ไม่สามารถเล่นวิดีโอนี้
ขออภัยในความไม่สะดวก
Silver nanoprisms grown into structural colors by high power LEDs
ฝัง
- เผยแพร่เมื่อ 8 เม.ย. 2022
- How to chemically synthesize silver nanoparticles, then grow them into triangular nanoprisms with light from a variety of LEDs. Each color LED creates a different size nanoprism, which has its own characteristic color.
CMDITR video: • Silver Nanoprisms Synt...
Multispectral LED driver on Github: github.com/ben...
Chemicals sourced from Amazon/eBay
LR-1 spectrometer: www.aseq-instr...
Micropipette set: www.amazon.com... www.amazon.com...
20ml glass vials with PTFE lined cap (do not use metal-lined): Environmental Express APC1675P Already gone from Amazon
pH pen (this cost more than I remember, but it works really well, and has lasted many years. Cheaper pH pens are often pretty bad) www.amazon.com...
Comparison of CD, DVD, Blu-ray discs with electron microscope:
/ 615327472909840385
Great way to find related papers:
www.connectedp...
Research sources:
• Silver Nanoprisms Synt...
sci-hub.se/htt...
sci-hub.se/10....
sci-hub.se/10....
sci-hub.se/10....
sci-hub.se/10....
opg.optica.org...
www.rsc.org/su...
sci-hub.se/10....
sci-hub.se/10....
www.cytodiagno...
Support Applied Science on Patreon: / appliedscience
![[UNCUT] ชาวบ้านถาม “นายกอุ๊งอิ๊งค์” ประชาชนมีกิน มีใช้ กี่โมง? I คนดังนั่งเคลียร์ I 16 ส.ค. 67](http://i.ytimg.com/vi/tKZvOzLIJJc/mqdefault.jpg)
the projects you come up with never cease to amaze me. great work!
You might have missed it but he spelt “subscribe” in nano silver particles.
MUM! It's that time traveler again!
I follow you both.... Very interest stuff and approach.
A collaboration between you two would be epic.
No way! That's too cool, man.
Fun fact, this triangular gran of the silver particles is why analog film companies call some of their stock, TRI-grain or Delta. Extar has a tri grain and illford Delta has this shape. it makes for a higher definition exposure.
I wanted to say this too.
And Kodak T-MAX films were made by using only tablate triangular prisms oriented perpendicular to the film plane This gave them maximum sensitivity and minimum grain size. Wonder how they made those crystals!
TIL
Fascinating! Throughout the video I was thinking about applying these to a paper or polymer substrate ie film. And I thought, hmmm I bet industry has checked this out.... thanks for the factoid!!
That was fun. Thanks.
A potentially fun follow-up experiment: Rig up a jig that lets you both illuminate the liquid with a specific wavelength OR sample it with the spectrometer in situ. This way you can turn off the illumination long enough to sample the absorption spectrum at regular intervals. It would be cool to watch the peaks shifting on the absorption spectrum in time-lapse.
The reference video did that over 26hr, but with multiple solution runs
Oh no. I have a centrifuge in shipping right now that I bought for dewatering nano particles. I did not realize until seeing that separation chart how much of a challenge is ahead for me and my $100 ebay centrifuge.
Really excellent project/video as always!
What are you working on?
I hope your particles aren't too small. In the past when I was making 4-8nm diameter water soluble nanoparticles, I used to have to use an ultracentrifuge to pellet them.
well if it's balanced and built well enough you can always try to supe it up a bit lol.
@@PedroDaGr8 I think I'll be ok. I'm looking for a range of particles from about 100-1000nm
You can get a refurbed air-driven ultracentrifuge for $100-200, which should be capable of 199,000 x g.
Ben, this is super cool, every time I see an upload from you I know I'm going to not only learn something new and cool but then end up down a rabbit hole for the next few weeks, lol so thank you :)
It also comes with a convenient stick that we can use to measure our own lack of accomplishment along the way.
I feel this same way. If Ben puts out a video it's going to be fascinating regardless of the topic.
9999999999999999999999999999999999999
We need Ben clones to have videos much often :D
I'm definitely saving this for later. In a few minutes the kids go to bed and roughly 45 minutes later I shall kick back in my recliner with this on the TV. Glorious Saturday Night Commence!
Nerding out with an Applied Science video after the kids are in bed. That sounds like paradise to me, no sarcasm at all. I love these videos!
I saw the notification of the upload when I was making dinner.. trying to hurry up so I can go relax and watch! 🙌🙌🙌🙌
If the reactivity of the particles is "encouraged" by a given wavelength of light, it makes sense to me that they'd therefore react and grow *until* they're no longer (strongly) absorbing that wavelength. If that's the reason, that also explains why irradiating with a lower wavelength of light doesn't shrink the particles. TL;DR: if we assume the reaction occurring here is strictly crystal growth and that reaction is *enhanced* by light absorption, then we expect both effects that you seemed slightly surprised by (absorption wavelength != exposure wavelength; and the inability to reduce the absorption wavelength.)
I agree. But have you found any resources that show size dependent absorption properties of silver nanoprisms? I found a calculator for mie theory of spherical particles which supports this theory, but it would be nice to check for prisms.
