This summer, I’ve joined the Doyle Group at Harvard and MIT Center for Ultracold Molecules (CUA) for the my second summer of work on a project managed by Dave Patterson, a senior scientist at Harvard. This project, in its most general sense, deals with the rotational spectra of large (or at least what physicists consider large, >4 atoms) molecules in the “cold” temperature regime. Rotational energy levels are very close together, so a huge number of these energy levels are populated at room temperature. When cooled to ~5 Kelvin the rotational spectrum is greatly simplified and thus useful for the identification and study of molecules. Throughout this post, I’ll refer to “pure rotational spectroscopy” as just “rotational spectroscopy,” though sometimes “rotational spectroscopy” is used to refer to ro-vibrational transitions, which is a similar but very different can of worms.
I built the innards of this box last summer largely from scratch. Now being used by the McCarthy group to study reactive intermediates.
There are two schools of thought about how exactly to accomplish the cooling necessary to perform rotational spectroscopy: seeded supersonic beams, and buffer gas cooling. Seeded supersonic beams allow for effective rotational cooling of large molecules by having a gas expand into empty space, which gives a relatively uniform distribution of temperatures and velocities. Spectroscopy can be performed on this expanding gas, though it has the disadvantage of needing some “recharge” time for the prior butt of gas to be pumped out before the next one is allowed to enter. Typical duty cycles are on the order of 60 Hz, and there is some fairly expensive tech involved in getting a small enough amount of sample into the vacuum chamber (referred to as “pulse width”) for the vacuum pumps to maintain the appropriate vacuum and thus the appropriate level of rotational cooling. Note that this isn’t cooling in the traditional sense, but cooling of only two of the molecule’s degrees of freedom, that is vibrational and rotational. The gas expands at a supersonic speed relative to someone standing in the lab, but are moving very slowly relative to the other molecules in the cloud. This was the first cooling method developed to reliably cool large amounts of diverse kinds of molecules to the cold temperature regime.
One of the other commonly used methods, buffer gas cooling, was developed by the Doyle Group at Harvard in 1998 and has many advantages over seeded supersonic beams. It is a constant source of cold molecules, rather than a pulsed source, allowing pure rotational spectra to be collected in a much more expedient fashion. Buffer gas cooling takes advantage of elastic collisions between the buffer gas (helium is used in the 6K temperature regime) and the introduced molecule to thermalize the incoming hot gas to 6K, and rotationally and vibrationally cool the molecule. The injected sample is of low enough density that the number of collisions between injected molecules in the cell is negligible. Interestingly, helium-molecule van der Waals complexes are not observed in rotational spectra of injected molecules. The exact mechanisms of rotational and vibrational relaxation have been studied in excruciating detail by the Doyle Group. Once molecules have been cooled in a buffer gas cell, an aperture can be opened from which molecules in the cell can escape and be directed into a beam for optical trapping, loading, or further cooling to form Bose-Einstein condensates. In the Patterson lab, we just cool the molecules enough to simplify rotational transitions and perform microwave spectroscopy without extracting them for further use. It is worth noting that molecules injected into a buffer gas cell stick to the walls after a fairly short time, about 20 uS, so more constantly has to be injected into the cell for continued data gathering .
Once the molecules are cold and floating around in space, the real work happens: broadband microwave radiation is poured into the cell, causing transitions from lower energy rotational states to higher energy rotational states. The current method to do this has only been around since 2006, when the Pate Group at the University of Virginia realized that advances in electronics allowed for “chirps” of microwave radiation to be introduced to the cell . A “chirp” is a frequency sweep, so that transitions at any frequency that the chirp accesses are made. There is some sacrifice of resolution in order to obtain this broadband ability, but it is well worth it as most transitions are strong enough for resolution not to be an issue. Simple rotational spectroscopy is useful for learning about bond lengths, rotational constants, and dipole moments of pure samples, but until very recently hadn’t been used as an analytical tool for chemical reactions or processes (I supplied the sample for the first time this is known to have been done). The Patterson Lab seeks to change this, and has shown that rotational spectroscopy can be a very powerful and specific mixture analysis tool, and the timescale and temperatures are fast and low enough respectively to see reactive intermediates. There is currently a postdoc studying reactive intermediates of ozone decomposition in the Patterson lab, with some very promising results about molecules that haven’t really been seen because of their longevity or rather a lack thereof. .
