Many of these siloxanes are optically transparent and can be used to seamlessly bond most substrates together very well so its unfortunate to hear that they even think this is anything special in terms of IP or knowhow.
Much of the problems I see with scalability of research chips for experimentation is poor planning in my opinion and lack of sufficient computational resources or related to simulation students will not have the ability or foresight to design their simulations in such as way that allows for parametric iteration of key feature geometries when referencing rebuild models and constraint specific meshing densities. With access to a decent cluster and suitable software package you can run parametric optimizations of vast feature libraries so long as the key geometric aspects defining the parametric of the model are adequately parametrized and are referenced to a performance simulation output adequately. Even with a simple algorithm you can build something that varies the permutation set in the direction of optimized performance until a target is met.
When it comes to experimentation with real research models, good planning is the best approach. Knowing exactly what scope of feature geometry permutations you want to test, and generating a mask comprising those libraries of designs from the start will save tons of time when done at wafer scale. Often students are too short on time and energy and just want to jump into fab without much planning or mask design and end up taking far longer than necessary. I have seen projects that take years to complete and never go anywhere get completed in 6 months with proper planning and permutation evaluation strategy. Even highly divergent cases, you can often sample a broad permutation space structure library using primitive sub elements of a design that can be stitched together during lithography (assuming you have a basic R&D stepper or direct write tool) into a broad combinatorial permutation set to be evaluated. I have seen such methodologies implemented in a practical directed optimization research study in which a large parametric design space was optimized down to a small number of highly efficient prototypes with record breaking Dean Flow separation capabilities as just one example.
Its unfortunate how different the level of capability and knowhow can be in various labs across the planet. People hold back a lot but there are usually things that are going on light years ahead of what you might be aware of.
@@Exascale the IP I was talking about concerns recipes for doing sodium silicate bonding. Sodium silicate bonding is incredibly finicky, because sodium silicate isn't shelf stable, so it really saves you a lot of time if you can start from an exact recipe, rather than the laughably vague methods and materials section in a paper. The recipes were published as an appendix in a PhD thesis. It was so fucking petty.
Concerning the planning and methodology, in my case, there was just an incredible lack of supporting infrastructure. My supervisors had no clue about planning and they viewed it as superfluous. In terms of fab know how: we had no fab technician on staff, so all important know how had to be transferred from PhD student to PhD student and there was a lot of know how being spilled on every handover. Plus: we were using machines and bays from other research groups, so we'd always be last in line for timeslots.
Concerning the parametric design in simulation: absolutely, this will speed up your work a lot. Under the condition that you have your very first proof of concept run of the actual fabbing done.
@@LReBe7 Ahh yes the flatness issue with bonding. Yeah what a nightmare that can be if your starting wafers are not near perfectly flat. You can get away with about 10um or better TTV in a 150mm CZ-Si wafer depending on your cavity density and layout but for glass you have to be within 5um TTV for a reliable bond. Even then without 1um TTV you are still going to often lose some die.
Sounds like you had a nightmare of a graduate experience which is not uncommon. Often advisors get into a tenured position without understanding the full scope of what they are managing within their own labs. This can create a disconnect in expectations as well as respect for the needs required to get things done. Most labs are underfunded and most students and post docs are just trying to get through their project asap so they can graduate and move on. The deep thought and genuine curiosity has been replaced long ago with the anxiety and stress of staying afloat, graduating, or getting that coveted recommendation letter. There is so much corner cutting and embellished data. Countless SI's that are nowhere near reproducible. Countless images of the one perfect square um area on a 1sq.cm die of garbage. As everything gets more economically constrained its only getting worse. Even at top schools its getting very bad. Best of luck and hopefully you have made it to a place that is not full of fakers/bullshyterz. These days more than half the students I see coming out of universities dont know a thing about much yet have excellent GPA's and LinkedIn profiles that make them appear like the next up and coming nobel laureate. Society is crumbling as a result and every year that goes by there are fewer and fewer talented folks that can execute.
I worked in years for microfluidics (under dr Ramsey and Henly- check the lit they’re big guys) I was an undergrad at the time so no papers for me. I don’t really think this is more than a dead end frankly. I tried so hard to solve a few problems. There are some surprising issues that make this much harder than you think.
Though it’s fascinating. I’m sure some of the early IC guys lost faith too. So take my opinion with a grain of salt. I worked on 2 phase separation - the first separated on hydrophobic proteins vs drag then injected into another CE separation where it separated based on charge vs drag.