I also noticed that in the third graph, the 405 peak has red-shifted somewhat. I think that aligns with what you're saying. (I'm also curious about the new large peak that appeared in the IR range after the first graph.)
this explain why when he use the 450nm LED, with intention to "revert it back to original colour" the red-shifting/peak-shifting actually do the opposite and is much stronger because more crystal from 400nm can grow bigger when provided with 450nm light.
This is the same hypothesis I had. Also, for the colour to shift back, it may be because the larger particles just fell out of solution while un touched smaller particles grew larger.
It is just an illusion of shifting back, the particles don't actually become smaller.
Ben, one of the ways that nanoparticles are prepared for TEM is to place a droplet onto your substrate, wait for some time for the particles to settle/adhere to the substrate, blot the supernatent out with filter paper, and optionally wash by dropping wash solvent and blotting. There are a lot of surface modification tricks to try to get more particles to stick, such as plasma cleaning, UV-ozone, spin coating charged polymer solutions, etc. It might take a few iterations of sample loading (e.g. apply multiple droplets) to get enough visible sample. Definitely not easy and unfortunately not well described in literature (generally passed down via institutional knowledge in labs studying nanoparticles).
TEM is definitely the way to look at nanoparticles. Time to build a TEM Ben!
I love your old JEOL JSM T-200. I worked for JEOL for about 10 years in the early 90's. I had several of that model in my service area, along with the later 300 series models. They are a great little scope when working properly. The 10nm resolution is only with Gold on carbon @ maximum KV very short working distance. The sample must be heated to ~ 100C before being transferred to the vacuum. Also must have very good chamber vacuum. low e-6 torr minimum. It is a pain, and I can remember the struggles when resolution had to be demonstrated.
One sidenote concerning the stability of the NaBH4-solution: You could basify it in advance (to like pH 10-11), which would increase it's lifespan by orders of magnitude. Just seems like a possibility, since the final reaction mixture needs to be basic anyway. So a mixed stock solution of NaOH and NaBH4 could be used instead of two separate ones, since the required amount of NaOH is also known (2mL 50mM NaOH stock solution)
Dang, I never knew how much I wanted a multispectrum monochromatic LED light, that is a super cool idea! I've got 9 different wavelengths of laser pointers which are also very cool for studying light absorbtion in different materials but now that I see this video I need the LED equivalent. It's amazing how much LED technology has evolved in just the last decade, being able to get this many monochromatic LEDs with such a high output power for the relatively low price they cost is incredible.
I also think a flashlight version would be pretty great too, with a wheel inside that rolls through all the different wavelengths so only one is on at a time, that would be neat...
A few years ago I bought a bunch of different wavelength LEDs to experiment with what things look like under different "equivalent" lighting conditions. Then I never actually wired them up. They had quite a variety of inconvenient forms, haha. Maybe I'll just make his. XD
Yyyeah, but compared to your lasers even the most monochromatic LED's are still pretty broadband. It's still a cool idea though.
Man I only have six wavelengths of lasers, unless we're counting frequency doubling as two.
Get a nice prism optical glass or diffraction grating
I wonder how hard it would be to get a dense enough series of LEDs so that by adding temperature control (IIRC peek wavelength shifts with temperature) you can get any peek you want?
I wish I had a friend like Ben IRL to just unashamedly totally nerd out with over spectroscopy, electron microscopes, cryogenics, lasers, x-rays, and superconductors.... 😿
RIGHT?! hahaha.
LoL who needs friends when you could tell GPT or another language model to be wtf you want it to be
When I needed to image citrate-terminated gold nanoparticles in SEM, I used to deposit them onto poly-electrolyte multilayers. Surprisingly simple to set up. Simply prepare your imaging surface (the CD ROM in your case) by alternatively dipping it in solutions of polystyrene sulfonate (PSS; a negatively charged polyelectrolyte) washing with water, then dipping into poly diallyl dimethyl ammonium chloride (PDAC; positively charged), and washing in water again. Repeat the process 3-5 times to ensure a good coating of the poly electrolytes. Make sure you finish the treatment with PDAC so the surface will tend to be positively charged, so that the negatively charged nanoparticles will adhere to the surface. Simply leave this surface in the nanoparticle solution for an hour or two and you should have a even monolayer of individual particles to look at.
10:08 I have been wanting to build one of these to test out astrophotgraphy light pollution filters. Mouser sells some very specific LED wavelengths, such as 656.28 for Hydrogen Alpha.
Something to filter out Musk's Starlink satellite's would be great...
For everyone else, my pic's would still be terrible! lol
The satellites only reflect sunlight, not really something to filter out optically
@@among-us-99999 I'll just filter out the all the wavelengths from sunlight then!.. ;)
@@timhooper1557 "Since the Dawn of Time, Man has yearned to block out the sun."
Two ideas for reversing transition:
1) Add precursor silver particles over time to encourage stealing atoms from plates.
2) Try going one wavelength step at a time, preferably try it with a smallest wavelength step you can get.
I was thinking that too. Maybe doing 2) in the first place would lead to a better conversion from the start as well. From the electron microscopy it seems there should be much room for improvement potentially.
