Most Detailed Photos of an Atom Yet

BuzzSkyline writes "Ukrainian researchers have managed to take pictures of atoms that reveal structure of the electron clouds surrounding carbon nuclei in unprecedented detail. Although the images offer no surprises (they look much like the sketches of electron orbitals included in high school science texts), this is the first time that anyone has directly imaged atoms at this level, rather than inferring the structure of the orbitals from indirect measurements such as electron or X-ray interferometry."

Comments

really?

[timmarhy]

looks like it was done in MS paint to me...

Re:really?

[ModernGeek]

The shadows are all wrong.

Re:really?

[Canazza]

Physicists are not photographers! They obviously had the shutter speed too low - Look how blurry the picture is!

Re:really?

[miro f]

can't really call it a photograph if it was taken with electrons rather than light.
I wonder how they can tell the electron is blue?

Re:really?

[nacturation]

Looks like a smurf sat on the photocopier.

Speaking as a chemist

[PatrickThomson]

This is amazing. We'd theorised orbitals to exist, and they worked very well. We could calculate the shapes of molecules and make detailed predictions that came true to 10 decimal places. Quantum mechanics as applied to electrons in atoms is the most successful and the most rigorously tested theory ever developed.

And yet, to finally see a real orbital, not a simulation. Looks like a 1s and a 2p, right there for the looking!

Re:Speaking as a chemist

[PatrickThomson]

We can only "approximate" orbitals in atoms with more than one electron, true. That's the same as saying that numerical methods won't "exactly" solve a function. We can still get really accurate results, even if it's computationally expensive. FT-IR of heteronuclear diatomics, anyone?. Orbitals still retain the basic shapes in multiple-electron atoms.

Re:Speaking as a chemist

[anarchyboy]

Depends what you mean by real, certainly splitting your many electron wave function into orbitals works well and allows an accurate approximation of the system as a whole. The orbitals do form a basis of functions for the system (with some acceptable approximations) so while your quantum state is the whole thing you can think of it as being built of oribitals. This is true in a mathematical sense its an expansion of the quantum function as a set of orbital functions. This is as valid as any other expansion like a taylor or powerseries expansion. So orbitals are real enough in that sense. You can then calculate observables for the individual electrons since your operators will act only on the single electron and your oribtals are normalised the rest of the electrons essentialy go away and you can calculate things like the average radial ditance etc and build up pictures of what that electron "looks like". Since the orbital functions are calculated (numerically) as a multi electron system even though the end product allows you to look at individual electrons as orbitals the overall wave function (all the orbitals combined) is still a very accurate picture of the system

In fact the experimental evidence showing a physical picture of these orbitals just goes to show that this is in fact a very sensible and useful way of picturing your atom.

Re:Speaking as a chemist

[PatrickThomson]

Basically, a chemistry education is very much like fast-forwarding through 300+ years of science history. Some dead-ends are skipped, but by and large, the simpler and more self-contained a theory was, the older it is and the earlier it's taught in school. The university-taught molecular orbital theory is (debatably) too rich and complex to be taught any earlier.

The moons-orbiting theory fit with all the available evidence at the time it was developed. Think of orbitals as clouds of probability where, if you tried to pin down the electron, it might be. A moons-orbiting theory would give this probability cloud as a thin donut around the atomic waist. The shapes of orbitals as depicted in wikipedia etc. are consequences of the maths of quantum mechanics. It's annoyingly non-intuitive.

Re:Speaking as a chemist

[PvtVoid]

Speaking as a chemist, could you explain what exactly this means? Up until this very moment I have been under the misguided notion that the nucleus of an atom was orbited by electrons within groups called "shells", and these worked very similarly to satellites around a planet.

You're thinking of the Bohr model.

So, could you in any way explain how we get from "think of it as a planet with many moons" to this or more importantly, what gives orbitals this shape?

It's because the Schrodinger equation is a Laplacian, and the hydrogen atom is a spherically symmetric problem. The natural basis for the Laplacian in spherical coordinates is spherical harmonics. The shape you are seeing is the characteristic shape of different spherical harmonics, corresponding to the angular momentum of the electron.

Re:Speaking as a chemist

[The_Duck271]

At atomic scales electrons cannot be thought of as points; instead they are smeared out probability distributions. They don't exist at any given point, there's a chance for a given electron to be found throughout a whole region of space, and the probability of finding it at any given point is given by a probability distribution. These probability distributions are called wave functions, and given an electron's wave function you can calculate the likelihood of getting different results when you take a measurement of the electron. It is a strange aspect of quantum mechanics that you can't calculate exactly what you will measure, you can only establish the probabilities of each possible outcome.

