Solid State NMR Facility

W. M. Keck Solid State Nuclear Magnetic Resonance (NMR) Facility

A multi-nuclear facility dedicated to the analysis of bio-geo-cosmo-mat sci-chemistry samples

The W. M. Keck Solid State NMR Facility, Geophysical Laboratory, on the Broad Branch Rd. Campus, Carnegie Institution for Science. This instrument is a three channel system with considerable versatility. We collaborate far and wide from Geochemistry, Biochemistry, Materials Chemistry, Cosmo Chemistry, Paleontology, ... we are game to collaborate!

INTRODUCTION: Solid State Nuclear Magnetic Resonance Spectroscopy is, in my opinion, one of the most useful tools in geochemistry, yet it is employed in only 4 Earth Science departments out of ~600+ in the US and just a few more geoscientists (in Asia and Europe).

Of course SSNMR remains well represented in chemistry and physics departments around the world- but I would argue that SSNMR facilities should also be in Earth Science departments. Below (and through my research) I will attempt to explain why.

So if, as I claim, Solid State NMR is so useful to Geochemists why is it that so few Geo-departments have SSNMR's and why is Geochemistry as a field so dominated by mass spectrometry and not SSNMR?

I believe two factors are involved:

1) learning to operate and run an NMR facility independently requires,

A) an introduction to Solid State NMR which is obviously rare to geoscientists - catch 22 or endless circle (I thank Pat Hatcher of ODU - my advisor at Penn State - for introducing me to this amazing analytical tool!- Pat learned SSNMR from Gary Maciel- who created "chemagnetics" which is the NMR that we have at GL!!!!).

B) a willingness to learn the physics that govern magnetic resonance, basically the quantum mechanics of angular momentum. Although, you really don't need to know quantum mechanics to run a solid state NMR, if you treat it like a "musical" instrument and know what to do, you can extract beautiful data. It is likely that most people who run NMR's focus on how to run them, rather the underlying physics of how they work. I will say that I encourage anyone doing NMR to try and learn as much as they can the underlying physics.

2) Overcoming a bias about what analytical facilities an Earth Sciences department should have. Departments are skittish about instrumental investment- as I agree they should be.

The funny thing is that setting up a Solid state NMR facility is much less expensive than running a ICP-MS facility with associated clean room/prep lab, let alone a laser ablation interface (well over 1 million $).

A really fine solid state NMR at moderate field (meaning not some crazy high field magnet that biomed uses!) will cost one no more than $ 500 K- and that gets you a very fine NMR, one can go lower- but not too much lower. As a bench mark a decent GC-MS will run you ~ $ 80K.

THE GEOPHYSICAL LABORATORY SS-NMR: Our instrument is a three channel Chemagnetics CMX Infinity 300 installed in 1998 and has been used continuously to augment research projects spanning Astrobiology, Organic Geochemistry, Biogeochemistry, Marine chemistry and paleo-oceanography, molecular paleobotany and paleontology, experimental high temperature and high pressure Geochemistry, cosmochemistry and material science. Pretty much any problem that involves solids!

NMR's require high quality RF and fast electronics- these have not changed since 1998 and our instrument can currently do whatever you want. What we have done is recently has had most of our instrument re-built (at the board level) and tuned by Kevin Goehring - the "mind" behind this amazing instrument. Rebuilt amplifiers, transmitters, receivers, and pre-amps. Basically I personally focus on , cooling fans, power supplies, capacitors and (in the case of the 1-H , 19-F High power amp) vacuum tubes. This amazing machine is currently purring like a happy cat and hungry for interesting samples :)

DETAILS: The Solid State NMR laboratory at the Geophysical Laboratory is well equipped to perform a broad range of experiments, the instrumental details are:

  • Super Conducting Magnet: Our NMR employs a 7.05 Tesla (300 MHz 1H) Oxford wide bore magnet with Resonance Research Shim controller. The choice of this field was derived as a compromise whereby we could provide first rate solid state capabilities for 13-C and 15-N (where advantages in sensitivity gained by moving to high field are significantly offset by requiring higher speed MAS, hence complicating cross polarization derived experiments) and yet still provide useful capability for various quadrupolar nuclei, e.g. 17-O, 23-Na, and 27-Al. Now, relative to high field solid state NMR (e.g. fields in excess of 14.1 T) at 7.05 T there remains considerable line broadening associated with higher order quadrupolar effects. However, in the case of many quadrupolar nuclei (e.g. 11-B and 17-O) we have found that in many experiments (e.g. multiple quantum magic angle spinning) at 7 Tesla actually provides considerable information due to the enhanced quadrupolar broadening associated with the lower field (Check out some of the papers by former post Doctoral Fellow Sung Keun Lee for a demonstration of some of these “low” field advantages as well as Katya Klotchko's 11-B study of boron oxides in biogenic carbonates). We also find (as Jaeger and others have shown) that considerable information is also available in the satellite transition sidebands. Of course in the case of quadrupolar nuclei, spectral data acquired at different field strengths is always preferable to a single field; thus we maintain valuable collaborations with other NMR laboratories with higher field strength systems. It is best to collaborate with regional facilities that support ultra high field magnet systems, just as we don't have our own synchrotrons :)
  • RF: The GL CMX infinity is a three channel spectrometer equipped with one high power narrow band amplifier (CMA)[ note this amplifier was completely exchanged with the best-available Chemagnetics amplifier back in 2005 and was recently completely rebuilt for performance in 2017- it is totally tuned for high performance] for 1H and 19F. We also employ two high power broad band (AMT) amplifiers capable of exciting resonances spanning a low frequency limit of 15-N (~ 32 MHz) and a high frequency limit of 31-P: these have been fantastic and we have two of them :). Also included are a pair of PTS frequency synthesizers [Just rebuilt one- will do the other when necessary- it is now in perfect shape :)], a three channel pre-amplifier [rebuilt- and perfect!] and receiver [also rebuilt!], a 400 MHz oscilloscope [also rebuilt!], and a Wavetek sweep generator for probe tuning [soon to be rebuilt-still works fine- fussy when first turned on]. BTW, what I am learning is that these instruments are very reparable- at a reasonable cost relative to initial costs- meaning not cheap but worth it!
  • Probes: We currently have five solid state MAS probes to serve a broad range of scientific inquiries. This provides for considerable versatility. Magic angle spinning speed is controlled automatically with a microprocessor MAS speed controller yielding ± 1 Hz. All are fully operational and I plan to cycle these through re-fab- re-spec in the next 3 years.
    • 1.) 7.5 mm double resonance probe-vespel housing: This probe supports large volume (~ 500 mg capacity) zirconia rotors. The maximum spinning speed is 7.0 KHz. We typically use this probe for 15-N, 29-Si and 31-P (when at low concentration) experiments targeted at elucidating the molecular structure of organic solids and silicate glasses. We also have the “Magic Angle Turning” accessories to allow for stable low frequency spinning for the isolation of isotopic and anisotropic signal via the elegant MAT experiments. This particular experiment helped us under key aspect of meteoritic organic solids, namely the abundance of furan moieties (Cody et al. PNAS 2011) and the possibility that IOM formed post Accretion from formose sugars- the solar system is sweet! (Add MAT Figure from PNAS).

Figure 2 (Right): The 7.5 mm probe is particularly excellent when signal is low and fast MAS is not required. Above we show variable contact time 1H-29Si CPMAS experiments revealing the differential rate of polarization transfer to Si (Q) molecular species with differing numbers of non bridging oxygens (e.g. Cody et al GCA 2004-2005). A short contact times Q2 and Q3 are the most intense at longer contact times Q4 is the most intense peak. The change in cross-polarization dynamics is a complex function of glass composition revealing hidden larger scale structural variation that could not be inferred from the distribution of Qn species alone.

Note: This image was made using MacRMN software written and provided by Phillip Grandinetti

  • 2.) 5.0 mm double resonance probe-vespel housing: This is our workhorse for 13-C analyses. With a maximum spinning speed of 12 KHz we are able to move the spinning sidebands completely outside the spectral window for carbon, affording high quality spectra. We typically employ the variable amplitude cross-polarization MAS experiments. In order to minimize background (from teflon) in single pulse experiments, we have a number of boron nitride inserts that can be used with the zirconia rotors. This probe is also a work horse for 2-H NMR studies (see Wang et al. 2015, and La Losq et al. papers - 2016-2017). These days we usually use this probe for 29-Si from Piston Cylinder experiments where the experimental volumes match the probe volume excellently.

Figure 3 (Right): Variable amplitude 1H-13C Cross Polarization NMR spectra of spruce at three different stages of degradation by the fungal micro-organism Gloephyllum trabeum; a ‘brown rot’ fungus. Note the selective loss of cellulose and hemicellulose indicated by the reduction in intensity of polysaccharide secondary alcohols.