The red-shift is due to inherent energy losses in the system (basically non-reactive vibrational modes). If you use flourescent nanoparticle systems as a model, I think it can help make sense of the situation. Basically, not all of the energy you put into the system is available for the output, be it fluorescence or building the particle. This is why the shift is always in the red (aka lower energy) direction comparted to the exciting wavelength.
This is probably the largest factor but it's also possible the other components are effecting the wavelength (glass, water etc)
Great video! I worked and made AgNP for my SERS research in grad school at UCSB. I made "bare" borate capped nanoparticles with just borohydride, but they were very tricky to get right and unstable once any contamination got into them. You can also use citrate directly as a reducing agent to make "nanoegg" ovaloids by heating a AgNO3 and citrate solution. These will provide stable citrate capped AgNP. Also, majority of absorbed light usually goes into creating plasmons (surface plasmon resonance). The electrons excited by this process can go on to participate in reactions, like reducing Ag ions in solution onto their surface. Might be an explanation on why the AgNP grow irreversibly with intense light.
I think photoelectric effect is interacting with plasmon resonances. Short wavelength (high energy) light causes electrons to be ejected from the particle surfaces and then the solvent captures a silver ion to neutralize the particle. This continues until the particle gets small enough that its plasmon resonance frequency equals the light's frequency; at this point, the plasmon resonance shields the particle's surfaces from the light. Once the particle is too small, no longer wavelength light can reach it to cause dissolution. You could check this model by adding some fresh solution to see whether the color could be reddened in that case. Adding fresh silver ions will allow all particles to grow until the photoelectric effect arrests the growth. Plasmon resonances will have normal modes in triangular prisms. Probably, the underlying fcc structure of silver plays a part in the triangular shape.
That could be potentially confirmed by looking at silver’s photoelectric work function to see if the redshift corresponds to that difference in energy. I think.
@@Scrogan Bulk properties don't usually apply to nanoparticles.
This would be an argument for shrinking the silver nanoprisms, but against growing them, right? "This continues until the particle gets small enough" would push particles to smaller sizes, but never to larger ones, which we clearly see in the initial illumination.
Please correct me if I misunderstood your comment.
*"Tunable Dipole Surface Plasmon Resonances of Silver Nanoparticles by Cladding Dielectric Layers"* -Xiaotong Liu, Dabing Li- Nature scientific Reports 2015, "[...] In this article, tunable dipole surface plasmon resonances of Ag nanoparticles (NPs) are realized by modification of the SiO2 dielectric layer thicknesses. SiO2 layers both beneath and over the Ag NPs affected the resonance wavelengths of local surface plasmons (LSPs). By adjusting the SiO2 thickness beneath the Ag NPs from 5 nm to 20 nm, the dipole surface plasmon resonances shifted from 470 nm to 410 nm. [...]"
Small particles interact significantly less with long-wavelength light and grow until the dissolved ions are depleted. Large particles emit photoelectrons and ions and dissolve making them interact less. Equilibrium occurs when the light's frequency equals the particles' plasmon frequency.
I love this channel. Every video is well researched and really well presented. The downside is we get a video every 2 months. 100% quality over quantity.
>every 2 months
i think it takes longer to produce these
What always fascinates me the most is how people somehow managed to figure out this works in the first place. Thanks for sharing.
Insanely obscure topics demystified and demonstrated is why I love this channel so much!
Yes, every video shines a light in an undiscovered corner.
If particles are catalyzed to grow when they absorb light, it makes sense that they keep growing until they are all too large to absorb the LED light. This would make the particle absorption peak longer wavelength than the LED illumination bandwidth.
The brief look I made into gold nanoparticles suggested they were much simpler to make, I think they used citric acid as both a reducing agent and an anti-clumping agent.
Hey Ben, you should check out this paper: Bensley, Robert D. “Natural Color Photography in Colloidal Silver.”. As the name suggests, it involves creating color photographs out of colloidal silver using a standard black and white emulsion. Very interesting stuff.
They fail to mention it in the article, but Bequerel, Herschel and others produced solar spectra in full color on metal plates coated with silver chloride. Some of the surviving examples are quite striking.
Many color films (in the past, maybe currently) use a layer of colloidal silver between the blue-sensitive layer and the red / green sensitive layer beneath as a filter. Kodachrome, in particular, was well known for this
@@johnkukla9522 The color images produced by Becquerel and others using the early processes are fascinating, but I believe that it operated more on the principles of Lippmann plates than colloidal effects. And you are correct about the colloidal silver in color film, but it merely acted as a yellow filter to block blue light from reaching the red and green (or complements if negative) emulsion layers.
@@Dukey8668 CNRS (the French National Centre for Scientific Research) did a study on these several years ago, and concluded that the colors were due to the presence of silver nanoparticles in the images. There's a press release available, but I have not found an accessible copy of the research paper. I'm not sure if Lippman - style interference phenomena would be possible with silver-chloride based materials that do not have (conventional) spectral sensitivity beyond the UV / Blue. As an aside, have you ever had the opportunity to see a Lippman photograph? I've never had the pleasure of seeing an original, but I've seen some modern work, and they are absolutely gorgeous beyond words. The emulsions are fairly straightforward to make, and this has been on my short list of projects to do for quite some time.