The big problem is the mixing problem and the non-orthogonal nature of those two methods.
And it is surely not “self powered”
We ran 10kV through it. In other words a crap ton. Small currents. But you require the voltage to create the electro-osmotic flow. The glass (borosilicate) becomes all negative charged because that’s what borosilicate does. Then the water being a dipole lines up the positive part to interact with the walls, all the positive outside water molecules run to ground . Because they are all the outside water in the “channel” they drag the rest of the water with mass flow.
So the channels have to be small.
We separated proteins from single cells. Normally if you do genetics on a biopsy you are analyzing the genetics of ALL the cells which are all doing different things.
Doing a single cell let’s you see what one cell is doing.
We also did DNA sequencing, by passing dna through a small pore and reading the resistance as it passes through one at a time. DNA has a negative charge so you can do this.
But clogging.
Dear god clogging . 95% of my time was Declogging my chips. It took days to make a batch of chips
By hand buy using old photolithography methods.
If one clogged, which happened all the f-Ing time it could take days and days and days to unclog.
Maybe it was dust. Maybe some of the salt came out of solution in your buffer solution, you would stare under a microscope for hours and hours and pass solvents through the channel to try and dissolve the clog and it never freaking worked
I’d be happy to have my “paper” (honors thesis) that explains things at a much lower lever since I started from fundamentals and worked
My way up. Mostly because I had a lot of pages to fill, wanted to show I understood the science, and the post doc assigned me a problem he couldn’t solve which was sorta a jerk move. So my “paper” had no “nove” findings and useless for anything other than the honors on
My degree. Anything I tried he’d condescendingly say I tried that. So I would scour research looking for ideas and I’m convinced the problem isn’t solvable on the purest sense. Smaller gate channels might help but then you run into lithography errors cutting such narrow channels and even more clogging. If only it didn’t take weeks to make more chips.
Ultrasonic for unglogging? What would happen?
A channel that opens it's full length to clean out.
is it possible to use a normal inkjet printhead (the resolution is already in the 10s of micrometers) and etching to produce micro fluidic devices? similar to how pcbs are made?
The biggest issue with microfluidics in the diagnostics (DX) space, excluding the limited cases where they are already used, is that they don't really do anything BETTER than current technologies. Most current DX systems already handle essentially all of the sample prep, sample control, and specific tets run. The operator just checks the tube of blood, scans the barcode, and inserts it into the system. The system handles the rest, right down to uploading the results.
Microfluidics emphasizes the concept of multiplexing tests, but there is little to no value in running unnecessary tests. An assay that runs 100 different tests when only two are needed is only worth those two tests; the other 98 are just wasted costs. Most modern DX systems will multiplex a handful (up to a dozen or so) of tests together which would logically be run together (e.g.multiplexing IgG, IgM, and Antigen tests for a particular infection). There are only so many tests that can be logically grouped together.
Lastly, for some DX tests, the usable concentration of analytes is so low that you can run into issues with minimum sample volume. This is especially true if a quantitative or semi-quantitative result is desired, rather than a simple binary result.
That reminds me of how most of the successful commercial development in the parallel computing space so far has been in single instruction multiple data (SIMD) processors, rather than multiple instruction single data (MISD). The former performs the same operation on a hundred samples simultaneously, while the latter performs a hundred different operations on the same sample. Most of the time, the former is more useful.
"Better" comes in a lot of ways.
The lab on chip is dirt cheap compared to big machines.
They also require no specially trained people to go around fixing and calibrating. No one going around cleaning and sterilizing.
Then there's the drastic reduction in electricity needed to run tests.
And let's keep in mind that if you have a box of 500 loc's, you can have 500 people being tested at once.
I think the Theranos event did set back the industry by a bit, because it did decrease capital inflows as the investors are more wary.
Not only that, but all the money they raised itself didn't go for startups that could have gone right. They took also time from that industry to move forward
Absolutely. A relative of mine works at an (actual) blood test company and they are very interested in solving this problem. But since Theranos, it's become incredibly hard to get funding for it, even if you try to solve it legitimately and have the track record to back up your claim of knowing how to develop and provide blood tests. Kinda sad because the technology has a lot of potential.
I'm not disagreeing with you, but I had never heard of Lab on a Chip - even while Theranos was advertising - until it blew up into a huge scandal that every news report covered.
