Thanks for your questions! We’ve worked mainly with two geometries: periodic arrays of nanowires (our ‘1D’ system) and square nanodots (‘2D’). The height of the structures is between 10 and 20 nm; the lateral dimension of wires or the side length of the squares ranges from 30 nm up to 1 um.
We certainly need to take care that our initial laser pump excitation is not causing too much impulsive heating; we keep our fluence small enough to avoid annealing, melting or any permanent change to the structures (at least when all goes well), which would certainly make models more complicated. It would also likely affect the size of our interfaces across which we measure heat flow as well as the efficiency of acoustic excitation. So permanent changes are certainly something we work to avoid.
Our diffractive measurement looks precisely at (dynamic) changes in the shape of our samples – so if there were some permanent shape change caused by our laser heating, it would show up in our static diffraction pattern. Thus we can have confidence that we have kept our pump power within a safe regime while our static diffraction pattern remains consistent. Furthermore we can repeat our measurements many times even with changing pump power and obtain the same results, and for a few samples we have AFM measurements both before and after measurement which are consistent with each other.
Please let me know if you have further questions.

Thanks for your question! The SAWs we measure are really propagating along the surface below our nanostructures, not within individual nanoscale particles. We need the periodic nanostructures to set the particularly short SAW wavelengths that we excite in the surface underneath, which can remain confined to a thin film layer at the surface, for example.
Please let me know what further questions you may have.

I believe your question seeks to focus on thermal transport properties in the nanostructured systems we study rather than acoustics (where so far we’ve focused on developing the measurement tool); if I’m mistaken about this, please feel free to follow up with more questions.

We have not tried to work with many different materials; you can see the comparison of fused silica and sapphire in the top right graph in the Conclusions box of my poster where the ballistic effect is much more significant for sapphire with its longer phonon mean free path. We’re working right now with new samples on silicon, particularly to follow up on some suggestions that significant non-diffusive heat flow can be observed even in nearly macroscopic systems with this substrate (characteristic dimensions ~10um; see, for example, Maznev et al., PRB vol. 84, 195206 (2011)).

But the material itself only governs intrinsic properties like thermal conductivity, thermal boundary resistance between two materials and, importantly, phonon mean free path. (Phonons are the primary heat carriers in all the substrate materials we’ve been studying.) We’re particularly interested in how the nanostructured geometries affect the heat transport – and this is primarily governed by how the size scale of the heat source compares to the phonon mean free path in the substrate below. This is the motivation behind the curves shown in the Conclusions figure I mentioned above, but that experiment only dealt with the simple 1D case of nanowire interfaces. We are working now on extending similar analysis to 2D nanodot interfaces, where we have early indications that the ballistic contribution to heat transport is even stronger for 2D interfaces compared with 1D interfaces of comparable lateral dimension. Unfortunately I cannot include any extra graphs in the discussion here, but I can send one to you directly via e-mail. We believe this stronger ballistic effect in 2D may be explained by the suppression of diffusive transport in yet another dimension: with nanowires, heat flow can remain entirely diffusive along the length of the wire with confinement only in the perpendicular direction, while 2D interfaces are confined in both directions.

As to which arrangements may be most useful, it depends on what you are trying to do. For something like thermo-electric cooling devices, thermal resistance must typically be maximized, so a system could be designed for as much non-diffusive transport as possible – perhaps with many small interfaces or small constrictions between larger reservoirs. On the other hand, for microelectronics systems, heat should generally be transported away as quickly as possible, and the thermal resistance should be minimized. Perhaps this would require the fabrication of more perfect interfaces minimizing the intrinsic (rather than geometric) boundary resistance; perhaps using extra layers at the interface to achieve better phonon mode-matching across the interface (as suggested by O’Brien et al., Nat. Mater. vol. 12, 2 (2013)) will become necessary. There are a number of interesting studies which explore ways to tune thermal conductivity through the introduction of different kinds of disorder or ways to control heat transport by design and even creating heat devices like diodes or computing elements. For some recent examples, see Bae et al., Nat. Comm. 4:1734 (2013) and N.Li et al., Rev. Mod. Phys. vol. 84, 1045 (2012).

The most important goal for us right now, though, is to better understand the role geometry plays in governing heat transport across nanoscale interfaces; with such understanding, the design of nanostructured systems can properly take this effect into account.