Awesome Project! About the red shift: Did you consider if the refraction index of the medium might have an effect? Or even the shape of the cylindrical container?
It would seem the crystals prefer growing over shrinking, like at 14:15. So if the crystals can vary in size, this could give the process a bias towards larger than optimal crystals.
Thinking of the reversibility, maybe it would work if you tried in small, few nm steps. The full cystals might be too far away from the wanted wavelength for it to have serious effect.
You are somewhat correct! The index of refraction directly influences the wavelength* and therefore directly influences how materials react to light. However, since we never look at the optical properties of the silver particles outside of solution this doesn't matter. The wavelength is shifted when illuminating, when measuring the extinction, and when looking at the colors of the samples.
The shape will likely focus the light a bit, just like a lens would. This may make the growth faster, but will not change the mechanism.
*It's the wavelength and not the frequency, because E=hv (Energy is planck constant times frequency) and the energy is conserved
It seems awfully similar to what goes on while using fluorophores in fluorescent microscopy. All fluorophores emit light at a longer wavelength than the wavelength they absorb.
Basically, instead of losing energy as a function of brightness, a fluorophore (in theory) could emit as bright of light as the one used to illuminate it, but energy is still conserved of course and the fluorophore emits as bright of a light but at lower frequency/longer wavelength.
Since the wavelength of illumination is depositing the amount of energy required for the particles to assemble to a given prism size, and no system is lossless, the prisms will assemble to absorb wavelengths longer than the wavelength required to stimulate self-assembly.
Perhaps like fluorophores, they conserve energy in a similar fashion - i.e. A solution of a given prism size may absorb as much light (in amplitude) as the brightness of the light used to create them but at a longer wavelength, thus still conserving energy.
@@puskajussi37 "crystals prefer growing over shrinking" This is in general true of all reactions, it takes less energy to make a bond than to break a bond. At least the ones that have a favorable forward direction. Additionally, in the case of nanoparticles, there is no guarantee that the "breaking" path follows the same path as the "creation" path. Often for optically active materials, irreversible oxidation pathways are a problem as they provide an alternate terminal pathway the reaction can take.
I would think that it would be the room temperature which is basically ir light. Would shift it in the red spectrum.
This is really cool. When I was in school we made something like this but with a rather more mundane size selection... it was CdSe nanocrystals (at the time we were still calling them microdots) and they were grown via Ostwald ripening... basically start with a mixture of Cd and Se mercaptan salts suspended in molten parafin, then set to a specific temperature for several hours and depending on the temperature, that would determine the size/color. I'm sure I've forgotten some critical details... it was almost 20 years ago. One thing I certainly won't forget though... the odor of bezenethiol (or thiolphenol for those who swing that way). It was like an unholy blend of smoke and anus.
I remember seeing a poster with that in the chem lab at my school -- a student had done the series showing the rainbow spectrum of fluorescence. Don't remember any of the chemical details though... just the pretty vials on the poster :)
Now there is an experiment you won´t see in any school anymore.
I've done hundreds (if not thousands) of CdSe nanoparticle reactions in the past. We could easily hit emissions at pretty much every 5nm from around 500nm-700nm with FWHM in the range of 15-25nm. Pretty much nobody uses Ostwald Ripening anymore due to its inherent issues (namely population broadening, surface defects, etc.). Xiaogang Peng had an excellent article from 2001 (J. Am. Chem. Soc. 2001, 123, 1, 183-184) which really changed the synthetic science in the field and others have built on that. The understanding about exciton confinement, non-radiative pathways, etc. have resulted in some amazing particle structures (CdSe Core/Intermediate Layer/ZnS Shell being the most common structure with various coating/encapsulation techniques).
Aww c'mon. You're blowing smoke up my 😂
I love watching your content when chronic pain wakes me up in the middle of the night and it takes 2 hours for a nerve blocker to do it's thing. Keeps me very focused on what your doing and talking about so I don't feel as much pain. Thanks again. Your awesome! 😎
2:15 I love the branding on that Sodium Citrate. "It's Just"
Always pumped to see you post. Hope you are doing well!
I always get excited when the electron microscope gets screen time! I think projects on this scale are some of the most interesting.
Ben, you focus on the coolest of engineering + physics
Neat! Sounds a lot like the chemistry of film photography!
You are one smort cookie. Truly one of a kind among the TH-cam STEM community. Thank you for the content
This guy made the very best video I have ever seen when he made a very low temperature beaker go below 100 degrees below zero in a two stage cooling beaker that makes the atmosphere separate into liquid components by repairing a unit that can even make dry ice.
"I'd like to show you these silver nanoprisms"
Truly a man of my own heart
It's great to see people so curious that they are willing to go through so much effort, basically just for fun.
I'd love to see a multispectral photographic imaging project using that light.
Put an ancient papyri document (or whatever document you have handy if you don't have any ancient papyri lying about) flat on a vacuum plate with a monochrome astrophotography camera above it with the light at low angle. Image the document at each wavelength. Rotate the light 15deg around the center of the document, image again, and repeat through a full rotation. Then rotate the angle of the light up 15deg and repeat the entire process again. Ideally the light should be collimated (maybe a bigish fresnel?), and maybe place some filters over the LEDs to narrow the frequency range.