So, while it did some damage in the short term, it may have actually helped the industry long-term by bringing awareness of the technology to the general public. When I started looking into Theranos to see what the story was about, it was pretty apparent in most coverage that I read, in spite of the fraud by Holmes, that the science behind the technology was in fact valid and being pursued by some of the world's premiere research institutions. Now I try to keep tabs on the industry, excited for the next real breakthrough.
I think the tech-knowledgeable investors are still going at this field. There have been quite a few impressive breakthroughs in specific areas of testing since this affair. Nothing as ridiculously grandiose as Holmes' BS. Theranos duped a lot of outsiders who's looking to make $$$. Notice she either failed or didn't go after health-technology focused investors.
@@JustJustSid If they had anyone that was truly educated on the science of the issue regarding blood testing and analyte detection then they would not be interested in solving this problem because it is virtually impossible for most screens of interest. The fact that there is interest in my opinion would be a red flag that they are not managed properly by scientific first principles.
The reason such technology does not exist has nothing to do with sensitivity. We have had single molecule sensitive assays for decades now. The issue is the statistics of the problem within the blood mixture, concentrations, and probabilities. For most analytes of interest you might have 1 molecule in a 10mL blood sample. You might draw 10mL and detect 1 molecule or detect 2 molecules of your analyte or detect zero with 3 sequential blood draws! So you see the pipedream of detecting any disease in a drop of blood is beyond folly. It is an absolute absurdity. No scientist worth his credibility should ever even be entertaining such research goals. If you see it you know they are either not scientists, poorly trained scientists that have no brain, total idiots, or charlatans.
The fact that Theranos got as far as it did is a testament to how much dumb money there is out there being thrown at people with letters after their name that came from anything remotely related to Stanford. What I have learned about people from stanford is that they are really good at marketing and bullshyting. They are professional bullshyters with no fundamental ability to do anything useful in reality. They will extract your money and charge huge salaries all while giving TED talks posted on LinkedIn on how smart they are.
There's another lab-on-a-chip concept that involves more biological instead of physical/chemical interaction. Growing tiny tissue clusters and then subjecting them to specific treatment (medicines, bacteria, chemicals, etc) produces an observable output that can help accelerate and unblock several new therapies that are either too expensive or risky to trial on human subjects.
ORGANOIDS!!!! This is what I do
@@kok2315 any interesting channels or resources i could check out? Thx😊
you mentioned flow cytometry briefly but it's a really interesting field, my mother worked as researcher for one of the most prominent experts in that field for over 20 years. flow cytometry is most useful because it can quickly and reliably count cells and other objects when they have a low concentration in a fluid that makes other methods impractical.
I use flow cytometry in my research all the time. It’s still a sorely misunderstood and under-utilised technique outside of medical research.
As a MEMS student, excellent video.
We're all students of memes.
We have that iStat device at my hospital. It is useful but I wouldn't call it "lab quality," it's mainly helpful for giving us an approximate answer very quickly e.g. when we have someone come in who might be in diabetic ketoacidosis and we want to do a blood chemistry. We still pretty much always send blood to the regular lab for definitive results.
I'm uncomfortable with manufacturer's claims of "Lab Quality". They muddy the water on these issue and it is interesting Abbot's i-Stat product webpage plaster a warning to me "BASED ON YOUR CURRENT LOCATION, THE CONTENT ON THIS PAGE MAY NOT BE RELEVANT FOR YOUR COUNTRY." Which probably meant Health Canada called their bullshit and restricted their approved labeling for medical device. I'm reasonably confident that their equipment can do a good job of accurate measurement ...but what is the quality of what they are measuring? I am not confident that a few drops of whole blood can be reliably reflect the actual intravenous status of patient. A pointless test given the sample limitation IMO, hope they put more money into speeding up and reducing the cost of actual lab testing.
Thanks!
Currently working on organ on chip model, and the chip in the lab is such a true thing, the amount of equipment you require to develop one, its better to optimize conventional methods but LoC holds great potential for diagnosis in the medical field. Great video by the way
Hey this is what I am working on at my job! We're a start up called Lightcast Discovery and I work there as a software developer, the potential is amazing.
Keep an eye on this field around early next year, we'll be making some very big announcements.
Just for some anecdotal historical perspective, the potential of LOC was amazing when I was a software developer in lab automation... 23 years ago 🙂 Best wishes for the success of your endeavours!
