[HN Gopher] Laser-based device can scan almost any sample of gas...
       ___________________________________________________________________
        
       Laser-based device can scan almost any sample of gas and tell you
       what's in it
        
       Author : PaulHoule
       Score  : 104 points
       Date   : 2025-03-07 16:06 UTC (6 hours ago)
        
 (HTM) web link (phys.org)
 (TXT) w3m dump (phys.org)
        
       | 01100011 wrote:
       | Now can we use this to do smog checks on cars as they drive so I
       | can avoid the biennial chore of taking the car in to a shop?
        
         | idontwantthis wrote:
         | Do they still actually analyze the gas? When I did mine they
         | just plugged into the computer and it was done in 10 seconds.
        
           | PaulHoule wrote:
           | The point of
           | 
           | https://en.wikipedia.org/wiki/On-board_diagnostics
           | 
           | is that the engine control unit in your car is monitoring the
           | operation of the engine and thinks it is in spec. If the
           | computer says it is OK, it is OK, at least in most states.
           | 
           | For years there has been talk about an OBD-III that would use
           | the cellular network or something so your car can narc on you
           | if it is out of spec
           | 
           | https://www.hotrod.com/features/obd-iii-the-proposed-
           | future-...
           | 
           | This has been stuck in purgatory with all the other proposals
           | for universal telematics not least because of privacy
           | concerns plus cell phone carriers being uninterested in
           | anything other than "what's convenient for us" coverage.
        
             | ziddoap wrote:
             | > _If the computer says it is OK, it is OK, at least in
             | most states._
             | 
             | The downside of this method (or upside, I guess, depending
             | on where you stand), over the sniffer method, is that this
             | is trivially bypassed. My race car, which is by no means
             | emissions compliant, can pass this check with the press of
             | a button.
        
               | Aurornis wrote:
               | States are beginning to roll out systems that do deeper
               | checks on the firmware. Those, too, can be bypassed
               | especially on older systems. However a lot of people are
               | getting caught at emissions time now with cars that have
               | been passing for years.
        
               | PaulHoule wrote:
               | Notably the functional test (measuring exhaust) can be
               | gamed too, see
               | 
               | https://en.wikipedia.org/wiki/Volkswagen_emissions_scanda
               | l
               | 
               | Emission control is a balance between emissions, fuel
               | economy and performance. You can give up some fuel
               | economy and or performance at test time and enjoy better
               | performance at the rest of the time if the system knows
               | it is being tested.
        
         | PaulHoule wrote:
         | I remember people from Sandia National Labs using LIDAR to
         | monitor pollution around Albuquerque circa 1990. Turns out this
         | been a thing for a long time:
         | 
         | https://ntrs.nasa.gov/citations/19720026861
         | 
         | and is still going on
         | 
         | https://iopscience.iop.org/article/10.1088/1748-9326/ab9cfd
         | 
         | https://idstch.com/security/lidar-technology-revolutionizing...
         | 
         | I think in those cases you know what the pollutant is that is
         | in a gas that you know what it is and want to make a
         | quantitative determination whereas the system described in that
         | paper is supposed to make sense of some random gas.
        
           | aerostable_slug wrote:
           | Sandia and other labs have indeed been working on this type
           | of MASINT collection for a long while. Their primary use case
           | is remote characterization of effluents to determine if a
           | given facility is being clandestinely used for things like
           | chemical weapons production or nuclear materials enrichment.
        
         | datadrivenangel wrote:
         | In Virginia they do this! They have devices that they'll put
         | near highway on/offramps and rotate them between locations
         | every few weeks or so.
        
         | BobaFloutist wrote:
         | No, I want to add headlight lumens/angle checks and window tint
         | checks to the smog check. It's an annoying chore but it's the
         | one time car owners have to (semi-regularly) get their car
         | looked at, we should take advantage of that.
        
       | dv_dt wrote:
       | It would be great to sample plastic and get composition
       | sufficient to determine a recycling category
        
         | bluGill wrote:
         | We already do that. However not all plastics and be determined
         | that way. Black plastic in particular often cannot be figured
         | out (since the laser as absorbed) and so even if the plastic is
         | easy to recycle it is often land filled.
         | 
         | There are open questions on if recycling plastic is worth
         | bothering with at all, but that is a different debate.
        
           | whatshisface wrote:
           | If you zapped it you'd have vapor to analyze.
        
             | aj7 wrote:
             | One time at the laser show, the booth across from us had a
             | methane analyzer.
             | 
             | You could tell when people farted.
        