Write a bit of software to display one of the ~2100 images from the set based on three inputs that select the lighting angles and frequency.
The idea is to reveal hidden details in the document.
I have never clicked on a notification quicker than this one.
I haven't either! Apparently I was the 1st comment, too. Let's me tell the people that just say, "First," that they're not.
Science aside for a moment, your opening clip was framed into an incredibly inviting scene. Seriously, let a few weeks go by and just pause at 0:00 to admire the artistic thought that went into the shot.
As a layperson with no education in science I watch pretty much all of your videos until the end. You're very good at explaining things and I'm always amazed at the breadth of your skills and knowledge and the amount of work you put into your projects.
I didn't really have time to watch this, but thought I will just watch some introductory part in the beginning. But I couldn't stop, watched the whole thing! Very interesting stuff!
Amazing video. I have a few comments and ideas:
1. You're talking about the *absorbance* of your sample at the end, but you are measuring the *extinction.* There is a big difference, because these particles *scatter* very strongly (as seen by the nice color range shown in the thumbnail), and what is transmitted is the source minus scattering and minus absorbance.
2. You could maybe filter the nanoparticles with a reverse osmosis filter. Just an idea, but you already got nice pictures anyway!
3. My theory for the growth mechanism is the following:
If you irradiate a sample of silver nanoparticles they will absorb some of the light. However, this absorption may depend on the size of the nanoparticles. If small particles absorb more light than big ones they get fed energy and are more unstable, leading to dissolution and agglomeration into bigger particles. This would explain why you can't shrink particles by illumination - the big particles don't absorb as much energy and therefore have no reason to dissolve into smaller particles.
This explanation does not offer a compelling reason for the spectroscopy results at 25:40, but I would argue that the spectrum just shows the emergence of smaller particles from remaining seed material, rather than the dissolution of the bigger prisms. The difference between the blue and black curve at 550nm is not that significant in my opinion (based on the shown sample size of 2, more spectra to compare would be nice)
I have not yet found a simulator for silver nanoprism spectra, but there are online tools to calculate mie theory for spherical particles*. The extinction spectrum matches nicely to your measured data, -but the absorbance of the spherical particles does not fully support my explanation, as the difference by particles size (for spheres) is not very large- [EDIT: scrap that, the simulator I used uses a varying y axis. The absorbance is significantly size dependent]. This may very well be different for prisms though, any resources for calculating the spectra of silver nanoprisms are welcome!
*TH-cam likes to delete comments with links to further resources, so I'll not link directly here, but instead try in a comment below this one. I'm sure you'll find one by googling "Javascript Mie scattering calculator" or similar.
Attempting to link: saviot.cnrs.fr/mie/index.en.html
You can set the material and the size of the nanosphere and it will show you a plot of the extinction, absorbance and scattering quality factor, which roughly corresponds to what the instrument in the video is measuring.
That's science for ya... the answer to any question provokes several new questions that you otherwise never would have thought to ask. This also illustrates the difference between an experiment and a mere demonstration.
I want one of those lights. Adjusting light color based on it's wavelength. it's so elegant. Genius. I couldn't build one if you gave me a year and unlimited budget.
Apparently bspp acts as a slow oxidizer in the original process. It breaks down over time making the photochemical reduction more and more preferable.
24:41 The way you explained the process so far, that's exactly what I think would happen.
The light of X wavelength gives the particles that absorb X wavelength a growth boost. So they would grow so big till they aren't absorbing it anymore.
Just, wow. Thank you so much. I'm going to be learning about silver nanoparticles for the next week
Hey Ben always good to hear from you.
An ultra sonic bath also works great for degassing solvents especially if you use it in short bursts. Heating the solvent also is a cheap and easy way to do it. It is great to see what you do in your garage, while we often struggle with stuff in a multi million dollar lab.
Super funky. I've only seen this with gold nano particles, but that was a while ago. Love this.
@14:10 according to your assumption, one could make an artificial paper with enclose microdroplets of silver nanoprisms, which then could be "imprinted" by projecting a multicolored master image at 3 (or more, here just to stick alongside the RGB concept) specific wavelengths after each other, and so "save" a colored photo on type of "nanoparticle" paper.
Great stuff Ben ! The optical properties of silver are fascinating. Photography, mirrors and now colorful nanoparticles.
I believe holographic bandpass filters are made in a similar way, except they do use a laser. The best description I could find was from a company called Tydex who says: "HNFs are made by recording interference pattern formed by laser beams in a layer of dichromated gelatin held between two plate glasses." Which doesn't really explain a whole lot, but it sounds cool I guess.
I did this experiment years ago with nano spheres. The interesting part is that unlike the butterfly scales in that case it was the surface area and not the shape that makes the color,
This man deserves respect. He is the Bob Vila of shop tech.
Might the redshift be caused by the heat produced by the LEDs? Lots of IR radiation coming off the glass and foil.
Great video. Ag Nanoparticles were always fun in the lab. We would grow plates by using specific ligands (much like the PVP shown here) and controlling the size with time and hydride concentration. This method is much better. Thanks for sharing.