@@JonWallis123 Everything is impossible until it isn't anymore :) I wish I could say more but due to the NDA, I cannot, but what I will say is that what we are doing was absolutely not possible 23 years ago, which is what makes this so exciting. If the potential was amazing then, just think of what we can do with it now!
But we are building off the hard work of people like you, so thank you for the encouragement. I remain hopeful that this succeeds, not neccessarily for the companies profits but for all the good it will bring.
Lab-on-a-Chip was "The Next Big Thing" when I was working in laboratory automation 23 years ago.
deterministic lateral displacement reminds me of tesla valves, straight up magic
I prefer compliant mechanism one-way valves.
Damn. Where do you work? This content is so diverse, and so deep.
I've been aware of MEMs, photonics, and (non micro) fluidics, for decades. This is a great intro, to microfluidics, for a general audience. Like most things, its more evolutionary, than revolutionary. Hopefully, Theranos did derail things, too much. Still a better investment than crypto.
This is a great and succinct summary of the technology and industry! I studied this for years and it’s as good as many introduction chapters of phd thesis I’ve read!
The guy's really good, isn't he!
Lab on a chip already exists. A number of POC devices uses LOC technology but often it's fairly basic stuff.
A big issue with LOC is typically the price of the device. Commercial lab analysis is all about price and just reducing the amount of chemicals necessary isn't saving enough money.
Great video about an interesting topic, as usual :)
At first glance, deterministic lateral displacement can be easily explained by fluid dynamics, or just make an experiment yourself:
On a stormy day, try to shield yourself from the wind by standing behind an advertising pillar - or any large, perpendicular, cylindrical object at your disposal. You will find, that the wind speed behind that object is in fact larger than the wind speed blowing at its front.
Due to Bernoullis law, an increase in speed of a fluid results in a decrease in static pressure. An object subjected to different levels of pressure experiences a force towards the lower pressure zone (think of buyoyanci).
The magnitude of that force is determined by the size of the area affected, therefore large particles experience a stronger force than smaller ones (assuming spheric particles, this factors in with radius squared)
But the inertia of the particles counteracts this force. The inertia factors in with radius to the third power, therefore the resulting acceleration of smaller particles towards the zone of lower pressure is larger than that of larger particles, assuming their density being similar.
You know there's a crisis brewing in science, when a random RUclips comment can explain in simple terms what top scientists in high paying positions can't ;)
Without the relatively simple blood glucose test strip, I'd probably be dead now!
Well done review. I worked on the product development that used a few technologies you described in the video with some success. But "thanks" to Theranos it was impossible to get funding for further development.
Over all very accurate review.
An example of microfluids on chip is DNA sequencing on a chip: nanoporetech
Corning developed microfluidic devices in glass as early as the mid-1960s for use in pneumatic logic circuits. They sold them into the late 1980s if I'm not mistaken.
I remember this was going to be the next big thing about 30 years ago when I was just out of college.
7:55 See? Sharks with lasers on their heads aren’t the Bad Guys they are made out to be.
Holy crap, wouldn't expect my Ph.D topic (and startup job) covered so well by someone outside the field!
Hi. I have been very impressed with the quality of the videos that you have been producing. I am very keen to learn what are the questions that you pose to yourself when you research an entirely foreign and new topic? You would try to dig into these topics to what extent? I would be much obliged if you can make a video to discuss the methodology you adopt in researching a topic. Thank you very much.
Thanks, this is a good review of the field. As a researcher in this field, I can say that the technology to miniaturize regular tests is there but the big question is economics. This is especially important when i-Stat already offers a lot of these advantages. Instead of making lab-on-a-chip, microfluidics will evolve to process biological samples which are erstwhile hard to process or result in poor yield by conventional bench-top technology.
7:05 I like when she look at sample without the eyepiece!
We are growing organoid in the chips and the major breakthrough was when the 3D printers got sufficient resolution to be able to print the molds (200um features). Otherwise there is absolutely zero tasks for microfluidics which can’t be done on the multiwell plates.
Thank you for the service you are doing to the world with this channel and newsletter.
Before I started the video, I wanted to comment immediately about Elizabeth Holmes. 0:44 seconds in, and a picture of Holmes shows up. I cannot get Elizabeth Holmes out of my head whenever microfluidics is discussed.