               | throwup238 wrote:
               | Simple hand held methane detectors have been around for a
               | few decades.
               | 
               | This was used to great effect in the Kenny VS Spenny
               | episode, Who Can Blow the Biggest Fart:
               | https://youtu.be/ash2NzL1IHo
               | 
               | (It's a Canadian show)
        
           | colechristensen wrote:
           | I prefer the trash to electricity route.
           | 
           | If you think about it, the oil just took a detour into being
           | a water bottle or whatever for a little while on it's way to
           | the power plant. I really can't imagine we'll be at 0
           | hydrocarbon electricity grid any decade soon and burning
           | garbage (and instead of "recycling" plastic) seems to be a
           | perfectly reasonable thing to do if you can do it cleanly in
           | a power plant.
        
           | Animats wrote:
           | > We already do that
           | 
           | Here's the machinery doing it.[1] It's just using RGB and IR
           | cameras; they're not using lasers to vaporize samples and
           | analyze the spectrum. That would be both overkill and a fire
           | hazard.
           | 
           | Overview of modern waste handling.[2] Most of the separation
           | is entirely mechanical, with screens, vibrators, rotating
           | drums, and air blasts. But HDPE and PET plastics are
           | separated using infrared sensing.
           | 
           | [1] https://www.youtube.com/watch?v=51S3ET7M138
           | 
           | [2] https://www.youtube.com/watch?v=6rkHt34AztY
        
           | jfim wrote:
           | Oooh, so that's why the local recycling facility doesn't take
           | black plastics. Thanks for sharing!
        
           | dv_dt wrote:
           | Thats good to know. On recycling or not, i dont think it will
           | ever be a universal yes or no. But if the tech was there to
           | reliably identify & pull out a processable stream that has
           | concrete assessable value. Its probsbly a harder requirement
           | on some sensing tech to do it with continuously on an
           | industrialized stream of refuse.
        
         | TuringNYC wrote:
         | >> It would be great to sample plastic and get composition
         | sufficient to determine a recycling category
         | 
         | i love this idea. i cant believe how tiny the marked recycling
         | categories are on the containers, i can barely see them. i
         | always wondered how they even achieve this at the town
         | recycling center.
        
           | colechristensen wrote:
           | >how they even achieve this at the town recycling center
           | 
           | They don't. A huge proportion is never recycled at all.
        
           | klysm wrote:
           | Plastic generally isn't recycled that much, and many kinds of
           | plastic cannot sustainably be recycled.
           | 
           | There's a lot of automation in the space, and none of it uses
           | the markers
        
           | ianburrell wrote:
           | The marks on plastic aren't recycling categories, they
           | identify the plastic resin. They partially indicate
           | recyclability but the form is important. I don't think they
           | are used at recycling centers which mostly go on common
           | shapes. Products are supposed to now use solid triangle to
           | reduce the confusion with recycling.
        
         | klysm wrote:
         | We do this already, but also for metals which have a much more
         | reasonable recycling story. In metals though, we use LIBS.
         | Tomra makes the best machines currently
        
       | aantix wrote:
       | If you wanted to identify the components of a liquid, could you
       | heat it and analyze the steam?
        
         | Keyframe wrote:
         | if it can survive phase transition untouched. otherwise, if
         | there'd be a chemical change, you'd be looking at something
         | new.
        
           | zdragnar wrote:
           | I apparently have watched too much explosions & fire /
           | extractions & ire, because my immediate thought on seeing the
           | parent comment was that "yeah, it's gonna decompose before it
           | vaporizes, and if it's hypergolic with itself that'll be fun"
           | 
           | Apparently, watching an Aussie with a backyard shed full of
           | mysterious chemicals and reagents has me assuming that any
           | mysterious, unmarked liquid that needs to be identified is
           | probably explosive.
        
             | Keyframe wrote:
             | well, given the right conditions and enough heat...
        
           | rtkwe wrote:
           | AFAIU most methods of spectroscopy work by measuring the
           | atomic composition of the materials anyways not the exact
           | structure and bonds so a chemical change shouldn't matter
           | much.
        
             | Keyframe wrote:
             | that'd make sense
        
         | kurthr wrote:
         | Usually, they use HPLC (high pressure liquid chromatography),
         | which uses a solvent and packed diffusion columns.
        
         | WhyNotHugo wrote:
         | Different components might vaporise at temperatures.
         | 
         | If you heat water and alcohol to 80deg, measurements would
         | indicate ~100% alcohol.
         | 
         | You won't know the right temperature unless you know what
         | substance it is.
         | 
         | Edit: I guess you might also have stuff which reacts below
         | their vaporising temperature.
        
