Stan's NMR Blog: 2008 (a) entries
Running INDEX | PREVIOUS page | NEXT page NMR in Stan's LIBRARY | Site Plan & SEARCH | Stan's HUB
Mail me a Query/Comment

See the running index for instructions how to cite/link an entry in this blog

This permanent page contains January - June 2008 entries

June 25, 2008

The largest J's ever measured

A June 24 entry on Glenn Facey's blog reminded me of an old idea of mine which might become a collective project.

It starts with the realization that the largest scalar coupling ever observed so far has the value of 284100 Hz (yes, over 284 kHz!) and occurs between the two equivalent 199Hg nuclides of the di-mercury [Hg-Hg]2+ ion. Of course, as Glenn points out (and as every student should know), when the two nuclides are truly equivalent, the J is not observable. In order to observe it one must break the ion's symmetry, which happens in some a-symmetric complexes of di-mercury with crown ethers. According to the exact pair of ether molecules used, the J between the two mercury nuclides varies quite a lot, so that its value in a "naked" di-mercury ion is still a mystery.

The record-breaking measurements were done by R.Malleier, H.Kopacka, W.Schuh, K.Wurst and P.Peringer [Chem.Commun. 2001, 51-52; doi, PDF] in 2000 and might be actually out-of-date today.

In the introduction to their paper, the authors mention also the largest known one-bond heteronuclear scalar coupling reported by M.Maliarik, K.Berg, J.Glaser, M.Sandström and I.Tóth [Inorg. Chem. 37, 2910 (1998)] for the 195Pt and 205Tl pair in (NC)5Pt-Tl; its value is 71060 Hz.

Another interesting case is the tri-mercury ion [Hg-Hg-Hg]2+ in which the central mercury is different from the other two (A2B spin system) so that the one-bond J's can be observed [R.J.Gillespie, P.Granger, K.R.Morgan and G.J.Schrobilgen, Inorg. Chem. 23, 887 (1984)]. Their value is 139600 Hz, smaller than in di-mercury.

The fact that the J's between covalently bound heavy atoms are so large has to do with relativistic terms in the quantum mechanical expressions for scalar coupling. While with light nuclides the relativistic terms are a tiny correction, in heavy nuclides they become very large and dominant (for more details, read the Review by J.Autschbach [Structure and Bonding 112, 1-43 (2004), doi, PDF)].

The whole thing made me think that it would be fun to try and build a table of record-breaking J's between all kinds of nuclides. For example, what are the largest reported two-bond H-H or C-C or C-H or H-F couplings? My practical experience is not broad enough to tell you. But this summer I will set up an on-line table of record-breakers and every time somebody mails me a new entry, I will cite him/her.
Preliminary contributions are welcome.

June 16, 2008


Group Delay
artifact in actual
Bruker data

  Selected LINKS:  
DSP Related
J.O.Smith III
An IET Article
An IEEE Article
Appl.Radio Labs
A patent

Are Bruker and Jeol NMR data wrong ?

In an entry entitled Why aren't Bruker FIDs time corrected?, Carlos Cobas has shown on his NMR Analysis blog that Bruker NMR data acquired with recent digital receiver hardware exhibit a massive artifact which makes the starting portion of an FID look very funny - certainly not the way it should look according to any NMR textbook. What is even worse, it complicates subsequent handling of such data and makes life more difficult for Companies like Mestrelab specializing in off-line evaluation of NMR spectra. There is the unsettled question of how many initial points of any given FID need to be discarded and whether/how they should be compensated by adding points at the other end.

Bruker acquisition software now sets an acquisition parameter named "group delay" ($GRPDLY). The term is borrowed from electronics, where it comes up in a natural way in connection with, for example, filter theory. However, for low-pass filters such legitimate group delays should not exceed about one to two dwell time intervals [one or two data points, with dw = 2*(1/NyquistFrequency) = 1/SweepWidth], while the values we see in new Bruker and Jeol data are one to two orders of magnitude bigger.

The number of points to discard depends on the operating frequency and spectral width in an apparently unpredictable way. I was told that some people use tabulated, empirical values for the corrections, an approach which is hardware dependent and therefore lacks generality. Bruker specifies that for DSP versions between 20 and 23, the group delay is given by $GRPDLY (as far as I understand, a floating point multiple of the dwell time). Whether this is precise and reliable is not quite clear. Mestrelab has developed an algorithm which uses just the raw FID data array to achieve the same goal and which sometimes leads to better results than the Bruker recipe.

The artifact is not present on Varian instruments but lately it started turning up also on Jeol NMR spectrometers, indicating a common conceptual root (though the numbers of data points to discard are not the same).

I have commented on the artifact, speculating that it looks like a piece of unfinished engineering of the digital filters chain and of the subsequent data sampling and accumulation (all parts of FPGA-based programmable hardware). If this is the case then I have myself run into it, but found a way to correct it at the hardware level. This may sound as a serious j'accuse even though it is not meant as such. Mine is just a hypothesis which can't be proved/disproved without the collaboration of the Companies involved.

A collaboration which, I believe, is due to the NMR community since the last thing we need is a new thread of urban gossip (plus a new acquisition parameter describing just an artifact). Bruker and Jeol should finally react to these blog entries and community rumors and explain to us what is really going on with their FID's. Needless to say, I would be glad to publish these explanations, hoping that they prove me completely wrong (alternatively, a link to their own page(s) would be more than sufficient).


2 July 2008: Aaron Rossini, University of Windsor, Canada

Hi Stan, ... I was reading your post about the "artifact" at the front of Bruker and Jeol NMR FIDs. Is the artifact simply not removed by converting the data from digital to analog? (Bruker's terminology, not mine) This can be accomplished by using the "convdta" command in TopSpin or X-Win. The "convdta" command writes the current data set to a new experiment number, without the annoying digital filter at the front. We usually process our Bruker data offline and working with this converted "raw" data is much easier.
Perhaps I am confused about the artifact that you were speaking about. Aaron Rossini.

3 July 2008: Stan

No Aaron, you are not confused, you are just too tolerant. Anyway, I have sent your comment to Carlos Cobas who has started this thread on his blog and he added:

5 July 2008: Carlos Cobas, Mestrelab Research

Yes, I know this "convdta" routine since many years ago. It is described in Bruker NMR Guide Encyclopedia: "The command convdta converts digitally-filtered Avance acquisition data to analog data format. The purpose of this command is to export data to external NMR processing programs not including the processing tools for digital data. Please note that by this conversion the quality of the baseline of the data may be slightly affected".
However, I do not like the sentence "converts digitally filtered data to analog data". I think this is incorrect and misleading. Basically, as I understand it, the command just performs a circular left shift by the appropriate number of points to remove the group delay data points from the FID. Evidently, there is no actual conversion from digital to analog, since that would not make any sense in this context. They probably adopt this terminology because the resulting FID's become similar to the correct FID's we were used to in the days of analog receivers. In fact, this command has been used very often in the past by people using old MestReC versions in which appropriate processing was not available (see this link).
I don't think that it can be considered a general solution to the real problem, since it is applied as a post-processing command, and the quality of the final spectrum is worse.

6 July 2008: Stan

Hmm ... let me see whether I am getting it right: you excite a spin system and, physically speaking, it responds with an analog FID signal which is fed into a 2-nd generation digital receiver and comes out as a digital FID which, alas, is no good for data evaluation except using Bruker (or Mnova) software. Bruker has a command which converts it into an pseudo-analog FID which is not really analog at all, but resembles a hypothetical digitized true FID you would get had you fed the signal into a 0-th generation receiver with analog filters ... WOW! It is becoming quite a 'Science', I must say!

A curiosity: Varian says that they have patented a way to handle the problem at hardware level. The funny thing is that while my search for the Varian patent failed, it unearthed a Bruker patent which claims to do the trick in a "mathematically exact manner". So how comes that Varian data have no anomalous group delays while Bruker data do?

6 July 2008, Jeetender Chugh, Tata Institute, Mumbai, India)

Hello Stan, ... I would like to say here that in our case, "convdta" command has never worked out. However, we use a figure of 24000 as a first order phase correction value in direct dimension during processing (in Felix or in NmrPipe), which does the job. Although I am not sure whether it deteriorates the quality of the signal. It may add to the ongoing discussion, so thought of mailing the comments. Thanks, Jeet.

7 July 2008: Stan

Jeet, have you tried whether your 240000 degrees first-order correction is still satisfactory when you change the sweep width? I bet it is not. It would be also interesting to know in which particular respect the "convdta never worked out".

8 July 2008: Andrew Craig, BlueGnome, Cambridge, UK

Hello Stan, I have also encountered this issue when working directly on Bruker FID data and after some searching I found this page:
This note gives a description of the cause of the effect and also a heuristic algorithm for correcting the FID. Implementing the procedure they describe worked well for me.
Apologies if you are already aware of this or if you are actually talking about a different effect.
Best regards, Andrew.

7 July 2008: Stan

Thanks Andrew, I had a printed copy of the page you mention but not the actual link to the original. And yes, we are talking about the same problem. However, you say that the note "describes the cause of the effect" which I don't quite see; I would just say that they describe quantitatively how the effect looks (I am sure that its true explanation has to do with unfinished electronic engineering). Another thing which I do not understand in this link is the insistence on "circular left shift". A circular shift inevitably moves the "dirty" data points from the start of the FID to its end and thus gives rise to additional problems (see point 2 of the Protocol at the end of the linked page). Why not a simple left shift with a replication of the last data point? Or maybe the softwares in common use do not support this simple operation?