The red-shift is most likely caused by the process not being 100% efficient from an energy perspective (nothing is). If you are familiar with fluorescent systems, it would be the binding equivalent to a "Stoke's Shift".
To investigate the energy gradient hypothesis you could check if you can always turn smaller particles into bigger ones.
I love how your electron microscope looks like it was developed at the Los Alamos Trinity Lab.
I've work with EM's and know the pain of sample prep !...cheers.
The "redshift" could be explained by the following: if the particles that match the LED wavelength are caused to grow, then they'll stop growing when they are longer than the LED wavelength
I think the red shifting is linked to the non reversibility you observed. The particules, by being close to absorbing the wavelength have still an energy advantage, so they continue to grow, until they don't have benefit anymore, thus overshooting systematically because they can't go back
Ben, this is super cool, as has become expectable from your channel. So I will give a good hint on the reaction mechanics:
Why Absorbtion wavelength is longer than the incident light used to make the particle:
The peak activity of the reaction is at the incident wavelength. This means that the peak growth rate is at incident and the growth rate drops to zero near the absorbtion wavelength. The variability may be due to agglomeration. I bet the math is similar enough to:
dLambda_PDF/dt = k1 * convolution(Lambda_PDF, Center_And_Shift(Lambda_Incident_Intensity_PDF), dLambda) * dt.
Where PDF is probability density function so as to account for the statistical aspects of the reaction mechanics and Center_And_Shift shifts to start at zero to mathematically preserve wavelength of activity. It looks like a convolution to me because a convolution would explain 2 peaks adequately as the reaction would never complete without a broader wavelength source, even in a perfect world with no byproducts. To improve yield, use a peltier device stack to vary the LED wavelength with temperature.
Awesome project! Makes me wish for a home lab and more time on my hands.
This guy is so smart, it always hurts me to watch while being interesting at the same time.
You're an icon of DIY science!
Thank you so much. I snagged all the parts on Digikey to build the broad spectrum LED board. I was planning to use a selection of noble gases for same but this is so much easier! This will also be great for the calibration of my telescopes optical spectrophotometer without having to point at known stars. Awesome content as always.
Really interesting video! It's also refreshing to hear which steps did not work in your process.
Rainbow llama blood, fascinating, real research , great choice pf project. a clear energy transfer , the mechanism unknown, exciting. , the frequency shift. Love seeing the SEM images. so the particles are not 600nm like the cd pits but ~10x smaller. , Inspirational science as always. I bow to your use of chemistry. Silver, highest thermal conductivity, made some silver heatsinks for a micromouse competition.
You might want to try exchanging the nanoparticles coating to sulphur based ligands such as undecylthiol. They will form a monolayer on nanoparticles and faciliate centrifugation, since nanoparticles will dissolve in aliphatic solvents and precipitate in polar ones. In my experience this is the easiest way to get rid of the bulky polymeric coating.
Your videos are all great, your attempts to taker it further are always interesting.
I absolutely love how this is explained like a tutorial, as if I’m going to need to do this at the end
I felt like I was supposed to be taking notes for an upcoming lab.
@@davidwillmore yeah! I kept reminding myself it was ok that I didn’t really know what was going on because I didn’t have to turn anything in
There's a sweet irony about mentioning the hydrated forms of chemicals, and using the bag that says "It's just sodium citrate, nothing else!" on it as an example.
There are many comments so this will probably get buried, but I did my graduate research in a lab that worked with plasmonically grown silver nanoparticles/prisms and my PIs graduate group is the lab that first published the work(not my work but in my lab, you can google my work if you want to). These are not colored by structural color, structural color is defined as the surface of a material being randomized (usually on the micro scale) as to produce changing reflection/diffraction (sort of like those weird films on credit cards that reflect colors strangely). These produce color because silver has localized surface plasmon resonance or we call them plasmonically active at these size scales (20-1300 nm or so) with increasing wavelength visible light where the physical height of the dorito shapes absorb at the wavelength proportional to their size (the small width absorbs at a different and much smaller wavelength!). The effect is caused by the wavelength of light coupling and oscillating the bands of the metal and causing a separation of electron/hole pairs at the surface (which are chemically active in the right environment!) and this is most effective near and below the peak absorptive wavelength for that size (metal species dependent). The effect is most strong at the tips of the doritos but grows out from there. The field of plasmonics gets way more in depth but you can look up more from there.
To answer some of your questions that im fairly sure of the literatures response.
1. The BSPP is a chelating agent (binds up free metal ions), a slight etchant, a surfactant (keeps them supported in solution and not stuck together), and a light sensitizer for the particles which helps them grow with the light (complicated, literature can explain), PVP works similarly, but isn't as strong a chelating agent.
2. The particles likely grow from a combination of Oswalt ripening (cannibalism of less stable crystals toward more stable less undercoordinated atomic positions) and PVP etching producing free silver ions, similar to your theory about growing into the preferred size and shape for the light, but this is forced by the active electron/hole pairs being split across the height of the particle by the plasmon resonance.