Excellent video
Never miss a video. Great content from a neat perspective.
15:57 " The Chip on a Lab remains elusive" 🤔🤔 Hmmm, yes
Some of the principles you mentioned have entire books written on them that are very thick and very dry. Great job on zeroing in on the important stuff and bringing the whole story together in such a short video 👍
really liked the general overview of the whole topic, and while you touched on digital microfluidic biochips, it's a whole interesting concept in its own, it addresses a lot of problems in traditional microfluidic biochips, but also have its own set of shortcomings, a lot of good work is being done on it now.
What a rabbit hole. Who knew it would lead down to that mega fraud. Nice video and it kept me interested. Hope to hear more
Oxford Nanopore voltrax should be in this video. Way more advanced than the examples
I am in awe of your overall knowledge of all science and technology.
Me on the other hand, can say that tech and I are not the best of friends, however I enjoy when I can feel I comprehend sophisticated stuff so I keep my interest high.
That is why I have been very interested in "truth seeking for dummies" which at its core among multiple controversial conspiracy theories, tries to analyze things at the basic level applying mainly facts, experimentation and most of all common sense.
Based on that premises, in the past few years I still have never seen real photographs of the earth that can be genuinely proven taken from space. Further more the photographic evidence provided by nasa shows we have never landed on the moon nor it can prove the legitimacy of most of the space missions going on in the past 6 decades, and when you link the dots and follow the money and political trail, all indicates deception and shady institutions running the show of making people believe myths, stories and fake narratives.
What I am implying is that sometimes sophisticated science at all levels serves as a diversion or distraction to keep the people dumb and their behavior automatized and predictable so they can remain asleep.
What about you, other than being good at all science, how far are you willing to go to dispute or debunk the establishment if you were certain you have been lied to?
Would you even dispute common knowledge if you knew its false or if there was a great degree of reasonable doubt in the way the narrative has been structured or taught.
What seems obvious and undisputable because we are such an advanced and capable species therefore the premise is to never dispute the established science of centuries simply because we are standing on the shoulders of giants past and present, could be a wrong assumption.
Anyway, let me remind you that gravity is still a theory, the laws of thermodynamics are always broken to explain how our world functions and for some reason blue origin and virgin galactic can only fly sporadically to aprox 60 miles
altitude to claim they take few passengers after 25 years of research and development. By the way those supposed few individuals only claim to see darkness up there. Should we investigate or be satisfied with what doesn't make sense?
Well, before we continue to understand the complexities of our advanced civilization don't you think that some introspection is required to answer some really dumb questions, and to demand explanation for millions of people waking up around the world that understand that a lot is hidden from them.
Finally, if you know things but keep them hidden away for personal convenience or you haven't put enough effort to investigate things you know dong make sense, does that make you a fake. ?
If there even less than 1% of doubt then that is too high of a percentage to just go along with the flow.
Thanks Asianometry, great Video. As a scientist working in the field, I like to make you aware that there a many microfludic based test for many diseases these days. For HIV here is a list from WHO: extranet.who.int/pqweb/sites/default/files/documents/211220_prequalified_IVD_product_list.pdf . Nearly all modern automated devices for medical diagnostic are using internally microfludics. However the periphery is usually makro sized, so that you end up with a bench-top device an not small chip. However I totally agree today we have a "Chip-on-the-Lab" more than a "Lab-on-the-Chip".
You might want to check Cuehealth's device. It is cellphone sized device without much sample prep required.
Thank you Asianometry, for introducing this technological concept of microfluidics to me and perhaps some others seeing this well produced short video for the first time. This topic is new to me.
Nice video, I know the video have a limited time, but maybe discussing drop based micro fludics was worth noting as it has the potential of being conuntitative analysis for the samples.
Man, your tech reviews are mind-blowing!
An established market, but lots of potential here. Would be great to see more of these LoC's in the household to take away some of the "market" of medical professionals.
We have a NanoOne 3D printer to custom make chips in our lab, the integration to the rest of the processes is super simple since the new chips have pumps and sensors integrated and all connectors are printed instead of glued on !
Just like MEMS packaging it solves the integration for us.
Just a suggestion, but I'd LOVE to see a video on josephson junctions and the josephson voltage standard. I think something that's so obscure, but so incredibly fundamental that it defines something most of us work with on a daily basis warrants a video. =) But I'm sure you've already got a huge list of topics you want to cover.