9 July 2008: Stan

At the EUROMAR meeting I have talked with an old Bruker acquaintance of mine who (i) indirectly confirmed what I have expected, i.e., that Bruker is inclined to insist their digital FIDs are correct while textbooks are wrong and (ii) promised to pass along the Bruker hierarchy my invitation to make a comment (with a pledge that I will not distort a single word).

June 13, 2008

30 Dec 2007
30 Oct 2005
Warren Proctor

The story of Fu Chun Yu, co-discoverer of chemical shifts

In my ongoing quest for biographical data about the discoverers of chemical shifts, I was lucky this year to come across two pieces of hard-to-find information.

The first one was the following e-mail from Aharon Loewenstein, dated January 10, 2008:

Dear Stan, I was invited by Warren Proctor to spend two months (about August-September, 1963) at Varian, Zurich (before their move to Zug). Warren served, at the time, as Director of Varian Branch in Switzerland. I did some research there that was published in '64.
I lived in Kusnacht, a small town near Zurich (along the lake) not far from the residence of Warren and his family. One afternoon I was invited to meet Yu who came for a short visit to Warren. Yu arrived from Bordeaux (where he attended a Conference, AMPERE?) and was on his way back to China where he was a Professor of Physics in the University of Beijing. I remember that Yu complained that his research facilities in Beijing are rather poor. We played table-tennis in Warren's garden (he was a good player). Yu was accompanied by a "Cultural Attache" from the Chinese Embassy in Bern who took him back to Bern in the embassy's car, not too late in the evening. Yu brought nice presents from China to Warren and his family.
I know from Maurice Goldman that Warren spent some time in Denmark (and France) and also that Warren was present at Abragam's 80th birthday ('94?). Maurice may have more information about Warren.
I should add that I had great respect for Warren who had a very impressive personality. I liked him much.
I hope that these lines can add a tiny bit to your Note.
Best regards, Aharon Loewenstein.

This excited me for several reasons. First, it contains the delightful human glimpse of Warren Proctor. Second, it put me on the track of Peking University as a reference for F.C.Yu. And third, for somebody like me who grew up in a communist state (in my case Czechoslovakia) the story of a scientist being closely escorted everywhere by a "cultural atachè" is all too familiar and revealing. I needed nothing more to understand why we have heard so little about F.C.Yu all those years.

Guided by the new knowledge, I have promptly found out that F.C.Yu started his scientific carrier at Peking University (then National University of Peking) in the field of nuclear physics and published a paper with that affiliation in Physical Review [1] already in 1942. His first paper published in the USA appears to be the one with W.G.Proctor in 1949 [ref.3 in the previous entry], with the affiliation to Department of Physics, Stanford University (assisted by the joint program of the AEC and ONR). The last paper he published while in USA [3] was submitted in February 1951. He returned back to his Alma Mater in 1951, less than two years after the People's Republic of China (PRC) was formed in October 1949. He is mentioned in the acknowledgements in a 1985 IEEE paper (Song Z., Yu J., Li R., Yuan Z., IEEE Trans.Nuc.Sci. NS-32, p.1826). Finally, the last paper published in a western Journal [4] which he signed indicates his affiliation with Heavy Ion Physics, Peking University, Beijing, PRC.

The Institute of Heavy Ion Physics at Peking University has nowadays merged into the School of Physics (there is an english Introduction page). I wrote to the Head of the School in March, enquiring about the whereabouts of F.C.Yu and about any papers he might have published in Chinese journals, but did not receive any answer (yet). However, I noticed that Dr. Gang Wu, now at Queen's University (Canada), is a Peking University graduate. I wrote to him and, on April 14, received this reply:

Dear Stan, sorry to reply your e-mail just now. Indeed I graduated from Peking University (but from a different department) where Prof.Yu taught in the Physics Department since his return from the US in 1951. Prof. F. C. Yu passed away on Feb. 12, 2003 at an age of 90. As you may be aware of, Prof. Proctor wrote an article in Encyclopedia of Nuclear Magnetic Resonance (Vol.1, edited by D.M.Grant and R.K.Harris) to describe their discovery of chemical shifts and his friendship with Prof. Yu. I hope this information is useful to you.
Best wishes, Gang.

This, in a certain sense, completes my quest (except for the still unknown destinies of W.C. Dickinson). From the Proctor's article in the Encyclopedia [ref.21 in the previous entry] I have finally learned the full names of Warren George Proctor and Fu Chun Yu, both of whom are now reposing in peace in Denmark and China, respectively.

Papers by Fu Chun Yu, other than those published with W.G.Proctor:
Notes following three stars indicate affiliation.

  • Ma S.T., Yu F.C., Electromagnetic Properties of Nuclei in the Meson Theory,
    Phys.Rev. 62, 118-126 (1942). DOI: 10.1103/PhysRev.62.118.
    *** Department of Physics, National University of Peking, Kunming, China.
  • Alder F., Yu F.C., On the Spin and Magnetic Moment of O17,
    Phys.Rev. 81, 1067 (1951). DOI: 10.1103/PhysRev.81.1067.
    *** Department of Physics, Stanford University, Stanford, California.
  • Alder F., Yu F.C., On the Magnetic Moments of Mg25, Re185, Re187, and Be9,
    Phys.Rev. 82, 105 (1951). DOI: 10.1103/PhysRev.82.105.
    *** Department of Physics, Stanford University, Stanford, California.
  • Yan J., Li Z., Bai C., Yang W.S., Wang Y., Zhao W., Kang Y., Yu F.C., Zhai P., Tang X., Scanning tunneling microscopy investigation of graphite surface damage induced by gold-ion bombardment, J.Appl.Phys. 75, 1390 (1994). DOI: 10.1063/1.356419
    *** Institute of Heavy Ion Physics, Peking University, Beijing, People's Republic of China.
  • Liang J.C., Xu Z., Le X.Y., Zhao W.J., Yu F.C., Liu Z., Chen R.T., Sun X.P., Zheng X.Z., Du Y.R., Zhao M.X., Shen L.F., Gu H.E., Li Y., Su Y.
    Chemical Shifts of 23Na in NaCl Crystal Specimens Implanted with Xenon Ions,
    Chinese Physics Letters 16, 563 (1999). Link to abstract
    *** Institute of Heavy Ion Physics, Peking University, Peking University, Beijing 100871.

June 10, 2008

eDISPA Poster

More comments

Old Swan's eDISPA blues

Old Swan is ironically asking what's going on with the eDISPA method for automatic phase correction which, according to him, can never work and would be easy and funny to demolish. This challenges me directly as an author of eDISPA and, since over a year has passed from its launch, Old Swan and - above all - Mnova users are indeed entitled to some answers about the development in this area.

First of all, the unsubstantiated statement that eDISPA does not work is grossly misleading. In many cases, especially the difficult ones characterized by broad and noisy lines like in in-vivo spectroscopy, it is the only algorithm which works reasonably well - even compared with a manual phase adjustment by a trained operator - and its expected immunity to noise is really formidable. It also works very well on spectra which have lines of comparable intensity distributed across a fair portion of the spectral width.

It is true that beta testing has revealed problems eDISPA has with those spectra of small-molecules in which strong lines are all grouped together in a relatively narrow band while any lines located outside that band are substantially weaker. This problem has been tracked to the fact that the present version of eDISPA is giving too much importance to relative line intensities. There are several obvious ways how to cure this behavior and we plan to explore them as soon as possible. Another problem is a loss of precision in spectra with severely under-digitized lines (a common defect in high-field spectroscopy) but that regards all automatic phasing algorithms.

Our principal problem, however, is that there are too many things to do. Right now, the priority of mine and my friends at Mestrelab is structure verification and elucidation. Automatic phasing, though important and improvable, is considered reasonably covered by several algorithms and therefore not so urgent. Nevertheless, eDISPA is slated for one of the next Mnova releases as an additional phasing option.

What puzzles me in Old Swan's comments is that he tends to be critical of other people's work (not just mine) but he fails to substantiate his opinions by something akin to mathematical algorithm analysis. There is no doubt that anything can be improved (especially a complex algorithm) but that is no reason for raising fake issues where the only laudable action is either to do the analysis or to propose alternatives (and, of course, analyze those as well). After all, a software algorithm can never really be an issue - if it works and people like it, they will end up using it; otherwise, they will be simply using something else. And, in any case, it is not as though I were in any way forcing Old Swan to implement eDISPA in his iNMR package.

June 8, 2008

NMR songs by Greg Crowther et al

Since it is Sunday today, you should visit Greg's science songs page and look up these entries (both the lyrics and the MP3 music):

Any comments or praise are superfluous and INADEQUATE. Thanks, Greg.

May 31, 2008

Related REPORT
from Uni Windsor
FCC Radio

FCC Frequency
Allocation Table

US Frequency
Allocation Chart


Versione Italiana

in chiave scherzosa
di questa storia



Home 2010
(labs are worse)

Magnetic Resonance and Radio Pollution

When it comes to polluting the environment, humanity has no equals. This applies to all kinds of spaces, including the electromagnetic ether. What follows is a contribution and a warning from Vanni Piccinotti, an engineer based in Florence, Italy, who has decades of experience with installing, testing, and servicing NMR and EPR spectrometers. The problem of RF interference has never been trivial and it keeps slowly but inexorably mounting. The spread of urban areas (combined with the tendency to install MR instruments in those same areas) and the mounting maze of wireless technologies penetrating our daily life spell nothing good for the future quality of our NMR signals. With concurrent progress in basic sensitivity of the instruments, time might come when environmental RF noise will be the single major obstacle to further progress. But let us hear what Vanni has to say:


Cross-talk between FM Broadcast Radio Transmitters (88-108 MHz)
and NMR Spectroscopy: A recent experience

Recently I had to install a 400 MHz (9.4 T) NMR Spectrometer. The system worked fine and, using an Indirect Detection Probe , met quickly and effortlessly the specifications. But later on, when the customer installed a 13C direct detection Probe, the S/N ratio turned out to be quite low and, on top of it, the sensitivity was subject to erratic and very large variations from 50:1 to 130:1 (manufacturer's specs give 155:1), without any apparent reason.