Thanks for the questions!

I appreciate the question. All of our nano-pillar measurements are essentially averaging over around 10^4 to 10^6 individual pillars. Certainly one or a few missing would be unlikely to affect the measurement. If the variation in the periodicity of the pillars were very high or the average quality of the shape were very low, we would not be able to observe a clean diffraction pattern. With our phase-sensitive diffraction measurement which is most sensitive to changes in the height of our structures, significant variation (a range of perhaps +/- 15%) in initial height of the pillars might confuse our dynamic thermal signal – although, again, we are measuring the average over many pillars.

The acoustic measurements will be sensitive to distributions in the nano-pillar periods (for the surface waves) and in the pillar heights (longitudinal waves). But here again, assuming some distribution in the dimensions around a mean, I believe this would primarily be observed by a broadening of a frequency peak and the data should still be interpretable – again basically giving us information about the average pillar.

One technique we hope to apply over the next year or so is to combine our pump-probe measurements with the coherent diffractive imaging (CDI) techniques used by other members of our research group (see for example, Seaberg et al., Opt. Exp. vol. 19, 22470 (2011)). CDI essentially offers a lensless microscope using femtosecond-pulse, short-wavelength extreme ultraviolet light for high spatial and time resolution. This would allow us to observe individual nano-pillars as they undergo dynamic thermal, acoustic and other processes.

Thanks for some excellent ideas and questions, Eamonn. I would say our model is semi-empirical – we observe significant deviation from diffusive heat transport and simply try to quantify that difference by allowing the thermal resistance at the boundary in a diffusive-model simulation to vary as a free fit parameter to our data. (Those are the data points shown in the top right of the Conclusions box.) So our approach was just to apply the regular diffusive model with an extra resistance which we attribute to a ballistic contribution. You can read more about this in Siemens et al., Nature Materials vol. 9, 26 (2010).
There are now more models which try to deal more directly with the phonon transport as it transitions through this quasi-ballistic regime, in particular some which try to separate phonon modes in a system into heat-carrying diffusive modes and ballistic modes. You may be interested in this recent review of many of the simulation and experimental techniques being applied to this problem: Luo and Chen, Phys. Chem. Chem. Phys. vol. 15, 3389 (2013).
One of the approaches discussed in this review does indeed involve molecular dynamics simulations. We have not yet tried to apply them to our work, and there are certainly some challenges still involved in such an application (unknown intermolecular forces at interfaces, particularly imperfect interfaces; relatively large systems to simulate in some cases). That said, we are always looking for new collaborators to help us explore new approaches such as this one.
Please let me know if you have more questions.

OK, thanks. Definitely some interesting stuff going on there. Keep up the good work.

Thanks for the compliments, Samson. That was certainly a book shelf I was proud to sit in front of. :)

Extreme UV, Heat Transfer, Nanotech, and Ender’s Saga! Loved the video Kathy!

It’s great to hear you enjoyed it, Carlos! And it’s always fun to discover more Ender fans.

Thanks, Bosheng! Glad you liked it.

Thank you, Joni! It’s great to hear that you’ve enjoyed something we put so much time and work into.

You’re welcome, Dennis. So glad you like it! And I hope it is helpful to friends and family.

Thanks for checking it out, Steve. I’m really glad you enjoyed it!

Haha – I’m glad you enjoyed this video. I’m afraid something about my name wouldn’t be quite so interesting. Just me from a Dutch American immigrant family (Hoogeboom) marrying someone from a Dutch Canadian immigrant family (Pot) and not wanting to give up my name entirely. Thanks for checking out my research!

Hi Kathy, you can take this whole thread offline and simply email me directly. I’m glad to see there’s a Canadian component. Is your partner also a researcher? I run the website www.science.ca and I just retired as in-house writer for the Office of the VP Research at Simon Fraser University in Vancouver. Recently I wrote about Nanotech and the question came up whether we describe it as “atomic” or “subatomic” level research. Personally, I thought it should be atomic, but we settled on using both words. Your video made me feel like perhaps we made the right decision. But not sure. To me, subatomic is particle physics, atom smashers and stuff, not nanotech. But what’s your take?

Thanks very much, Margery. The potential medical uses of nano-particles – particularly with heat – are certainly a very interesting subject that is part of our motivation. Thank you for sharing our work with him!