Great show really informative and so well put together
Hello! Thank you for these videos. I find them to be very calming yet very intriguing. I often have problems staying focused on any type of content, but these videos are very interesting and keeps me glued. The topics you bring up are great and you always seem to be well versed yet down to earth, really makes me curious to learn more about the topics. Keep it up :)
TAM = Total Addressable Market
I love your use of the mountain streams joining in Taiwan's Toroko National Park to visualize laminar flow :D
Very nice field you explored similar to MEMS
The number of papers is far higher than 2000, that figure was year on year. It's in the tens of thousands. Otherwise nice video - I'm a biomedical engineer working in this field.
I am 40 and this is a new stream of science that I am falling in love with…Can I switch to this industry after studying….what can I do?
Microfluidics was used for high throughput RT-PCR testing for SARS-CoV-2/COVID-19. I think it worked well in this case because there was a need for large batches of a single test. Far from a lab on a chip, but there have been some practical applications.
@Asianometry i have a question and i would appreciate your answer. Do you research about the subjects you make videos(sepetially the ones on tech stuff) or is it part your field of study, and if it is the latter and if you dont mind whats ur field of study/research/work, because u seem well rounded in many topics(espetially microships, but also material sciences ,optics, pharma, tech, ai.. and even economics) but you also able to post frequently which i suppose leave you with little time to possibly research ur videos ?
Possibly, it would be worthwhile to have one chip to perform all of the tests that are now performed on different chips. A single fab is less expensive than several even when it fabricates a more complex chip.
To hell with that, I'm subscribing to this channel, at last.
And thank you for all these videos!
For classic organic synthesis this method gives high yields. Compare it to a bulk method...every molecule have to shake hands inside a big hall, but if you restrict them to a hall-way, it gets easier.. aka higher yields
It think that a lot of people including you look on this topic from the wrong perspective.
If there are chemical tasks to be done then the (main) functionality does not lay in the chip itself but of course the chemicals used and this is the area where research has to go into and not the chip itself that is just performing supportive and/or protecting tasks. This is the case because as you rightfully point out a lot of tasks in both detection and especially in preperation just cant be provided with a chip and need a lab anyway. The only way to work around this is developing analytical chemicals that perform these tasks since they are the only part here that can be scaled down with no differences/ loss in functionality.
But since investors, politicans etc. just see the word "chip" and ignore the "lab" part ...
theres yet much time needed I guess.
You don't need to add the "but first" when shouting out things like patreon, it scares viewers into thinking there's gonna be a long ad
(Love the vids)
Framing the patreon + substack as ad reads makes it easier to transition into other plugs if he gets sponsors later on, and shows prospective sponsors where their ad would appear in videos. It's not really for you, it's for Jon and them.
Was the blooper at the end deliberate?
Your videos should be shown in schools
Super great video, concise yet very in depth, easy to understand! Awesome job!
15:57 : "Chip on a lab"
There's a bigger reason these don't work out so well - after you've placed all these super sensitive structures and devices on a wafer....how do you dice the darn thing without applying chemicals or water that'll break or clog the device? This is where Stealth Dicing holds a big edge on other methods - without the need for chemicals like ablation lasers or plasma, nor the water jet and vibration of a blade saw, the NIR-Laser's use of creating internal cracks within the silicon has been an important step in helping that industry. I worked with such "lab on a chip" concept, but it had problems - the NIR can't function through metals, and since most traditional schematic to layout designers placed PCM's and other test structures along the dicing lane with the assumption they'd just use a blade saw, their miracle on a chip turned into a nightmare, and the most awkward five minute teams meeting meltdown I've ever seen, as their senior engineer just screamed **** repeatedly and started crying... There would be no "remasking" - this was supposed to be it, and they'd blown all their money on prototypes that hadn't foreseen the backend issues, and now they had the world's most sophisticated lab-on-a-chip, but no one in the world could properly(and cleanly) singulate the darn thing.
Microfluidics research in microgravity is very promising
Innovators: 'Been there done that'; study & manage your surface tension of the fluid and the "Surface Energy" of the device and you'll do well, cheers !
This is your 2nd video I've watched. I dont subscribe to many channels but your videos are very well done. Im in.
spectroscopy I think will eventually be the gold standard when it comes to diagnostics
Reminds me about the recent xkcd comic about the universe wasting effort on fluid dynamics.