Since the nominal 13C observe frequency at 9.4 T is 100,568 MHz, right in the middle of the range of commercial FM broadcasts, I have immediately suspected that the spectrometer was picking up one of those radio stations. In fact, using a cheap FM radio receiver, a strong station was quickly found at 100,60 MHz. At this point, I have connected a simple audio amplifier ending with a loudspeaker to the output BNC of the observe receiver which was there apparently just for this purpose, and all of us were listening to the radio using a 200.000 Euro NMR spectrometer, except that the audio quality was really poor, much worse than from the above-mentioned gadget radio (a shopping mall gift).

The problem is well known from the old times, when the highest field was 2.45 T and the nominal H1 frequency was close to 100 MHz. One of the first Italian NMR spectroscopists to experience it, back in 1974, was Prof. L.Lunazzi at University of Bologna, on his brand new Varian XL100 spectrometer, and the radio station was Radio San Luchino, well known to anybody living in Bologna, which broadcasts from the top of the nearby Saint Luca hill.

The obvious solution is to change the magnetic field, and thus all resonance frequencies, in order to get out of the modulation envelope of the interfering transmitter. But this is not always easy, since the range by which one can move the magnetic field changing just some software parameters is usually limited to a few tens of kHz in the frequency domain. If larger variations are required the poor engineer has to work on the superconducting coils of the magnet, which is a no-trivial job entailing the risk of a total or partial quench.

Bitter experience shows that persuading the involved radio station to change its operating frequency is a time consuming, frustrating, and apparently quite impossible task.

Being well aware of the problem, my preliminary spectrometer checks always include some blank acquisitions taken before running up the magnet so that there is no chance to observe an NMR signal. The resulting dataset should be pure white noise, without significant spikes. This was done also in this particular installation but, as usual, in the days following the energization the magnet drifted a bit, getting closer to the radio station carrier. Furthermore, the usual 13C spectral widths are quite wide which makes things even worse. Murphy's Law has no exceptions!

But we are just at the beginning of my real troubles. Before putting one's hands on the magnet, one should better know how much, in which direction, should the field be moved. I have therefore used a good Spectrum Analyzer (Tektronix model 2710) to check the frequency spectrum around 100 MHz, ready for the worst. And the worst was what I got! The band was filled with FM signals, evenly spaced by 250 kHz and with modulation envelopes as wide as 100 kHz, so that when I got far from one station I started receiving the next one; accounting for folding and aliasing effects, there was no chance! The only somewhat free region was at 100,120 MHz, but this implied proton frequency of 398.100 MHz. So now the spectrometer is no longer a "400"!

Before installing a spectrometer, you better get a Spectrum Analyzer and check for the presence of RF fields in the instrument room, taking care to explore the areas close to the observe frequencies of all the most important nuclei. Don't forget the lock: at 14 T (nominal 1H frequency of 600 MHz) 2H resonates at 92,095 MHz, once again in the FM broadcast band. The lock channel receiver has quite narrow bandpass filters, so hitting a radio is a really bad luck, but it had already happened, resulting in fast lock level variations and totally malfunctioning Gradient Shimming which uses deuterium as observe nucleus!

Needles to say, the extremely high sensitivity of an NMR Spectrometer shows up. The signal from the guilty radio, as observed on the spectrum analyzer inside the spectrometer room, had very low intensity level of about -70 dBm, some microvolt/meter, but that was enough to almost completely hide the quite strong 13C signal from the ASTM sample!

The radio was clearly picked up by the Probe (closing the Preamplifier input with a shielded 50 ohm RF load, all signals disappear) but, quite surprisingly, there is almost no shielding effect attributable to the metal body of the magnet, which is after all an almost completely closed cylinder all around the Probe. Most probably a good deal of the signal leaks in through the Shim Coils which are mounted very close to the Probe and, together with their connection cables to the Console, constitute a quite good antenna.

Too bad the Shim Coils are essential, and effective shielding of the instrument with a Faraday's cage is always difficult and expensive (it may be almost impossible once the spectrometer is installed).

Before concluding, let me venture some additional advice based on my experience:

= Install the spectrometer in the best shielded room you can get; the best choice is once again in the basement, where you have the whole building above the ceiling and its [grounded] foundations all around the rest, done in iron-reinforced concrete, amounting to a good Faraday's cage at no extra cost.

= If possible, avoid top floors. If you can't avoid going upstairs, take a good look out of the window: if you see nearby transmission antennas, get ready for troubles proportional to their dimensions and closeness (to my knowledge, however, mobile telephony antennas cause so far no harm).

= I'm sure that an exchange of experiences and/or suggestions regarding this matter would help a lot to solve many existing installation problems and prevent ones yet to come. Stan's Blog is an ideal location and, needless to say, I will be absolutely glad to cooperate.

Vanni Piccinotti, Firenze, 11 April 2008



1 June 2008: Editor's Note

I can only add that it will be a honor for me to host such a serious and consequential discussion on these pages. This being more an e-zine than a true blog, contributors need to submit entries through e-mail, but all well meant comments and suggestions will be published and properly acknowledged.

9 June 2008: Stan

Electromagnetic interference (EMI) and compatibility (EMC) problems in NMR and MRI are due not only to radio stations but to many other sources. In particular, the following categories of devices can produce nasty artifacts in the spectra:

1) The instrument itself, with all its RF circuitry (self-pollution), generating f-domain spikes.

2) Mains power switches which are not "decoupled" by means of good-quality anti-sparking capacitors. This regards, for example, the room lighting, but also ON/OFF contacts of relays in air conditioners, compressors, next-door centrifuge, etc. The arcs on such contacts are well known to generate high-intensity RF bursts of milliseconds duration with frequencies peaking around 10 MHz but extending up to and beyond 100 MHz. They cause corresponding noise bursts in FID's and thus extra noise in the spectra. It is really unfortunate to have a high-quality overnight accumulation on a low-γ nuclide ruined simply by switching ON the nearby corridor lights in the morning!

A systematic revision and decoupling of all electric power contacts within 20 meters from the instrument (including the vertical direction) can sometimes do miracles to the reproducibility of your S/N ratios, while getting rid of all gas-discharge and halogen lights in the same area (even when out of sight) can help to improve it.

3) High-voltage power lines are also powerful RF noise generators, due to sparking. Ever heard the crackling noises under cross-country power lines on a wet, drizzly day? And ever noticed what it does to your car radio when you drive under them? Well, be warned: that noise is what you will get right in your spectra of low γ nuclides (or in your MR images) if you install your instrument anywhere close-by!

4) Today, however, the dominant source of EMF pollution are computers - plus anything that contains microprocessors. I have encountered the problem for the first time back in early 80's (before the advent of IBM PC's). I and a colleague were mounting an external lock circuit on a Bruker WP80 instrument and there was most of the time (but not always) a terrible noise nearly covering the lock signal. After two days of desperate search for a bad ground contact in our new accessory we have noticed that the noise was there only when somebody was in the next-door lab and, after some more detective work, we have traced it to a Z80-based table-top computer. The RF it radiated got picked up by the body of the Bruker variable temperature controller (especially when switched OFF) and passed straight into the sample area along the sample-temperature thermocouple (always a major entry port for all kinds of external interferences).

Computers are nowadays ubiquitous, run much faster and, notwithstanding regulations, radiate like fireworks. A single PC uses a bewildering range of frequencies. It is not just the main chipset clocks but also all the local busses and communication devices (typically, there are a dozen of them and each has a different clock standard). Worse still, all these busses carry highly modulated data, spreading their spectrum over extremely broad intervals. And do not forget that, from this point of view, a printer, an iPod, or even a digital camera are also computers. Every time a new e-gadget is born (be it fire-wire, wi-fi, RFID, or the next thing), NMR users and manufacturers alike should start sweating cold!

If your S/N ratio is systematically worse than that of a colleague with the same instrument, start looking around your - and your neighbor's - labs. Often the main culprit responsible for 90% of the problem is not the instrument but a single piece of digital electronics humming half forgotten in a closet that nobody has opened since five years.

The same holds for the situation when accumulations do not seem to improve S/N ratios the way they should and even ruin resolution. In this case, you might have RF noise affecting the lock signal which actively generates noise in the magnetic field which, in turn, ruins the accumulation efficiency and post-accumulation signal resolution (for details, click here).

May 27, 2008


NMR Spin-noise

of doped water
at 63.78 MHz.
Hoult, Ginsberg
Spin noise image

of a phantom
at 500 MHz.

Spin-noise radiation and NMR spectra/images without any RF pulses

All NMR textbooks tell us that, in order to measure an NMR signal, we must first "excite" the nuclei by means of a suitable RF pulse. As it often happens, however, textbooks are wrong !!! The spectrum shown below has been obtained acquiring and evaluating just the noise at the output of an NMR probe containing the sample, while the transmitter was physically disconnected!

Spin noise spectrum of a mixture of 80% isopropanol, 10% DMSO and 10% DMSO-d6
10 minutes acquisition at 500 MHz, using a proton cryo-probe.

Left trace: example of a noise sample. Top trace: spin-noise spectrum recovered from a set of such noise samples by means of a special self-correlation algorithm. The Figure has been adapted with permission of the Authors (NM) from the graphic of their ENC-2008 poster [1].