3. The particles can't grow back into smaller ones because the large particles can't absorb the shorter wavelength light as they are not resonant with those shorter wavelengths, so nothing happens. Also the larger particles are more stable so its hard to go the other way without adding some other chemicals and starting over.
4. These prism are often made into right bipyramids by accident, meaning they would have 2 LSPR (absorptive) wavelengths in the visible spectrum, so while a distribution is the most likely case due to extremely slow ripening and etching at that size due to particle stability for those 470 particles they could also be becoming bipyramids (trigonal pyramidal shape with a flat bottom).
5. The LEDS you are using are likely fairly narrow wavelengths (+ or -) so the particles probably aren't growing too much bigger because of that but that could be some of the cause of the redshift, but plasmons exist outside the particle itself so they will likely outgrow the plasmon excitation and become somewhat larger than the wavelength of light (then stop growing) so that will redshift the absorbance as well.
6. More intense light does speed up the reaction kinetics if there is free silver in solution post NaBH4 reduction, but this could cause bad effects and different products because faster doesn't often produce the same thing faster, but something else faster. in this case where Oswalt ripening/etching is the likely silver source, the limiting reagent is those processes so intense light would only help the kinetics to a point.
7. Standard workup for our particles was 1 mL spinning at 6-9K for 3-10 minutes to produce those black pellets, decanting the supernatant with a pipet, re adding back DI water and spinning it down 1-2 more times and repeating. The final pellet was then resuspended in a tiny amount of water and sonicated or vortexed to disperse it in that tiny bit of DI water. That little bit of water with suspended particles is really concentrated so you only needed like 1 uL of it dropcast onto silica wafers or TEM grids for imaging. (they are small, the bigger particles you can probably see with a tungsten filament SEM, but a field emission SEM is preferred. I think you would need a resolution at least 2-3 nm to get a decent image of them (the field emissions are often around 1 nm) but most TEMs would be fine, but many $$ for any of those)
These reactions were always picky as heck and didn't like to behave sometimes so other stuff often happened (water purity and vials themselves could ruin the reaction). If you have any more nanoparticle questions or would like references, I can probably find them or point you toward someone who is better versed on the topic.
This is fascinating on so many fronts. I would be interested to see what the energy difference between the LED wavelength and absorption peak wavelength is plotted against the input LED wavelength. Without hyperbole, I would have watched the whole thing even if it was 5 times as long. Keep up the great work!
GREAT little nugget of info at the end there... It almost makes it seem like you can use the initial LED to get to around the right size then shift back down a bit to target a particular size. The real challenging question is then how do you separate them based on size to isolate certain sizes to block out specific wavelengths.
Love your channel, Ben. Would love to see you collab with Tom on Extractions and Ire. I bet you could help him his cubane synthesis series, he is having trouble with an IR reaction. Glad to see you’re still making videos!
This is literally like my biology, physics and chemistry classes from school but for adults that I actually enjoy 👌
You didn't have to make the PCB for your light machine so beautiful. You could have just hacked it together on a breadboard or done a rough solder job. Good show!
24:00 Could it be that the reason the side lengths are longer is because the prisms are actually growing from the light interacting with different chords through the triangle rather than the edges? Perhaps the absorption spectra for the formation process is some bell curve between side length x and x*(sqrt(3)/2), which would explain why the 470nm light's absorption spectra is between 470-543 (470 * sqrt(3)/2)
It might be slightly different due to the depth of the prism itself but I suspect the difference is negligible, and that would explain why extended exposures likely don't increase the sharpness of the peak as well, meaning lazer light probably wouldn't have an appreciable effect on the prism growth so a broad spectrum with large exposure area is likely favourable.
I haven't done any math bar googling what sqrt(3)/2 is but it seems plausible to me! I thought of the height because I was thinking of what the light is more likely to be absorbed by, hitting the face of the prism rather than brushing the edge? Not sure if that line of thinking is at all valid but the number does seem promising!
Sorta makes me wonder in which dimensions the prisms are growing, how quickly
@@Erhannis Schema 4 in "Mechanistic Study of Photomediated Triangular Silver Nanoprism Growth" by Can Xue et al shows that the particles can grow in two different ways, depending in the wavelength used to illuminate the sample (Dipole is longer wavelength than quadrupole, but both vary with particle size, the larger the particle, the longer the wavelength). The sides grow faster than the face.
EDIT: I would link directly to sci-hub, but TH-cam doesn't like links.
I'm wondering if it has more to do with the area than the side length. There's no reason to expect the light to correlate with a specific linear dimension.
What would happen if the light was polarized?
The peak at longer wavelength represents a long axis in the nanorods that you have made. The shorter wavelength corresponds to the short axis of the nanorod.
It's cool that you're investigating by using the tool you have, electron microscope.
The particles seem like a *really* good opportunity for diffraction studies.
I don't know; this is my first video I've watched from this channel but when I heard that "...pretty crazy. Right?", something clicked in my mind.
In order to reduce the time of exposure, you may want to add a scattering agent (TiO2?0) during the conversion phase, which you may be able to filter out. additionally you may use a modified integrating sphere to as a perfect internal reflector.
This reminded of interference effect pigments where mica or borosilicate platelets coated with TI02 with the thickness determines the color. I used to work for a pearlescent company for a few years.