I've wondered, with microfluidics and the way fluid channels can be used for controlled chemical reactions with fluidic logic gates, would it be possible to use the technology for recreational or "fun" purposes?
is there any way to take parasights out fast enough through blood tranfusion of yourself recieving your blood without parasights
Bruh finally a channel about real engineering
Might be expensive but spamming sensors and chips distributed to up the scale of required data sets to monitor the entire human body (or anyliving entity)? Having a larger compute with AI overstructure can organize these sub structures like a broodmother with a swarm of sensors/SoC's in a body?
15:00 there is no patient-oriented reason to minimize the blood you need to such a small amount. Once the patient has been stabbed then really you can spare a decent amount of blood. Repeatedly. It’s the stabbing that people don’t like.
If somebody can do even a 10th of what Theranos promised, the blood test market will expand way beyond its current $20 billion. Right now even basic blood work requires an expensive doctor's appointment. It's something most people do only when they are sick, or maybe once a year for a physical if they have good insurance.
If it can be simplified to where each test can be done with unskilled labor for less than $10, a lot more people will be doing blood tests on a monthly basis. And it doesn't even have to be some creepy fully automated vending machine like Theranos was promising - a Walgreens lab tech will be just fine. Basically, if you can cut anybody making over $20 an hour out of the loop, it will reduce test cost 10x.
Would love to hear your take on Archer Materials biochip tech.
The fortnight laser shark glider made up for the lack of one punch man reference after mentioning “sitama”try
Sharks with frickin laser beams!!! Thank you for the call back
Why are they shooting for a whole lab on a chip given that you need to be specialized for each field of medicine?
Apparently because the economics of chip fabrication - It's cheaper to produce ten million chips that does a hundred tests than it is to produce thousands to a million chips for each of a hundred different tests. The first chip of a new product is insanely expensive, but if you produce enough the price drops through the floor.
Sharks with frickin laser beams!
Wow dang I spied Ferro fluid on my hand is that bad?
Thank you for this video! It helped my research for a school project
No mention of the Covid test strip ?
Which do you have in mind? lab-on-chip tech isn't the primary blocker for adoption of rapid (molecular) field diagnostic kits. lyophilization, primer/probe multiplexing, heat source miniaturization, and overall price competition with lateral flow ag strip style tests are.
Yeah! the cross section picture of the strip at 5:17 remind me of the home Covid test strip I used last year. I was amazed that it can confirm my fever was positive within 15 minutes. During the pandemic, I have to walk in serpentine line in a parking lot, got poked way up my nose then wait 3 days to get the result by email. This video make me aware that microfluidics technology has touched my life.
Covid or ELISA tests are established technologies similar to pregnancy tests. They are easy and cheap to make. Antigen testIng where tests require positive and negative are easy. Tests where you need a concentration (eg. Sodium level) is much more complicated
Have you checked out Nuclera they have a chip on lab product that seems to be gaining traction with good results.
i-Stat only measure lactate, pH, bicarbonate, base excess, oxygen saturation, the partial pressures of oxygen and carbon dioxide and total carbon dioxide. Not so useful to determine the existing of specific bactery or viruses. I work in a company that currently develop a small device that can be used to detect those, and we are looking for clinical trial right now. Perhaps we will launch this device in next 2 or 3 years. What people think that not yet reachable for the moment, but the acceleration of R&D world makes this happen quicker than we thought.
new micro fluidics machine idea this is your goal
Idea: Tesla valve might mix fluids well
TTL to CMOS.... This study is more like Hydrolic to Bio metrics. A Fluid has partical contaminatin. Electricity has Noise. Fluid activity response is compromised by relitive conditions. Overly complicated.
Capillary electrophoresis is fairly established in bio labs but I don't know if you could put that on a chip. The chip would have to be very long I guess.
The intestines are long but in a smaller area
@@terrafirma9328 Yeah I think that would mess with the electric field though.
we really do need this tech for the better of humanity
Oh god he said Laminar Flow. Where is Destin from SmarterEveryDay 😂
Aren't DNA Sequencers another example of success?
Think we need a face reveal. The voice leaves a lot of unanswered questions.
hey jon i was wondering if you could do a video on the progress of chinese domestic CPU/GPU makers and compare them with Nvidia
Over 2000 papers? I can easily see 10,000 in that graph above (and maybe even more).
I used to work on microfluidics back to the graduate school.