Note: the selective removal of cellulose by fungal organisms is not at all a simple process- in fact is likely that the detailed mechanism(s) remains unknown- and yet! This is likely one of the most significant reactions governing the greatest carbon flux back to the atmosphere - period. The respiration of plant debris back to CO2 by fungal organisms provides one of the biggest fluxes of carbon in the carbon cycle period - and we do not really understand the detailed mechanism. Even less is understood about anaerobic degradation of cellulose- obviously it is not easy- hence the reason why ancient boats are pulled out of the anoxic depths of the black sea. And perhaps the only reason why carbon sequistration in sedimentary basins is possible at all- Coal and Oil Shale.

Begs the question as to why organic preservation only seemed appear in the geologic record (Cambrian forward) after the innovation of polysaccharide biosynthesis.

Figure 4 (Right): 13-C solid state NMR reveals substantial differences in the electronic environments of carbon in cellulose and hemi-cellulose. The shift in frequency of the anomeric carbon (100-105 ppm) is particularly helpful in allowing one to distinguish and quantify the relative distribution of these two important biopolymers in biological materials. 13-C NMR was very useful in a collaboration with Jill Banfield and her group in establishing whether iron mountain bio-films were actually synthesizing cellulose (see Jiao et al 2010) (Proteomics suggested major cellulose synthase activity but was cellulose actually being synthesized?) NMR to the rescue!

13C SSNMR was able to show that disordered cellulose might well be synthesized in these biofilms, but for me this study introduced me to biofilms period- these are super complex systems and I think that NMR will have considerable application here- looking into it!

Actually, it is very cool to compare 13C NMR spectra of pure culture (see below) with a dried biofilm- there are distinct differences.

A SOLID STATE NMR VIEW OF BACTERIA (freeze dried- E. coli). Note that as expected many different carbon functional groups from alipatics at ~ 0 to 40 ppm, C-N (~ 55 ppm) and C-O (~ 75 ppm) then unsaturated carbon (sp2) from 110 up to 180 ppm.

We can fit the fairly complex total E. coli spectrum (BLACK) with a representative protein (Bovine Serum Albumin- BSA)- in RED and a energy storage biopolymer - glycogen - a polysaccharide. in BLUE.

What we see in this beautiful NMR (13C) spectroscopic snap shot is that E. coli is largely composed of ~ 70 % protein, ~ 10 % lipid (membrane + intermediary metabolites), ~ 10 % glycogen and ~ 5 % RNA + DNA (probably mostly ribosome.)

How cool is that? The cool bit is when you compare this spectrum with that I acquired of bacterial biofilms that "age" naturally (see publications) and are much more rich in polysaccharide (EPS) and lipid

( I suspect cellular debris that is not re-utilized- too expensive- thus biofilms can get "fat" too!) Will show this spectrum in short order...

  • 3.) 5.0 mm CRAMPS probe-vespel housing: This probe is dedicated to Combined Rotation Multi-Pulse, CRAMPS, excitation experiments for solid-state proton NMR. This is a great probe, but frankly I have never used it because 1) our fast MAS probe does 1-H NMR really well and 2) CRAMPS is not at all a trivial experiment- I am not sure who is doing it these days. But if you want to do CRAMPS we have totally amazing CRAMPS dedicated probe! with accessories.
  • 4.) 4.0 mm Triple resonance probe-vespel housing: This probe has a maximum MAS frequency of 16 KHz. A given sample can be excited by three frequencies simultaneously (e.g. 15-N, 13-C, with 1-H decoupling). This probe is perfect to support experiments such as TRAPDOR, REDOR, and solid state HETCOR- (XY), the 4mm rotors allow MAS up to 16 KHz but with decent volume providing some advantages to certain samples with MQMAS. I built this NMR for capacity way beyond 99 % what I do, and this probe provides it- This is an amazing probe if you have the need for its capabilities.
  • 5.) 2.5 mm double resonance probe-vespel housing: This is a great probe!!! The maximum MAS speed is 30 KHz (nearly 2,000,000 rpm!)-this makes this probe ideal for solid state 1-H and 19-F. The extremely small rotor size results in a terrific filling factor, thus plenty of RF power delivery to the sample- great for exciting multiple quantum transitions. Consequently we find this probe is ideal for 17-O (enriched) and 27-Al (Nat abund.) MQMAS experiments.

Figure 5 (Right): The 2.5 mm probe is excellent for providing substantial RF power that aids experiments like the mutiple quantum – single quantum (MQ) MAS experiment. The contour plot shown above is an 17-O MQMAS spectrum of albite glass acquired by Dr. Sung Keun Lee (former GL Fellow and now Assistant professor at the Seoul National University) using our 2.5 mm probe. Clearly defined are the two oxygen environments. NOTE: in the "MAS" dimension one observes the Quadrupolar (MAS) powder patterns that records the symmetry of the electric field gradient around the 17-O, which for Si-O-Si is largely due to bond angle.