         | egberts1 wrote:
         | Every phase transistion (or phase change) of a molecule alters
         | its emission lines, at least as detected by gas chromatography
         | equipments.
        
         | aj7 wrote:
         | You could taste it.
         | 
         | But working on a good dog interface is the way to go.
        
         | DecentShoes wrote:
         | Man, the next Vessyl is going to be intense...
        
       | xeonmc wrote:
       | Similar concept to FTIR in terms of de-convolving a known
       | agitation pattern to enable more robust signal collection.
        
       | rz2k wrote:
       | I wonder how the mirrors are cleaned or kept clean. Maybe that
       | isn't an issue when detecting concentrations that would take a
       | mile of light to measure, and existing NDIR detectors are
       | sufficient for things like detecting different gasses in
       | pollution.
       | 
       | It would be nice if there were general purpose gas measurement
       | sensors that could identify more gasses rather than being
       | specialized to a specific gas. Could this be done by replacing
       | the circuitry of a typical evaluation board with an interface to
       | something like an NVidia Jetson Orin Nano?
        
         | condensedcrab wrote:
         | Usually cavity mirrors would be somewhat sealed to prevent
         | dust/junk build up.
         | 
         | Problem with general purpose spectroscopic measurements is that
         | you need a broadband source and detector. This adds complexity
         | compared to a targeted wavelength range.
        
         | jmusall wrote:
         | You usually mount mirrors vertically so dust does not settle on
         | them that much (also it is safer to keep laser light
         | propagating horizontally at only one height). Additionally,
         | overhead ventilation is often installed which provides a
         | vertical Laminar flow of dust-free air. If you have to clean
         | the mirrors eventually, you'd use compressed (clean) air or
         | special mirror cleaning wipes. Thorlabs (lab supply company)
         | has a detailed guide on optics cleaning[1]
         | 
         | [1]
         | https://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=90...
        
       | aj7 wrote:
       | People can't even tell how much cheap oil is in olive oil, let
       | alone "extra virgin olive oil" vs. el cheapo olive oil. Take it
       | from this old laser spectrocopist. When there's 75 mirror mounts
       | on a 4 foot by 8 foot table, with a Chinese graduate student
       | handcuffed to the table to adjust them, it's at least a decade
       | away. When you can buy a frequency comb laser on Aliexpress, now
       | we're talking.
        
         | j45 wrote:
         | Available on Aliexpress really is a solid indicator.
        
         | kylehotchkiss wrote:
         | Why did you handcuff a grad student to the table?
        
           | generalizations wrote:
           | Presumably they require constant high-precision adjustment
           | and a lot of domain knowledge. I'd read it that the only way
           | to do it now is with a high-maintenance setup and a lot of
           | technicians.
        
             | khazhoux wrote:
             | Yea, buy why handcuffs? Is it to limit range of motion when
             | taking precision measurements?
        
               | nartho wrote:
               | I believe this was a figure of speech to illustrate that
               | the student has to stay right by the table, not that the
               | student is actually being handcuffed
        
           | analog31 wrote:
           | It's like how Odysseus demanded to be tied to the mast of his
           | ship, so he wouldn't be tempted by the Sirens.
           | 
           | The grad student doesn't want to be tempted by the software
           | industry offering to quintuple his salary.
           | 
           | (Disclosure: Former physics grad student, who tended a
           | similar optics bench).
        
             | caycep wrote:
             | forget software, they all go into starting boutique
             | bakeries
        
           | contingencies wrote:
           | To stop them leaving academia for the private sector.
        
         | Eduard wrote:
         | searching for "frequency comb laser" on Aliexpress doesn't
         | yield fitting results currently for me.
         | 
         | https://de.aliexpress.com/w/wholesale-frequency-comb-laser.h...
        
           | gregschlom wrote:
           | yes, that's the point.
        
       | ck2 wrote:
       | Shame this won't work lightyears away
       | 
       | very hard to detect oxygen around exoplanets apparently but they
       | came up with a new method recently
       | 
       | https://news.ucr.edu/articles/2020/01/06/scientists-develop-...
        
       | zerealshadowban wrote:
       | It seems that the frequency comb laser is easy to diy! "NIST
       | Shows How to Make a Compact Frequency Comb in Minutes" (2013)
       | [0]. I can think of many applications apart from the medical
       | diagnosis they talk about.
       | 
       | The varying mirror cavity is today's innovative step, plus the
       | computation to identify atoms and molecules from the observation
       | of color absorption.
       | 
       | [0] https://www.nist.gov/news-events/news/2013/07/nist-shows-
       | how...
        