Another aspect of quantum mechanics is that if you measure, say, the energy of an electron in an atom, you can only get one of a certain set of discrete values, and never any energy in between those values. The energy of the electron is quantized. In general, if you measure an electron's energy you have a certain probability to get a result corresponding to the first energy level, a probability to find it in the second energy level, and so on. This is also the case for some other things you can measure, like angular momentum.

However, there are certain wave functions that correspond to exactly one value of energy; that is, if you have an electron with this wave function, you are guaranteed to get a certain energy value when you measure it. In fact, there is a special set of wave functions with the following three properties:

• They each have a definite energy level.
• They each have a definite total angular momentum around the nucleus.
• They each have a definite angular momentum around the z axis.

These wave functions are the atomic orbitals that are so important in chemistry. If you calculate the shapes of the wave functions that satisfy these properties, you get the shapes shown on the Wikipedia page. They are listed in a table indexed by the variables n, l, and m. n corresponds to the energy level, l corresponds to the total angular momentum, and m corresponds to the angular momentum around the z axis. For example, you can see that orbitals with high m (angular momentum around the z axis), like the ones on the very right of the Wikipedia table, are sort of flattened out by the centrifugal force from spinning fast around a vertical axis.

Re:Speaking as a chemist

[locofungus]

It's wrong to think of the electron as a particle when it's "orbiting" in an atom. Instead you should think of it as a probability density. This is Schroedinger's cat all over again, the electron is "smeared out" all over its "orbit" but instead of being "half dead, half alive" it's x% here, y% there.

This is also like the two slit experiment. The electron doesn't go through one slit or the other, it goes through both slits (not 50% dead and 50% alive; 50% went through that slit and 50% went through this slit) but when it hits the phosphor screen it's a particle as its "where is it" probability function collapses to a point.

The "wavefunction" is (as far as we can tell) a mathematical curiosity that when squared gives us the observed probability function. The probability distribution is real, the wavefunction gives us a convenient handle to calculate probabilities and how they evolve.

But now that I've said that the wavefunction is an imaginary curiosity, imagine a sine wave on a string and then join the two ends of the string together. There will only be a few discrete lengths of string where the sine wave will "join up" correctly (ok there's an infinite number but the length of the string is limited). It turns out that, with rather a lot of unpleasant maths (see the wikipedia page for spherical harmonics), our wavefunction works like that sine wave and we find that there are only certain orbitals where the wavefunction is well behaved.

Tim.

Re:Speaking as a chemist

[master_p]

Could it be that we can't pinpoint the exact position and velocity of an electron at the same time because they are interlinked with all the surrounding particles? i.e. the act of measurement affects the outcome.

Re:Speaking as a chemist

[Anonymous Coward]

Based on your previous posts, your lack of education was never in doubt.

Re:Speaking as a chemist

[Barterer]

Very informative, thanks. When you say the electrons have a definite momentum about the Z-axis, do you mean just one chosen axis (depending on your perspective and axes you assign) or does it have something to do with which way is "up" or gravity?

Re:Speaking as a chemist

[brian0918]

You can turn down the beam current in the two slit experiment until you're talking about orders of magnitude less than one electron in the apparatus at any one time on average and you still get the diffraction pattern.

That's not correct. See experiment and photos here (Figure 2). Single electrons produce single dots. It's only after you dump many electrons through that you get a pattern - that's simply because the electrons follow wave trajectories rather than the standard trajectory visualized from classical motion. In reality everything follows these same wave trajectories, it's just that for macroscopic objects, the individual oscillations of the individual particles cancel out.

I don't know why it's any more conceptually obvious that a "variable" should be smeared out than an "electron" should be smeared out.

The phrase "smeared out" conveys nothing, so it should be no surprise that it can be used for situations that are completely different.

Similar Pictures From Switzerland

[Maddog Batty]

"Leo Gross and his colleagues at IBM in Zurich, Switzerland, modified the AFM technique to make the most detailed image yet of pentacene, an organic molecule consisting of five benzene rings"

http://www.newscientist.com/article/dn17699-microscopes-zoom-in-on-molecules-at-last.html

 

Re:Similar Pictures From Switzerland

[L4t3r4lu5]

Your article is much more impressive, IMHO. All I see in the original story is three blue blobs. You could have told me it was false-colour cellular mitosis, and I'd have believed you. I understand that the detail in the story is much higher (imaging one atom instead of a whole molecule) but seeing hexagonal Benzene rings with my own eyes just excites me more.

Re:Similar Pictures From Switzerland

[junglee_iitk]

It is anything but similar.
* That article was about taking picture of big but fragile molecules, even though atoms have been pictures before with ease.
* This article is about even detailed picture of atoms.

Unscaled photo link

[UPi]

The unscaled photo is here:

http://insidescience.org/polopoly_fs/1.918!image/671260397.jpg

Re:Unscaled photo link

[PGC]

Unscaled, wow. That is one HUGE atom.... no wonder they were capable of photographing it.