The above Figure comes from a poster signed by a group of Norbert Müller and Alex Jerschow and presented at the last ENC meeting. It was dedicated to several puzzling experimental aspects of spin-noise radiation emitted by samples placed inside tuned circuits. Another poster presented at the same meeting [2] by Tim Field and Alex Bain used a phenomenological model to discuss the relationship between S/N ratio and spectral resolution of spin-noise spectra on one side, and noise acquisition parameters such as sampling rate and number of data points on the other side. Though many aspects of the spin noise phenomenon are not yet well understood, both presentations implicitely demonstrate what might come as a surprise to many of us: that spin noise and its spectral features are easily observable and therefore merit attention. This is futher underlined by a recent paper [4] by Norbert Müller and Alex Jerschow in which they show experimentally that spin noise is all one needs to acquire MRI images !!!

Over a year ago I had the luck to listen to a talk given on this topic by Norbert at the 22nd Valtice NMR meeting [3] in which he has presented, perhaps for the first time ever, a high-resolution spin-noise spectrum of a coupled system similar to the one shown above and discussed some of the associated implications and "mysteries".

One of the most striking implications is the fact that, at least in principle, spin noise radiation is proportional to the square root of the the number n of nuclei present in the coil rather that to n as in "normal" NMR. This means that, while for samples we currently use in NMR the standard approach gives a much better S/N than spin noise, when the sample amount decreases one should eventually hit a break-even point below which spin noise techniques become more sensitive. Other markedly different aspects are the frequency and temperature dependencies. Since spin noise is due to random, incoherent emissions, there is no need for the spin system to be polarized. Hence no Boltzmann dependence, no saturation, and no direct temperature dependence. The electronic noise does vary with temperature, but the spin noise does so only indirectly through variations in spectral density functions (the same ones which determine relaxation times).

However, some of the above statements are a bit uncertain because spin noise (incoherent spontaneous radiation emission?) still presents a number of theoretical challenges and shadow areas, particularly when it comes to bridging the gap between quantum mechanical and classical descriptions of spin phenomena. The doubts are of the same nature as those which afflict our in-depth understanding of the normal FID (coherent spontaneous radiation emission?). These problems date back to the early days of NMR [5,6,11-16] and many physicists feel that they are still not satisfactorily settled. Understanding spin noise is also closely related to the unfortunate concept of radiation damping. I say unfortunate because the term is used somewhat improperly [6] to denote both a spin-system's self-interaction feedback via a pick-up coil and the unrelated (but semantically more appropriate) phenomenon of progressive energy loss due to spontaneous emission.

The existence of spin noise radiation was mentioned by Felix Bloch [15] who also predicted its √n dependence. The history of the first, wildly oscillating estimates of the intensity of spontaneous radiation emission from a spin system and the various contradictory points of view are well reviewed by D.I.Hoult [5]. The first experimental observation of NQR spin noise (working at 4°K and using a SQUID detector) is due to T.Sleator and E.Hahn [9,10], followed by room temperature NMR studies of M.A.McCoy and R.Ernst [8] and M.Guéron and J.L.Leroy [7]. In 2000, given the persistence of deep-seated theoretical doubts [6], D.I.Hould and N.S.Ginsberg have undertaken a formidable experimental study [5] which has resolved some of the reservations, but not all.

Yet it now appears that, through a simple application of relatively straightforward noise self-correlation algorithms, this area has suddenly reached a qualitatively new stage. We now know how to measure noise radiation spectra and it turns out to be easy and affordable (no special equipment is needed). Consequently, we will be no longer groping in the dark regarding noise radiation intensities and their dependence on all kinds of factors (frequency, spin, gamma ratio, temperature, relaxation times, spectral densities, coupling, etc). There is a lot to be done and there are many rather fundamental questions to be answered - but we now have the tools to do it.

Another question is whether spin noise radiation will ever become an important technique in actual applications. For the moment it can't compete with conventional approaches. The sample quantities at which it would become advantageous in terms of sensitivity are too low to allow detection by any of the two techniques. But who knows - sensitivity is being pushed up continuously, so a day may come when we will measure picogram quantities and do so simply by monitoring the RF noise generated by the sample's spins.

Principle references:

  • Nausner M., Schlagnitweit J., Smrecki V., Jerschow A., Müller N.,
    Nuclear Magnetic Spin Noise Spectra,
    Poster presented at the ENC-2008 meeting (Asilomar, March 9-14, 2008).
  • Field T.R., Bain A.D.,
    Spin Noise in NMR: Theory of its Origins and Implications for Experiments,
    Poster presented at the ENC-2008 meeting (Asilomar, March 9-14, 2008).
  • Müller N, Jerschow A.,
    Recent Progress in Spin Noise Spectroscopy,
    Talk presented at the 22nd NMR Valtice meeting (Valtice, Czech Republic, April 15-18, 2007).
  • Müller N, Jerschow A.,
    Nuclear spin noise imaging,
    Proc.Natl.Acad.Sci.U.S A103, 6790-6792 (2006). DOI link. PDF.
  • Hoult D.I., Ginsberg N.S.,
    The quantum origins of the free induction decay and spin noise,
    J.Magn.Reson. 148, 182-199 (2001). DOI link
  • Hoult D.I., Bhakar B., NMR signal reception: Virtual photons and coherent spontaneous emission, Concepts Magn. Reson. 9, 277-297 (1997). DOI link
  • Guéron M., Leroy J.L.,
    NMR of water protons. The detection of their nuclear spin noise, and a simple determination of absolute probe sensitivity based on radiation damping,
    J.Magn.Reson. 85, 209-215 (1989). DOI link
  • McCoy M.A., Ernst R.R.,
    Nuclear spin noise at room temperature,
    Chem.Phys.Lett. 159, 587-593 (1989). DOI link
  • Sleator T., Hahn E.L.,
    Nuclear-spin noise and spontaneous emission,
    Phys.Rev. B36, 1969-1980 (1987). DOI link
  • Sleator T., Hahn E.L.,
    Nuclear-spin noise,
    Phys.Rev.Lett. 55, 1742-1745 (1985). DOI link
  • Macomber James D.,
    How does a crossed-coil NMR spectrometer work?,
    Spectrosc.Lett. 1, 131-137 (1968). I can't get hold of this one. But see this and this and this.
  • Bloembergen N., Pound R.V.,
    Radiation Damping in Magnetic Resonance Experiments,
    Phys.Rev. 95, 8-12 (1954). DOI link
  • Dicke R.H.,
    Coherence in Spontaneous Radiation Processes,
    Phys.Rev. 93, 99-110 (1954). DOI link
  • Hahn E.L.,
    Nuclear induction due to free Larmor precession,
    Phys.Rev. 77, 297-298 (1950). DOI link
  • Bloch F.,
    Nuclear induction,
    Phys.Rev. 70, 460-474 (1946). DOI link
  • Purcell E.M.,
    Spontaneous Emission Probabilities at Radio Frequencies,
    Phys.Rev. 69, 681 (1946). the Am.Phys.Soc.


3 June 1008: Richard Upton, Glaxo Smith-Klein R&D, UK

Dear Stan, ... We have a couple of questions on your recent articles...
For the spin noise spectrum of isopropanol in your recent entry, how much of that signal is actually from genuine spin noise and how much is from the RF pollution of the environment, if that is possible to say? Obviously, your following article prompted this question.
Even theoretically, would spin noise ever reach a point where it was more sensitive than normal techniques, as taken to the limit the least you can have is one nucleus present and the square root of 1 is 1. With this particular factor, is then spin-noise signal equivalent to normal method response? We are not theoreticians, so we may have missed a quite fundamental point here!
Richard J. Upton.

6 June 1008: Stan

Dear Richard, your first question regards primarily the authors of the original spin noise research, to whom I will pass it for comments. What I can add is the following:

In their paper, Hoult and Ginsberg made it clear that they were acutely aware of the danger of contamination by an unwanted, low-level RF leakage into the sample coil area and went to great lengths to make sure that no such thing was going on at any detectable level. Muller and Jerschow, according to what Norbert told me, were also very conscious of the problem and made sure, for example, that not only was the transmitter disconnected from the probe and its input clamped by a 50 Ohm resistor, but all the transmitter hardware was actually kept off during the experiments.

For what regards environmental RF pollution, I imagine that it can be kept under control, though it may require quite a bit of care. After all, the nuclides react to local magnetic fields of any kind and have no way to distinguish whether they are due to local molecular motions or to remote radio stations. Any such interference shows up as a noise when doing standard data acquisition and evaluation, but it might come out as a spectrum when using noise correlation methods. After all, spin-noise spectroscopy can be viewed as a variety of stochastic NMR and MRI with the noise-excitation generator replaced by the intrinsic thermal noise of the sample. In this picture, an external disturbance becomes simply an additional (free :-) noise source.

The second question you raise could heat most theoreticians up to the deflagration point. As I see it, whenever one has an isolated spin system (or just a few of them), one enters the uncharted region of quantum physics of single systems where the statistical Copenhagen convention is not applicable and paradoxes abound. The only rule which seems to hold is that whenever one does any measurement traceable to a single quantum system, one always finds exclusively the values pertinent to its eigenstates and never to any mixed states (this, in particular, regards radiation absorption & emission, but also, for example, magnetic force microscopy where they manipulate single electron spins). But if this were true also for ensembles, then spin 1/2 systems could not exhibit any transversal magnetization component and thus generate a signal in a pick-up coil. The fact that they do so has evidently to do with coherence of the phases of the individual systems, but I doubt whether somebody really knows how to convincingly reconcile the "comfortable" ensemble quantum mechanics with the counter-intuitive single-system quantum physics. Theoreticians are still squabbling over the Einstein-Podolsky-Rosen paradox!