Part of this progress towards a useful mixture analyzer is the development of chiral resolution techniques, which is the current bread and butter of the Patterson lab . This technique relies on the phenomenon of double resonance and sum frequency generation to produce signals that change phase with enantiomer (species of chiral molecule), which in turn relies on the dipole moments of a molecule. It is best to think, from a spectroscopist’s point of view, of a molecule as a “dipole in a brick.” There are three orthognal dipolar moments, and for a vast majority of molecules, these dipole moments are all different, which makes it an “asymmetric top” as opposed to a “linear top” or a “symmetric top.” It is worth noting here that molecules where the linear combination of all dipole moments are zero, there is no rotational spectrum because there is no transition moment. Chiral molecules can be thought of as with both enantiomers sharing two of the dipole axes, and the third being mirrored.
The three mutually orthogonal dipole moments in the molecule are taken advantage of by orthogonal fields in the experimental cell. Each field corresponds to a rotational transition in the “triangle” of resonances: one corresponds to “a” type transitions, one to “b” type transitions, and one to “c” type transitions. “a” and “c” type transitions are driven by electric fields resonating on order of magnitude 100 MHz and 10000 MHz, respectively, with the third side of the triangle being emission of microwaves at the frequency that is precisely the sum of the “a” and “c” frequencies (this is the sum-frequency generation part). The critical bit of this sum frequency generation is the phase of the resultant microwave emission. This emission is collected by a horn, and amplified into a voltage V, done in such a way to preserve the phase, which is dependent on enantiomer.
So far, this just sounds like an incredibly complicated way to go about doing the same thing that a simple polarimeter can, so why go about doing all of this? It all comes down to specificity and sensitivity. The sensitivity of this technique is high because the signs of the phase of emitted radiation are exactly opposite so racemic samples have exactly zero signal when done using this three-wave mixing technique. This allows very very small enantiomeric excess to be measured (2% enantiomeric excess can clearly be seen, where the limit for a polarimiter is much higher). In addition, it is possible to make these measurements in complex mixtures with >20 components, and still assign the specific enantiomer, as well as other chiral compounds in the mixture.
My goal is to take this to the next level, by showing that it is possible with this technique to resolve molecules that are chiral only because of isotopic substitution. To do this, I spent many many hours in the Broadrup/Blase lab this spring synthesizing both racemic and chiral samples of 1-pentanol, with a deuterium substitution at the 1-position. This creates a chiral center in a molecule that is otherwise completely achiral, and it is only barely chiral (this deuterated species has essentially the exact same physical properties as the non-deuterated achiral species). This was quite a challenging synthesis because pentanol has a relatively low boiling point and has no chromophore, so is fairly difficult to extract from a mixture of solvents. One other possible molecule is the benzyl alcohol version of this, which has synthetic advantages in the chromophore and spectroscopic advantages in that it is not a long, floppy chain like pentanol. At the temperatures dealt with in this experiment, several conformers of pentanol are stable and possible, and each conformer has different rotational and centrifugal distortion constants, cluttering the spectrum.
Thus far, we have taken a full spectrum of a racemic sample of deuterated 1-pentanol, assigned rotational constants from that spectrum, and confirmed that I did indeed make deuterated 1-pentanol in the enantiospecific synthesis. There is much work yet to do, and precious little time in which to do it, so I’ll be off to the lab.
I’ll try and post about once every week and a half with some interesting story from the rotational spectroscopy world, whether it’s equipment malfunction or smashing success.
This is a strong 1-D-Pentanol line, confirming that the asymmetric synthesis worked!
Postdoc Sandra Eibenberger pounding away at MatLab scripts
 Molecular optical spectroscopy with supersonic beams and jets, Richard E. Smalley; Accounts of Chemical Research 10(4):139-145 (2002)
 Spectroscopy of buffer-gas cooled vanadium monoxide in a magnetic trapping field, Jonathan D. Weinstein; The Journal of Chemical Physics 109(7):2656-2661 (1998)
 Cooling molecules in a cell for FTMW spectroscopy, David Patterson; Molecular Physics 110(15-16):1757-1766 (2012)
 Buffer Gas Cooled Bose-Einstein Condensate. Doret et al. Phys. Rev. Lett. 103, 103005 (2009)
 The rotational spectrum of epifluorohydrin measured by chirped-pulse Fourier transform microwave spectroscopy. Brown, G.G.; Dian, B.C.; Douglass, K.O.; Geyer, S.M.; Pate, B.H. (2006). J. Mol. Spectrosc. 238: 200–212.
 Sensitive Chiral Analysis via Microwave Three-Wave Mixing. Patterson, D.; Doyle, J. M. (2013) Phys. Rev. Lett. 111, 023008 (2013)