Amazing detail and tenacity as usual. Fabulous to see really interesting chemistry in such complete "stories"
I would speculate that the reaction is not reversible at all. Unreacted molecules in the solution react with the new wavelength of light creating a solution with not one but two preferentially sized particles depending on the light wavelengths used. The overall process is very similar to that in photography using silver chemistry. Once the process has occured, i.e. the film has been exposed to light, development and fixing prevent any further transformation of the materials. However multiple exposures to light do not result in the overall process reversing at any stage.
A variable that could be interesting to explore is temperature. The growth rate should increase (significantly) at elevated temperatures. So you wouldn't have to wait days for the particles to reach the desired size.
I saw this months ago, but just had a thought here regarding growing the red particles. Since the absorption is so low initially, I expect to see improvements by using a series of colors with the wavelength extending towards red to maximize absorption and efficiency.
Quite a cool experiment. Almost like the particles are "learning" what light to absorb based on the light shown to it.
Be REALLY careful when anthropomorphizing...anything really.
It is SUCH a convenient tool, especially for explaining things.
But it also invariably ruins the explanation, and ends up causing more confusion than it solves.
Even passively anthropomorphizing is a problem.
Example 1: "Machine Learning" and Artificial Intelligence" sure do FEEL like those names literally apply. But they don't.
In fact, out of those 4 words, only 1 has anything to do with what they describe. And even that one is a stretch.
...
(And that word is 'Machine". Just to be perfectly clear.)
Example 2: So few people actually understand what 'Evolution' describes [†] that the term should never be used if you actually want to talk about it properly.
It's so deeply rooted, and deeply wrong, that even I have to stop and remind myself each time I think/hear it, because otherwise it's just going to keep merrily rolling along in the wrong direction.
Now, I'm no Idea-ologist, but I'm pretty comfortable attributing this complete lack of understanding to explanations like "How a species adapts to be more successful in it's environment." [‡]
[†] I'll award 42𝑖 internet points to anyone who even comes close to a reasonably accurate, and TERSE explanation.
[‡] Clearly this is not the right answer since I'm pointing it out as the root of misunderstanding. I would even go so far as to say it is probably as far as you can get from the right answer, without being the opposite of it. (Hint. The terms are not opposite. But the ideas motivating the USE of those terms might be...)
ALSO: Yes. I did footnote a TH-cam comment.
@@Prophes0r Dude. I've never had anyone write such an indepth response to one of my simple comments before. I would agree on your machine learning example for sure. That's pretty close to the kind of "learning" I was referring to, for lack of a better term. Indeed for both there is no intelligence on the part of the chemicals or the machine. Only that of the scientist or the programmer.
@@Prophes0r Thank you for posting this. I think excessive anthropomorphism is causing real issues in our society right now especially around "machine learning" / AI. This, for instance, causes people to vastly overestimate the capabilities of self-driving cars for instance. They SEEM to be "intelligent" in a human sense, except that they really are not...
@@Prophes0r this makes me think of how we view our own intelligence... Always giving a positive value if we can arrive at something useful or something we, for some reason like and vice versa. The concept of 'intelligence' gets hijacked by our individual or collective whim and then used as a reward or punishment to modify peoples thinking and behaviour.
We all want to be 'clever' not seen as an 'idiot' while actual intelligence requires hard work and in this day and age hard work is not clever...
Hi Ben, I am an Electron Microscopist and I can help you in capturing the structure with my normal and cryogenic TEMs. Also I suggest to boost your resolution with STEM detector. There are two types passive and active. Passive STEM detector is simply a gold coated plate at 45 degrees below the sample. Active has an electron detector: you can strip a photo detector and use it as a electron detector. Both are very easy to build.
That is nutty! I suspect this is something to do with why they used silver nitrate in photographic film, it's light sensitive.
This is the kind of content that keeps me on youtube.
Love the idea of nanoparticles and nanostructured it feels like the bridging technology between where we are now and building designer atoms
Electron microscope resolution is its best possible spot size. Vibrations, electrical/magnetic noise, column contamination, etc. will blur the actual image beyond this ideal unless they are located and eliminated. The best resolution will also improve at higher acceleration voltages since electrons repel.
Nile red beakers, I love how you teachers intermix.
I'm surprised such a low density of a few shattered snowflakes make such good structural pigments. Thanks for the great science again.
Neat! Similar to the blue butterfly, the first time I learned of this visual phenomenon was when I found out the Bluejays I gave peanuts to at my window were actually brown!
Pretty cool as always! 👍
Just a small note: A PCB with white or black solder mask would probably be preferable for optical reasons.
That's really interesting, especially that in the last solution the "peak" shifted away from the 400nm mark further.
Did a similar experiment in college. Used photospectrometry to measure the quantity of silver nano particles
This actually has a lot of research done in the sphere of photography
Gurney and Mott model of latent image Ref. 29. AgBr remains in ionic form Ag + Br − in the crystal of the grain. Radiation produces ionization of Br − to Br+ e −. These electrons make the speck negatively charged. The Ag+ migrate to neutralize the speck and forms a lump of Ag aggregate on the speck.