Fortunately, I don't think that when it comes to spin-noise radiation, we need to enter this area. My very crude estimate of the break-even number of nuclides below which spin noise radiation might beat standard methods is somewhere in the region of 1e9 to 1e11 (say 1 microliter of a 10 nano-molar solution). This actually regards single-scans, and since spin-noise NMR spectroscopy does not need any relaxation delays, there may be another order of magnitude in its favor when referring to a fixed accumulation time interval. So pushing our worries down to 1 nuclide smacks of masochism. After all, the √ n rule is statistical and needs a decent-size n to hold.

But you will probably be interested why I put the break-even point around n = 1e10 spins and not n = 1. My reasoning goes like this: Given an ensemble of n spins with magnetic moment m and randomly oriented vectors (this is crude, so forget quantization), the r.m.s. of the resulting, randomly fluctuating vector is m√n (as pointed out by Felix Bloch) and the resulting spin-noise signal should be proportional to that. The classical FID signal is proportional to m.n.β, where β = [1-exp(mB/kT)] is the tiny Boltzmann factor. In both cases, of course, there are other factors but those should be either similar (transition probabilities) or equivalent (coil filling factor and Q). Consequently, the break-even point should occur when, approximately, m.√n = m.n.β which, depending upon the temperature T and the main field B, leads to the numbers I have indicated, give or take a factor of 10.

At least, I hope so :-)

6 June 1008: David Hoult, NRC, Canada

... Yes, we were acutely aware of the risk of any interference and of course all the experiments were conducted in a shielded room with the transmitter not only off, but disconnected. The evidence that there was no sample excitation from external sources having a noise temperature greater than 293 K lies in the facts that 1.) in the absence of the sample, the noise spectrum of the coil and pre-amplifier was as predicted from the RF noise model for the transistor, its noise characteristics having been experimentally measured with hot/cold shielded sources; 2.) the noise figure of the pre-amplifier (0.55 dB) was low so stimulation of the sample by noise current in the RF coil was low; 3). there was good agreement between theory and experiment. These facts put my "comfort level" with the results of a very difficult experiment at about the 95% level.
By the way, a practical point for anyone attempting to reproduce our results at a field greater than 1.5 T: GaAs field effect transistors lose gain and noise performance in high fields unless they are oriented correctly.
With regard to the "break-even" point for spin-noise versus conventional spectroscopy, I concur with your estimate. A simple-minded formula is n = (2kT/hv)^2 ~ 10^10. Finally, to be really nit-picking, one should not talk in this context about spin-noise "radiation". The term radiation implies emission of energy and as I hope our paper shows, this is truly a negligible phenomenon. Rather, it is spin-noise induction ...
David Hoult.

7 June 1008: Stan

Wow, somebody agrees with my numbers! Thanks, David :-) More seriously, let me wrap up how I interpret your papers on the spin radiation-induction impasse:

We can't explain NMR spectra without sharp energy quanta. Yet we know that when it comes to standard NMR phenomena, true radiation of electromagnetic waves is negligible. So the coupling between a spin system and the coil is non-radiative (induction) - but it still involves exchange of energy quanta at well-defined frequencies. If one wanted to model those, one would have to resort to quantum electrodynamics with its virtual photons and their creation & annihilation operators - a feat nobody had yet the courage to carry through.

I believe that Max Planck would have liked this! He started with energy quanta, not with photons, and he was not quite comfortable with the equation energy quantum = photon. Actually, he nurtured quite a few ontological doubts about photons [M.Plank, A Survey of Physics, Methuen & Co. 1925, reprinted as A Survey of Physical Theory by Dover Pub. 1960,1993]. So, who knows, NMR might yet help to resolve this fundamental puzzle. Spin noise radiation (pardon, induction) might be nothing more (nor less) than a sequel to the black-body story.

May 6, 2008

Rheology and NMR

Looking at the technical program of the forthcoming huge XV-th International Congress on Rheology (ICR, August 3-8, Monterey, Califirnia), one might be puzzled by the fact that while the opening lecture by Paul Callaghan is entitled "From molecules to mechanics: nuclear magnetic resonance and rheological insight", NMR does not appear anywhere in the 14 Keynote Lectures and is barely mentioned in just one of the 17 associated Mini Symposia (see the one dedicated to New Experimental Methods). This contradiction merits a few words of explanation.

Rheology is a complex science of utmost importance in an amazing number of technological areas. Actually, the technologists who most frequently use rheological phenomena are housewives: working a dough, whipping a cream and pouring honey from a jar are all processes which require an empirically learned rheological feeling.

NMR has been somewhat slow to enter this field, possibly because of its inherent complexity. Rheological systems are often chemically complicated, have badly resolved NMR spectra and, to study them properly, require a theoretically demanding combination of spectroscopic, relaxometric and diffusometric NMR techniques.

During the last decade, however, the situation has been changing and there are now several groups directly concerned with rheological NMR applications. Of these, two are recognized as leaders in the field: Paul Callaghan's at MacDiarmide Institute (New Zealand) and Claudia Schmidt's at Paderborn University (Germany). A more complete list of groups is available on the dedicated Rheo-NMR web site which contains also a list of publications and other useful information.

Knowing the adaptability and the penetration power of NMR as an investigative tool, I bet that the XVII-th ICR (if not the XVI-th) will reserve to NMR a Mini Symposium of its own.

April 30, 2008

Valtice Castle ...

and its
wine cellar!

Impressions from 23rd NMR Valtice

I am now back from the annual meeting of the Central European NMR Discussion Groups (the 23rd NMR Valtice) held, as always, in the castle built by the Liechtenstein family in Valtice, a tiny town in southern Moravia (Czech Republic).

Compared with last year edition, there were considerably more participants (97 from 14 Countries). Though most of the contributions still centered on applications of NMR spectroscopy to the structure and dynamics of relatively small molecules, presentations regarding NMR of large biomolecules are on the rise (about one third). This is no doubt due to the nearby Czech NCBR (National Center for Biomolecular Research) in Brno with its well equipped NMR Lab. In any case, there was certainly something for every taste, including classical structure-crunching, tautomerism, conformational changes, residual dipolar couplings, chemical shift tensors, quantum calculations, magic angle spinning, diffusion studies, multiplexed phase cycling, saturation transfer difference NMR, pulse sequences optimization, data evaluation, NMR of gold nanoparticles, matabolomic-like analysis of air-pollution aerosols, plain metabolomics, small proteins, DNA blocking and RNA relaxation and dynamics (abstracts of all the presentations are on the conference site).

My subjective fancy was caught by four distinct topics:
- Two presentations by Gerhard Zuckerstätter and by Judit Schlagnitweit from Norbert Müller's group in Linz, Austria, regarding the simple but neat trick of getting more info from equal-time acquisitions which is now known as multiplex phase-cycling,
- A talk by Marc-André Delsuc from CNRS in Montpellier, France, about the fractal dimensionanlity of large molecules and the way it can be determined by means of self-diffusion studies,
- A presentation by Jan Sýkora (no relation to me) from the ICPF NMR Group in Prague about NMR analysis of aerosol particles (like in air pollution) which I perceived as a case of methodology transfer from metabolomics to environmental chemistry, and
- A presentation by Lothar Brecker et al from Austria (Uni Vienna and Graz), about a biochemical application of saturation-transfer difference NMR. Inevitably, the simple principle of STD-NMR made me wonder why have we not thought about it decades ago?

I have presented a talk co-signed by Carlos Cobas of Mestrelab Research entitled NMR Spectra Processing, Verification and Elucidation: Challenges and Current Progress (see the abstract and slides).

The meeting was enlightened by two special moments. One was a brief, impromptu presentation by Vladimír Zeman of his memoir Vzpomínky na NMR v Brnênské Tesle (Recollections of NMR in Tesla Brno, so far only in Czech language) which was recently published in Stan's Library. Vlád'a was present and said a few words about the times when a group of NMR enthusists within the state-owned behemoth Tesla was producing NMR spectrometers. They were the only such group in the ex East Block and they produced about 500 instruments! Since I am servicing perhaps the very last operative Tesla spectrometer (installed in Bari, Italy), I have prepared a few slides to document that, 15 years after the demise of Tesla, it is still running and producing useful data.

The other moment consisted of the assignment of a prize for the best presentation by a young Author. The prize comittee selected two equally meritable presentations: one by Katerina Maliñáková on 13C Chemical Shift Tesnsors in Regioisomers of N-Substituted Adenins, and the other by Pavel Kaderávek on 13C Relaxation Study of RNA Dynamics: Application to UUCG Hairpin Loop (both from the NCBR at Masaryk University, Brno). The Prize involved two bottles of excellent local wine, offered by NMRtec, so it was easy to split it in two. By the way, NMRtec is a young French NMR service company; they started with NMR software and are now moving into high-quality services and consulting, a market which is still little developed in Europe, especially compared with the USA.

The meeting's social program included an evening at the Czech National Wine Center, located in the wine cellars of the Valtice castle (offered by Scientific Instruments Brno) and a guided tour to the nearby Mikulov with its complex and interesting history.

If I were to characterize this meeting by a single word, I would say pleasant. It is anything but pompous, it has a reasonable size, and it is both interesting and relaxed. Thanks for that are due to the organizers from Masaryk University and Pliva - Lachema (both in Brno), and from the Czech Society of Industrial Chemistry.

April 27, 2008

home page

Invitation to an EU-NMR Workshop

Interested in biochemical NMR and in the EU-NMR network of European large-scale NMR facilities? If so, you should register for the EU-NMR Workshop which will be held afternoon on July 3 in Athens, Greece, as an appendix to the joint FEBS-IUBMB Congress.

Not only is the registration free, but the first 50 registrants will get their dinner and one extra hotel night paid by the EU!