Re:Unscaled photo link

[fastest fascist]

Not to pick nits, but is a picture that is the result of electrons striking a surface actually a photograph?

Re:Unscaled photo link

[miro f]

I have embedded the unscaled photos in this post.

Millions of them

Re:Why is this significant?

[Genda]

The ability to directly measure electron density is quite an old technique. STMs and AFMs have been doing this since the very beginning.. I agree with the researcher's quote in the article that it's good to develop a complementary technique(FEEM) abd at best that's its contribution. I'd be happy to hear what else it contributes. though I don't quite agree with his or the editors spelling! ;) "it's always good to have complimentary approaches,"

In this particular application, its simply a very cool thing to be able to prove theory with direct measurement. In the future I can imagine viewing electron orbitals for test samples of high temperature superconductors or producing high resolution images of the electron cloud density for a protein (get a better idea of the quantum component for protein folding) might prove extremely useful and interesting.

In my experience, no sooner does someone come up with a better device for viewing, then someone comes up with a exquisite need for that device.

It makes me very suspicious indeed.

[Kupfernigk]

Why? Because the "orbitals" are actually solutions of the Schroedinger Wave Equation. They are images or a probability distribution in abstract space. Electrons are not clouds or points, they are things we don't really understand but describe by means of quantum mechanics. So I am deeply suspicious of the picture, because there is no physical object of that shape to image.

Re:It makes me very suspicious indeed.

[vikhyat]

It's probably like one of those long exposure photographs.

Re:It makes me very suspicious indeed.

[Anonymous Coward]

The article was extremely superficial when describing the actual experiment, but essentially a current was passed through a small chain of carbon atoms by applying a voltage across the chain. The current caused the the carbon atom at the tip to give off electrons to a phosphor screen. I would suppose that these "given off" electrons were integrated (summed) over time and this formed a pattern that reflected the shape of the probability distribution, i.e. orbital. Each electron that was "given off" constituted a sampling experiment regarding electron position, and the sum total of the samples would, over time, give rise to the orbital shape. In the case of an s orbital, electrons were given off in all radial directions. For the p orbital, certain angles gave off no electrons. This behavior corresponds to the quantum equations.

mirror

[Anonymous Coward]

.
(not actual size)

Wow

[Anonymous Coward]

So this is what's powering my netbook!

(do not mod)

[aepervius]

ignore. Reply done to undo bad moderation

Re:(do not mod)

[machine321]

MOD PARENT UP!

Not quite a of an electron in an orbital

[Richard Kirk]

They do look like the classical orbitals, don't they?

However, there are some problems with interpreting the image as a photograph of an orbital. What the FEEM does is to charge up a very sharp point. The actual voltage may not be very big, but the local field strength depends on screening and curvature, so you can get very large electrostatic fields around sharp features, and if you get the balance right, electrons will leave the sharp points, zoom down the field lines, and get imaged. I remember seeing a sharp tungsten needle in a FEEM back in the seventies, and seeing the individual atoms. This sort of thing provided the first real evidence of a screw dislocation. You got a strange projection of the tip of the needle, as the electrostatic field tended to map the roughly spherical tip onto a flat plane.

So what is happening here? Our field stripping an electron from the orbital. We are getting a map of the electron flows as focused by the electrostatic field. We calculate the trajectory back through the electrostatic field and guess some sort of map of emission. They must have stripped hundreds or thousands of orbital electrons from the same atom, and replaced them to get each image. However, if an orbital 'pokes out' of the atom, or forms a 'sharp feature' (inverted commas because they are wave functions, so these concepts are a bit hard to define) then we get a bright spot. The really cool bit is getting the atom to go back to the same hybridization state hundreds of times, so we got the two-lobed picture.

It's dead clever. However, for my money, the atomic force probes are cooler as they can measure the fields without stripping the electrons. But, as the reviewer said, it takes all sorts...

Pity this is AC

[Kupfernigk]

Thanks for responding. This could do with some mod points but I can't mod and post...so I'll respond. It's interesting to think about what is happening here. It's possibly unhelpful to refer in the same sentence to "current" and "electrons" but I know what you mean, though I would rephrase it a little to help my own understanding. The "current" did not cause the carbon atom to give off electrons; rather, the potential difference enabled some electrons to pass along the carbon chain until they left the tip, and the path of the emerging electrons was probabilitistically interfered with in a way that reflected the solution of the Schroedinger wave equation for the outer electrons of the end atom. That's a very interesting experiment. The benefit of using carbon atoms in a molecule is that the bond angle presumably locks the orientation of the P orbitals sufficiently to enable the experiment. So for many atoms it simply wouldn't work, and what we are seeing here is not an image per se but something more like the result of the Rutherford/Geiger/Marsden experiment. It looks like a significant experiment, but the summary is quite wrong as to what is being shown.