         | aj7 wrote:
         | Yes. And a surgeon can remove your appendix in minutes too. A
         | 777 can be landed in a crosswind, in minutes. Why a pass can be
         | intercepted and run back for a touchdown in seconds! All of
         | this easy. I've seen it on TV, even.
        
       | pjs_ wrote:
       | It has been possible to buy pretty nice handheld Raman
       | spectrometers for a while now:
       | 
       | https://www.bruker.com/en/products-and-solutions/infrared-an...
       | 
       | This paper is obviously trying to go much further. But if you
       | just want to buy a tricorder, you can
        
         | multimoon wrote:
         | I was unable to find a specific price for the model you linked
         | in my few seconds of looking but these devices generally run in
         | the $10,000 USD range, so I don't think it's achieved the "if
         | you just want to buy it" territory yet.
        
           | pjs_ wrote:
           | Good point. You have to ask yourself -- would you rather have
           | a '97 Camry or a raman spectrometer and a pushbike ;)
           | 
           | That said this is getting into Christmas present territory:
           | 
           | https://www.ebay.com/itm/167319991806?_skw=handheld+raman+sp.
           | ..
        
             | throwup238 wrote:
             | I love the juxtaposition of thinking a '97 Camry is
             | anywhere close to $10k while at the same time a $3k piece
             | of lab equipment is "Christmas present territory" :)
             | 
             | It's one banana, Michael. What could it cost? Ten dollars?
        
               | Loughla wrote:
               | It is genuinely disheartening when people in this board
               | start talking about things they buy when they're WILDLY
               | out of my price range. It's almost like living in a
               | flyover state and working in education isn't optimal for
               | income. Weird.
        
               | pjs_ wrote:
               | Dude I have three thousand dollars net total in my bank
               | account. I don't own a car or a house, don't have kids
               | etc. I literally used to co-own 97 camry with bees in the
               | trunk. I paid about $2500 for it. Hence why I say it
               | costs the about the same as the spectrometer on eBay.
               | $2500 is about $3000. However you and I both know that
               | there are people all over this board who buy their kids
               | macbooks for christmas
        
               | LoganDark wrote:
               | > there are people all over this board who buy their kids
               | macbooks for christmas
               | 
               | Kids? I bought a MacBook Air for an internet friend of
               | around 2 months because I was so offended by the fact
               | that they didn't own a good computer.
               | 
               | My performance standards are higher for myself of course,
               | but the goal was just to get them any form of macOS.
               | 
               | (They did in fact end up liking it a lot better than
               | Windows)
        
       | reaperducer wrote:
       | _Laser-based device can scan almost any sample of gas_
       | 
       | Oh, sure, it can tell you who smelt it. But the real money is in
       | figuring out who dealt it.
        
       | Vox_Leone wrote:
       | Why this progress is significant for AGI:
       | 
       | Multi-Sensory Integration: Traditionally, AI models have been
       | trained on a limited set of sensory data, mainly acoustic
       | (speech, sound) and visual (images, video). Adding chemical
       | (olfactory) and tactile data allows AI to engage in multi-sensory
       | integration--mirroring human perception much more closely. This
       | can help machines interpret the world in a more holistic way, as
       | humans do, where smells and touch inform emotional responses,
       | memory, and decision-making.
       | 
       | Challenges:
       | 
       | Data Collection and Tokenization: Tokenizing chemical signals and
       | tactile sensations in a way that AI models can understand is no
       | small feat. With smells, for example, creating a "universal"
       | vocabulary or set of descriptors that an AI can use might require
       | advanced pattern recognition and mapping of these signals to
       | emotions, objects, or experiences.
       | 
       | Complexity: The integration of multiple senses is complex, and it
       | might take time to fully understand how chemical and tactile data
       | interacts with visual and auditory data to form a coherent model
       | of the world.
       | 
       | + Ethics and Privacy
        
       | mannyv wrote:
       | Contamination?
        
       | akomtu wrote:
       | Can it detect the amount of PFAS in tap water?
        
       | Beijinger wrote:
       | Sounds trivial. Without opening the link, let me guess: QCL?
        
       | amelius wrote:
       | Can it detect medical disorders, similar to how (according to
       | popular belief) dogs can smell cancer?
        
       ___________________________________________________________________
       (page generated 2025-03-07 23:01 UTC)