April 17, 2008

Eiichi Fukushima
(click to enlarge)

NMR and Antarctica

Since philosophy tells us that all things are interconnected, there should be also a connection between NMR and the Antarctic continent. In fact, if you look for "NMR Antarctica" in Google, it comes up with over 70000 hits (of which 0.1% are pertinent) and thus puts philosophers off the hook. However, most of these hits refer to molecules and tissues from cold-adapted species of bacteria, algae and fish living in Antarctica, but studied in cozy NMR labs elsewhere around the world. When it comes to actually bringing NMR to Antarctica, I know about just two sets of initiatives.

One is that of Paul Callaghan (MacDiarmid Institute), Craig Eccles (Magritek), Robin Dykstra (IFS at Massey University) and other New Zealanders who love to spend their summer holidays at McMurdo Sound doing all kinds of NMR experiments on sea ice and, in general, enjoying the ultra low level of ultra low field (ULF) perturbations characteristic of the area. They did it in 1994, 1995, 1997, 2002, and again in 2006. At least that's what they confess but I have a feeling that they went there a few more times 'privately', just for the fun. Here are a few selected references (1),(2),(3),(4), a pertinent pdf, and Google results searching the Magritek site for 'Antarctica'.

The other initiative is that of Eiichi Fukushima (photo on the left) but in this case the connection is not so direct since the NMR he brought to Antarctica was all in his head. I am sure that all readers of this blog know the immensely popular book Experimental Pulse NMR: A Nuts and Bolts Approach he and Stephen B.W.Roeder wrote in 1981. Few know, however, that Eiichi was a member of the very first team which in 1966 conquered the highest peak of Antarctica, Mount Vinson (4892 m) - and that he repeated the feat once again 40 years later! Have a look at the amazing photo of Mount Epperly he has taken during the 2006 expedition, the text which accompanies it, and also at ABQMR, his recently established NMR company. I wish I had his stamina!

Maybe the trick is going first to Antarctica and soaking up some of the ULF silence!

April 4, 2008

The best educational MRI site ever

Sometimes it just happens - in a galaxy of stars one goes supernova and dwarfs the others.

This is what happened among all the MRI sites carrying educational content (including the little in that direction I have on these pages). The MRI Step by Step course by Denis Hoa and Antoine Micheau on the e-MRI site is a beautiful example of what internet education could be if only the world were perfect!

This is what the Authors say about the course and the site: MRI Step by Step presents an interactive course about the basic principles of MRI. This course is intended for students, MR technologists and radiology residents. Its content is highly interactive, with many animations, experiments, and quizzes so you can have fun while learning MRI physics! This site is free of charge and no registration is required.

All I can add is that it is all 100% true and that I am sorry it took me over a year to discover this treasure. e-MRI is a subdomain of Campus Medica (Montpellier, France) which is a sub-domain of the learning management system Dokeos. Thanks to all.

COMMENT added on September 2011:

The e-MRI courses website is still active but it changed affiliation and URL.
Use this updated link.

April 2, 2008

NMR Wikis

I am a bit uneasy about the NMR Wiki which Evgeny Fadeev is setting up. The feeling has nothing to do with Evgeny, for he is enthusiastic, young, and very earnest about the venture and there is no doubt that he should go ahead. It has also nothing to do with resources since, like the legendary Orbitsville (Bob Shaw), cyberspace is all but inexhaustible and the memory capacity of its servers grows much faster than the contents stored on them.

I am uneasy because, while I do not want to undercut Evgeny in any way, I am not sure whether wikis can really work. From what I know about them (which is not much), they are public, cooperative projects run by automatic software's under only limited human supervision. This raises some doubts about these points:

1) Irresponsible input from jokers, vandals, and embittered vendicators (this makes a human moderator mandatory).
2) Subjective bias. Even the Wikipedia, which is very particular about its Authors, presents many MR-related items in a way which I am not quite happy about. On the other hand, if I wrote them, many readers would certainly object to my points of view. This is irrelevant or even welcome on a personal site, but it requires a sharp Editor in a cooperative project.
3) Availability of quality input. It is difficult to find knowledgeable would-be contributors because nobody has the time to do it just for the fun. For example, when I finally write something, I prefer to publish it in a Journal or on my personal site.

I will follow Evgeny's project with curiosity in order to find out whether my doubts are real, hoping that they turn out to be imaginary. In the meantime, I have instituted a list of Wikis in my own NMR directories. And I recommend that you visit NmrWiki and possibly contribute to it. You will already find there a number of interesting items, such as useful links, manuals, an online utility to draw pulse sequences, and much more.

March 31, 2008


Grab your copy of Nature Milestones in Spin

The Nature Publishing Group, thanks to a sponsorship by Organic Spintronics has made us all a unique free gift - a package of educational materials pertinent to all aspects of spin.

It includes well written, chronologically ordered articles describing the milestones in our understanding of spin as a phenomenon, a number of historic articles starting with Peter Zeeman's 1897 paper, a selection of spin-related articles published in Nature, and even two podcasts, one by the Nobel prize winners Frank Wilczek and Richard Ernst discussing spin-world marvels, and the other by Hideo Ohno and David Awschalom discussing the recent conquests of spintronics.

But hurry up, since the free bonanza has a deadline (August 31)

About the sponsor: Organic Spintronics is an Italian high-tech Company based in Bologna and a spin-off from the Italian National Research Council (CNR). If you are not quite sure what is spintronics, please visit their site. And take into account that one of the many modern applications which combine spin with electronics is the giant magnetoresistance effect for which Albert Fert and Peter Grünberg were awarded the 2007 Nobel Prize in Physics. What is it good for? Well, ever heard about hard disks ???

March 29, 2008

Richard Ernst
during a lunch break
(photo Roberto Gil)

2nd Edition; by
Malcolm H. Levitt

Kazuyuki Takeda's
OpenCore NMR

...and a 7T spectrum
of l-alanin it

Monterey Bay

Glimpses of the 49th ENC

First, let me stress that this 'report' is personal and subjective. If you do not agree with me on anything, let me know your own point of view and I will publish it.

As meetings go, Experimental NMR Conference is quite large - this year's program book listed almost 1100 pre-registered participants and there were some more who, like me, registered on-site. Yet it manages to convey the impression of being cozy and human-sized, thanks to the lucky synergy between the organizers' skills and the fractal, nature-friendly environment of the Asilomar State Park with its historic wooden lodges hidden among the pine trees. Where else can one walk out of one of the two unobtrusive halls crammed with a thousand attendees and, within 200 meters, encounter a group of free-roaming deer or, heading the other way, observe a group of fishing cormorants?

Given the size of the meeting, some of its Sessions had to be parallel, something that I dislike since, invariably, many talks among those which interest me most are held at the same time and I can't attend them all. This time it happened again, though the parallelism was kept to a minimum and there was every day a plenary Session before the program split into two parallel ones.

Since ENC is the most important North-American NMR meeting with nearly half a century long tradition, it is the natural forum for presenting new ideas, discoveries and phenomena. This time, however, I have spotted few surprises and those which came close were not quite in the maistream (let me return to it at the end). But magnetic resonance was running fast during the last decade and the fresh approaches and novel phenomena yet to be fully explored are so many that to expect more every year would not be reasonable. As far as I can tell, most of the many ideas which turned up during the last five years were generously represented in the talks and dealt-with in many of the more than 500 posters. The feeling of novelty and first-hand exposition was thus most pervasive.

The meeting comprised:
- The opening address by Angela M.Gronenborn, the presentation by Richard Ernst of the Gunther Laukien Prize awarded to Malcolm H.Levitt, and the respective Malcolm's lecture on Symmetry in NMR which touched also upon the latest Malcolm's hobby, the long-lived spin states.
- Four (!) Sessions dedicated to solid-state NMR (1 plenary)
- Two Sessions on liquid state NMR (1 plenary), dominated by proteins and DNA
- Two Sessions on relaxation and dynamics, also dominated by bio-macromolecules
- One Session each on In-vivo and in-cell NMR, MRI techniques, Instrumentation, DNP and sensitivity enhancement, Data acquisition and the respective Spectral processing, Small molecule techniques, Metabolomics, Interaction and binding (bio-systems), NMR in very big systems (bio-systems), Exotica and new concepts, and NMR Education in Principally Undergraduate Institutions.
- Two tutorials on Paramagnetic NMR and on Hetero-decoupling in MAS NMR
- One workshop on Sparse data acquisition in multi-dimensional NMR
- A side meetings of AMMRL (Association of Managers of MR Laboratories)

Since I am not competent to judge the ubiquitous bio-you_name_it applications, let me comment primarily on contributions to NMR physics, technology, and software. These I have divided into the following categories:

Advances in solid-state NMR
As testified by the number of Sessions (and also the number of past and planned dedicated solid-state NMR meetings), high-resolution multi-nuclear solid state NMR has reached a maturity where it can successfully compete with liquid state NMR spectroscopy. It took over 40 years of efforts, but finally we are there and there is an ever increasing area of applications which would be impossible in liquid state. There are still many things to refine (like solid-state decoupling techniques) and/or properly exploit in practice (like the relatively new double-rotation probes which extend solid-state high resolution to high-spin nuclei). All these aspects were discussed and some significant advances were shown.
Sensitivity enhancement through spin ensemble hyper-polarization
Conceptually, today's techniques of producing spin ensembles whose polarization exceeds that of thermal equilibrium (hyper-polarization) are heirs of the decades old CIDNP (chemically induced dynamic nuclear polarization) and NMR studies of chemical kinetics employing para-hydrogen. We now witness a renewed interest and an intense search for novel ways of carrying out each of the three phases of such experiments: ex-situ or in-situ hyperpolarization generation (plain high-field and/or low temperature pre-saturation, electron-nuclear or nuclear-nuclear polarization transfer, optical pumping, long-lived spin states), delivery (mechanical transfer, RF-assisted cross-polarization, spin diffusion, chemical exchange), and exploitation (high sensitivity 1D spectra, single-shot 2D spectra, 13C imaging, etc). Judging from this meeting, this whole field is now in a frenetic turmoil. Progress, though laborious, appears to be steady and studded with formidable achievements (like 1D and 2D spectra that could never have been obtained otherwise) and emerging systematic applications (like the use of hyper-polarized inert gases to study air flow in lungs).
Sparse data acquisition and evaluation strategies for multi-dimensional NMR
n-Dimensional NMR methods are notoriously time consuming. For n > 2 this is true even when single-scan S/N ratio is very good, due to the enormous number of points needed to uniformly cover the whole nt-domain space. Another - but closely related - problem is the dismally bad digital resolution in indirect dimensions. It has been shown that these obstacles can be alleviated to some extent by exploiting the sparse character of nD-spectra. It is in fact not uncommon that only a portion (say 1/3) of a 1D spectrum contains a non-zero signal, in which case the corresponding 3D spectrum, for example, is non-zero only when when all three f-domain coordinates fall inside a non-zero area of the 1D spectrum, which boils down to less than 4% of the 3t-domain points (1/3)^3. Though this example is grossly over-simplified (usually one must consider more than one 1D spectrum), hopefully it conveys the idea. There therefore arises the question of whether such an a-priori known sparseness of the nD-spectra can't be exploited to reduce the number of required data samples in the nt-domain. The same question has been extensively studied in MRI and the ideas formed there, properly modified and expanded, are presently being transfered into nD spectroscopy. Actually, the problem is exquisitely mathematical rather than physical (much less a chemical or medical one). As far as I know, there are yet no rigorous mathematical theorems regarding its globally optimal solution(s). Consequently, many different approaches are being explored, with acquisition-time saving factors ranging from 2 to 5.
Any sparse-sampling (or compressed-sampling) strategy is indissolubly associated with a corresponding data evaluation method to carry out the mapping from the nt-domain to the nf-domain. The ENC Conference dedicated a Workshop to the various sparse data-acquisition and evaluation approaches (pseudo-random sampling, reduced dimensionality, projection reconstruction, G-matrix FT, Bayesian and maximum-entropy reconstruction, multiway decomposition, nonuniform FT, filter diagonalization, covariance NMR). Moreover, the Session on Spectral Processing was also entirely dedicated to this topic. Again, progress appears to be steady but laborious.
Improvements in front-end hardware
Despite decades of efforts, the classical induction coil has still a few trumps up its sleeves. Thus advances in micro-printing and nano-technologies are pushing viable micro-coil volumes down to the few nano-liters range (this will likely combine with inductive-coupling techniques whose hidden potential has been recently brought to attention by the appearance of the Sakellariou MACS rotor). Completely new micro-designs are also emerging, based either on strip-line waveguides (magnetic component detection) or on in-plane resonators (electric component detection). These, apart from reducing the required sample size, are likely to find their way into stop-flow HPLC-NMR and other hyphenated techniques.
Sincerely, I have expected to find much more happening in the "probes" category. I am convinced that there is still a considerable leeway in terms of both special designs and sensitivity, and a factor of 10 in S/N ratio over the next decay or two - on top of cryocooling - does not appear unreasonable to me. Yet I have noticed little in the way of such logical "next steps" like multi-coil designs, incorporation of complete on-chip digital receivers inside the probes, use of high Tc semiconductors, optical I/O coupling, etc.
Novel signal detection principles
Today, inductive coils and resonators are by far not the only way to detect NMR signals. At very low fields, SQUIDS are now the most sensitive, well established method, boosting sensitivity by orders of magnitude and opening new application areas.
Another detection alternative is magnetic force microscopy (MFM) known for its enormous sensitivity sufficient to detect a single electron spin or a clot of about 50 nuclear spins. I have particularly enjoyed the talk by B.Meier on the use of MFM to directly follow spin-difusion in a solid. I was already aware of this recent work (Phys.Rev.Letters 99, 227603, 2007) by K.Eberhardt's group at ETH (Zürich) but is was nice to hear the details about this first-ever achievement. In addition, The MFM group of IBM (C.Degen et al) have presented a nice talk about force-detected MRI with 10 nm resolution - a huge step beyond the few tens of micro-meters limit of conventional MRI and another demo that MFM is reaching maturity.
Still another detection principle is the spin-induced optical plane rotation (provisorily, I call it the Savukov-Romalis effect; see Nature 442,1021,2006). There was no talk dedicated to this phenomenon, but there were several most interesting posters (look up the names of I.Savukov, M.Romalis, and A.Pines) based on it. When somebody knowledgeable talks about S/N of 3:1 for about 1 μl of 0.1 mM solution at zero field after a pre-polarization at only 1 T (poster W-Th092), I think that we should all listen.
Last, let me mention advances in the excitation-less spin-noise spectroscopy associated with the names of N.Müller, A.Jerschow, D.Hoult, A.Bain, etc. There was no talk about this topic, but there were two very interesting posters (M-T 082 and M-T 087).
Al these marvelous detection techniques and spin phenomena were relegated to the single Session on Exotica and New Concepts. Exotica indeed! Shame on you, organizers! Well, I am kidding of course; it is just that in my physicist's heart this section overshadowed some others. I plan to write more about these phenomena as soon as I can.

I do not want to talk here about the posters I and my collaborators have presented (see the previous entry). They attracted even more attention than we have hoped (warm thanks to all those who dropped by), but they regarded 'just' conventional spectral data processing.

Among other things that attracted my attention was the talk by K.Takeda about his home-made, FPGA-based NMR spectrometer (see also OpenCore NMR). Considering the nearly a year old entry I have written about this topic, I cannot but applaud (I am actually in the middle of a similar, though more modest, project). Maybe we are approaching the times when managers setting up a new NMR lab will first contact a next-door hacker rather than Bruker, Varian, or Jeol. Believe me, it is feasible - and all digital parts of a whole NMR spectrometer do fit into a $50 Cyclone III chip (or an equivalent)!

What impressed me as a truly new breakthrough meriting to be singled out, was the presentation by Vadim Zotev at al (Los Alamos National Laboratory) entitled Microtesla MRI of the Human Brain with Simultaneous MEG. They record usable MRI images at 46 μT (using SQUID detectors) after a 1s pre-polarization at mere 30 mT and, on top of it, combine their MRI apparatus with magneto-encephalography (MEG) using shared detectors. This is not only a top achievement in ultra-low-field MRI (ULF-MRI), but also the first time we hear about a hyphenated MRI, namely MRI-MEG!

To conclude, I have enjoyed this meeting a lot and I can't wait coming back for the next year's 50-th edition.

March 06, 2008

Peak-Picking using

NMR spectra processing at 49th ENC

This entry is somewhat personal. I am leaving for the 49th ENC meeting At Asilomar, Pacific Grove (CA) and, this being the first time I am going to participate at an ENC, I am quite excited. It will be fun to write about it as soon as I return.

I and my friends from Mestrelab Research have just finished three posters we are going to present there. They describe some novel data processing algorithm we have developed - the Resolution Booster, the J-Correlator and the Bayesian DOSY and ROSY Transform. I have put the posters on this site (click the links) so, if you wish, you can already have a look at them and then come and discuss them with us during the ENC poster sessions. We will be also available during two breakfast meetings at Forest Lodge Suite on Tuesday and Thursday morning (8 AM), ready to discuss these and other data processing techniques we are currently developing as stepping stones on the road towards automatic molecular structure verification and elucidation.

As for me, I will be glad to discuss any kind of NMR at any time. There is so much to learn all the time and discussions are probably the best way to do it.

February 29, 2008


Spin: 0
  4.00260325 au


Spin: 1/2
  3.01602929 au
  1.38 ppm
Gyromagnetic ratio
  -32.436 MHz/T

Helium scarcity: can it affect NMR and MRI communities?

During the last 5 years helium prices kept growing and there were numerous worried reports [1,2,3,4,5,6,7,8,9] about its impending scarcity. So the question is: should we worry about our supercons going dry and quenching ? Before telling you what I think about it and why I do so, let me review a few facts about helium and its technology.

Helium (He) is the second most abundant element in the Universe, just after hydrogen. Despite this, it is very rare in the Earth atmosphere (about 5.2 ppm) because its light atoms achieve velocities large enough to escape Earth gravity. While Earth is leaking helium into space, it also produces it as a side-product of the natural decay of its radioactive elements (every alpha particle eventually becomes a helium atom). This amounts to a "production" of about 4 liters of helium per cubic kilometer of the Earth crust every year and accounts for the fact that all porous rocks, underground cavities and, most importantly, natural gas, contain considerable amounts of helium. Unfortunately, only an infinitesimal fraction of such fresh production becomes economically accessible. The helium we consume in ever increasing quantities is recovered from natural gas (He content: 0.5 to 7%). It is essentially "fossil" in the sense that most of its atoms were formed billions of years ago and, as such, it should be considered a non-renewable resource. At least until we will be able to tap the giant planets like Jupiter and Saturn whose atmospheres are full of the stuff ...