This is something that I believe is actually not exploited enough- Like Prof. Sung Keun Lee, I believe that we can use SSNMR to understand some deep Geochemical questions and no one but a few of us are doing this. It just takes time! Saddly NMR is not fast! But it is very powerful!

And we spend the time! A lot of time- some experiments take weeks!

Note: This image was made using MacRMN software written and provided by Phillip Grandinetti

Figure XX: 17-O MQMAS 2-dimensional NMR spectra of NAS Glasses as a function of pressure: My crazy analogy is that these spectra are like looking down on a time-sequence of a blob of ice cream melting on the sidewalk :) But Really! - this what high pressure does to NAS (sodium-aluminum-silicate) melts quenched to glasses a various pressures: 17-O Multiple quantum Magic Angle Spinning (MQMAS) NMR 2D spectra of sodium aluminosilicate glasses synthesized at 1 atm, 60 K atm (6 GPa), and 80 K atm (8 GPa). The horizontal projection is the standard MAS projection, the vertical projection provides the "isotropic" spectra. One observes that from 1 atm to 60 K atm the primary densification mechanism is reduction of Si-O-Si bond angle (the isolated peak on the right). From 60 K atm to 80 K atm Si-O-Si bond angle continues to be reduced and four coordinated Al is partially transformed to six coordinated Al (seen as intensity growing in between Si-O-Si and Si-O-Al), signifying a liquid-liquid phase transition between 6 and 8 GPa. Note that peaks corresponding to Si-O-Al and Si-O•Na are largely unperturbed with pressure- implying no role in the densification mechanism. So the pressure induced "melting" involves Si-O and Al coorindation only- all other species are unaffected - How cool is that ? I trust you understand my analogy- nothing is actually melting-there are only glasses- but you are seeing differential responses to deep compressive strain.

These samples were made by Dr. Sung Keun Lee using one of the three multi anvil presses at GL (lab maintained by colleague Yingwei Fei) and analyzed via NMR by me using the W. M. Keck Solid State NMR facility. (Sung Keun is more than capable of running his own MQMAS experiments, but he realized that my "little field" accentuates the 17-O interaction and is helpful- this is one more example of why high field is not always better- best is to have access to many different field strengths. Depending on what you want to know- what a concept :)

Choice of magnetic field is a complex issue that Non-NMR people have difficulty with- meaning- higher field (bigger magnets) makes sense for some experiments but not for others.

Note: This image was made using MacRMN software written and provided by Phillip Grandinetti

  • Computer: Communication with the NMR relies on a Sun Ultra-5 workstation running Spinsight Software. This computer is not particulary special (and it does not have to be) it does nothing more than allow the user to communicate with the CMX on board computers (four in the VME cardcage) and is completely backed up on mirrored hard drives.
  • Variable Temperature: Our NMR is equipped with a variable temperature controller, allowing for high temperature (~200 C) and low temperature (liquid Nitrogen temperature) experiments. This worked at demo, but I have never used this- it works but low T work would require considerable attention (like sleeping bags and timers). I can imagine some applications for low T, but have never felt I had to do this. It is nevertheless possible if you need it :)

Figure Final: And by the way ... NMR is a time machine! REALLY- NMR experiments are performed in the time domain and spectra are recovered through Fourier transform to the frequency domain. For 2D experiments we have to acquire in time - time space. The figure above shows raw data in time-time space (time 2 is horizontal, time 1 is vertical) of a typical MQMAS experiment (this one is 27AL MQMAS of kyanite Al2SiO5). Noise is purple and signal is blue to yellow. The MQMAS experiment aims to have the quadrupolar interaction refocus at set times creating a signal (we call an "echo"), the echo signal propagates forward through time-time space as a slowly decaying signal moving from lower left to upper right (obvious signal). But! there is another "echo" that propagates up and left moving backwards in time and intersecting the y-axis approximately 1/5 up (weaker signal). This weaker echo is moving backwards in time and thus the NMR spectrometer is in fact a time machine. I'll leave it to you to puzzle this one out as to how this could be so. I assure you that this does not violate any laws of physics. Clue: A single quantum coherence (what we measure) is a wave and the directionality of time is defined by us! We are, of course, constrained to move in only one direction of time.

So at Carnegie Science we have two types of time machines: 1) Telescopes that our astronomers use to study light from ancient time- going back billions of years and 2) our NMR that generates events that propagate backwards in time- what a great place to work!

Note: This image was made using MacRMN software written and provided by Phillip Grandinetti