Here are some of the reasons why helium is so special (there are many more, of course):

  • It is the material with the lowest boiling temperature (4.214 K). Hence its use as a portable cryogen which we are best acquainted with (20% of total consumption).
  • It is chemically absolutely inert (there are no known compounds of helium). Consequently, it is non-toxic which, combined with its low solubility in water (about 8 ml/l), accounts for its use by divers in breathing mixtures (unlike nitrogen, it does not cause embolic crisis). Its chemical inertness makes it also non-flammable and non-corrosive, features which make it ideal for purging, protecting or pressurizing industrial vessels, chemical and nuclear reactors, rockets (both with and without fuel), etc. Every launch of the space shuttle, for example, disperses in the atmosphere over 100'000 cubic meters of helium.
  • It is very easy to detect by portable spectroscopic instruments. Together with its inherent safety and its small atoms, this makes it ideal for closed-systems leak detection.
  • It has very high thermal conductivity. This accounts for its use as a safe cooling medium, such as in nuclear reactors where it has the additional advantage that its ultra-stable nuclides do not get activated by thermal neutrons.
  • Its atoms move very fast, accounting for very high speed of sound in helium. This property, matched only by the lighter but unsafe hydrogen, makes it ideal in high-speed gas turbines (such as in NMR-MAS) since no gas can impact a higher speed to a rotor perimeter than the average speed of its molecules (the speed-of-sound barrier). For the same reason it is used also for special experiments in wind tunnels.
  • In its gaseous state, it is extremely light, much lighter than air. This accounts for its use in all kinds of lighter-than-air devices, including party balloons. The USA Federal Helium Reserve, for example, was originally set up in the 20's because of the expected strategic importance of helium in air transport, a forecast which did not come true.
  • It has very low dielectric rigidity. Don't let it envelope your NMR coil while pulsing since it would make it spark! On the other hand, this property facilitates welding (18% of current use) where it also renders the arc hotter and provides an inert protective atmosphere.
  • It is the only material which remains liquid at absolute zero temperature. To make it solid (at 0.96 K), one must apply a pressure of over 1600 kPa. Admittedly, so far, only physicists appreciate this feature (and also the superfluidity of 4He below 2.174 K).

A few decades ago, when there were few superconductive magnets for NMR spectroscopy and no MRI, and when welding was less sophisticated, the consumption of helium was a fraction of what it is today. One of its strategic uses was expected to be in super-cooled super-conductive super-computers (another erroneous forecast) and its major producer was the USA, with production in Texas, Kansas and Oklahoma and a large Federal Helium Reserve in Amarillo, Texas. Some more was produced by the Soviet Union and Poland for the internal needs of the communist block.

Some misinformed journalists keep on writing that Amarillo is still the helium capital of the world, but nothing could be more wrong today. Provided it ever was true, considering that the largest US helium supplier is BOC, a Company of British origin. Today, BOC is part of the Linde Group (a behemoth of German origin), the US Federal Helium Program is being phased out, competition from the French Air Liquid is fierce, and the world production relies on refineries in Algeria (Linde) and Qatar (Air Liquid). It was a delay in the construction and start-up of the latter plants which caused the world shortage - they were supposed to be operative five years ago, but they are not up to full capacity even today.

So the helium market - both production and consumption - is now totally global. There was a temporary shortage due to coincident technical problems in the production centers in Algeria and Qatar but that should be now over. However, helium supply is closely linked to that of natural gas and will last only as long as the latter, while helium demand is on a steep rise. Which means that you do not need to worry about quenching your magnet (yet), but you should also not expect to see helium prices go down (ever).

To keep an average NMR spectrometer (400 MHz) running takes today four refills a year, each amounting to 60 liters of liquid helium (including a security margin and losses due to manipulation). In Italy, liquid helium is presently sold to end customers at 8 - 16 euros per liter, depending on how important and/or gullible is the customer. This amounts to a yearly expense of about 2000 euros which is a fraction of the total operating cost of the instrument. In medical MRI, though the liquid helium needs of a whole body magnet are several times higher, the proportion between the helium cost and the running expenses of a whole MRI facility is even smaller than for an NMR spectroscopy lab. In the USA, these ratios should be still smaller since, I am sure, liquid helium is substantially less costly over there. Which, paradoxically, is a problem since the cost of liquid helium for the MR crowd may be too low to stimulate recycling.

Recycling is very easy using balloons attached to the helium dewar vents. Wholesale prices of liquid helium are of course much lower than the end user prices - well below 3 euros per liter. Over two thirds of that goes for the stuff itself, while liquefaction accounts for less than a third. Which means that an end-user can expect to get back about one euro per liquid-liter-equivalent of recycled helium - not much, but enough to stimulate recycling at large facilities like particle accelerators with hundreds of superconducting magnets [10]. But it is insufficient to make care small consumers like NMR spectroscopists so perhaps, considering that helium is non-renewable, its recycling should be made compulsory.

What about the [not so] distant future? I see two trends. One is magnets with actively cooled liquid helium tanks and drastically reduced consumption (for example, one refill every five years). The other one is cryogen-free magnets (or at least helium-free ones) which are becoming a reality. I expect cryogen-free superconducting magnet technology to make MR of all types completely helium-independent in a matter of 20 to 40 years. But this is another topic ...

February 13, 2008

Charles P.Slichter


Field-Cycling NMR and History of Superconductivity

In a recent entry I have coined the term MR splinter events represented by sporadic but important presentations by key scientists. An event of this kind will be the invited talk by Charles P.Slichter, to be delivered at the fast approaching 2008 APS March Meeting (American Physics Society, March 11-14, New Orleans, Louisiana). The presentation, entitled NMR and the BCS Theory, will review an episode in the history of NMR which contributed in a decisive way to our understanding of the phenomenon of superconductivity and helped to verify the Nobel winning Bardeen-Cooper-Schrieffer (BCS) theory.

Today, everybody knows that superconductivity is important in NMR and MRI because it permits the construction of high-field magnets which are an essential part of our instruments. With all the present accent on chemical and medical applications, however, not all MR users know that NMR played an essential role in elucidating superconductivity as a phenomenon. The historic Hebel-Slichter-Redfield experiment, and the Hebel-Slichter effect which it permits to measure, helped to put superconductivity theory on a correct track and, even today, it still plays an important role in the study and development of superconducting materials. The experiment itself is probably the first published case of fast-field-cycling NMR relaxometry (FFC-NMR) with an actively switched magnet.

The key papers in this area were published by C.P.Slichter and his student L.C.Hebel back in late 50's (Phys.Rev. 107, p.901,1957; 113, p.1504, 1959). Preparing this entry, I have read them again - and was again struck by the daring simplicity of the experimental set-up, and the elegance of the physical insights which transcend the ageing formulae (even then, the authors themselves were not completely comfortable with the math, but pretty sure about the essence of what they saw). Having refreshed my memory, I have attempted to sum up the experiment in an article for those who are not acquainted with it.

February 2, 2008


Raman, ?RS

Hyphenated NMR: 
Hypernated NMR: 
A forthcoming 
NMR Hyphenation

A past 
NMR Hyphenation

Hyphenated and Hypernated NMR

A reader is asking what is hyphenated and, especially, hypernated NMR.
Let me attempt a brief answer:

There exists today a whole array of techniques suitable for molecular analysis (see the left column). When one faces the task of determining the molecular structure of an unknown compound, it is therefore logical to combine as many of them as possible. Assuming that one has access to the respective instruments, this is in some cases elementary (for example, Raman spectra can be acquired in the same sample tube used for NMR) and sometimes not. Though most of the techniques are non-destructive, there are often practical compatibility problems (such as those due to sample and container shapes, solvents, etc.) that need to be solved to make such combinations practical. When a combined solution involves a special technical device and/or methodology, it is natural to denote it by joining the acronyms of the primary techniques (example: UV-NMR).

In this context, the condition that an analytical technique should be non-destructive is not as stringent as it seems. Some techniques, though destructive, are so sensitive that the amount of sample they destroy may be negligible. A typical example of this kind is mass spectroscopy (MS). Moreover, if the very last one in a chain of analytical techniques is destructive, it does not interfere with the others and its use may be tolerable.

All this is still too elementary to make much fuss about it. A qualitative jump is achieved by bringing into the picture also some of the many separation techniques developed during the last fifty years (see the left column) and combine them with the various analytical methods. In some cases, it is then possible to design analytical chemistry systems which combine the power of a modern separation technique with one or more analytical methods.

One of the best known precursors of such systems is LC-NMR which combines liquid chromatography with NMR spectroscopy. Since the LC elute flows inside a flexible capillary tube, it is possible to run it through several detectors, including the probe of a repetitively scanning NMR spectrometer. Though a practical realization of such a system is not trivial, there are today efficient, commercially available LC-NMR accessories and there is intensive work in progress on other combinations, most notably LC-NMR-MS. Needles to say, this whole instrumentation area faces a rich and bright future.

After this somewhat lengthy introduction, the answer to the original question is simple. Since all these combinations of a separation technique with (one or more) analytical techniques are designated by multiple acronyms connected typographically by hyphens, the combinations themselves are called hyphenated methods.

Now, if you think that this is a semantic joke, you are absolutely right! I do not know who started the terminology, but the egg hatched and developed into an even more bizarre linguistic monster. When the combined techniques are many, like in LC-UV-NMR-MS, we are in the presence of a hyper-hyphenation or, abbreviated, hypernation and hypernated methods. Browsing through the literature, I have got an impression that hypernation starts from three hyphens up (unless it has already been done, I suggest that some international standardization body makes it official :-)

The research areas where hyphenated and hypernated techniques are most appreciated include, for obvious reasons, screening of natural products, drug discovery, metabolomics, food research, and many other branches of biochemistry.

Selected References:

January 24, 2008

Bruker + Bruker = Bruker

"Once upon a time there was a charismatic physicist named Günther Laukien who fathered four sons ...". Well, that could be quite a tale, but what do I know? At the time I left Bruker, the four were still kids.

At this moment, the story's closing paragraphs would no doubt regard the acquisition of Bruker BioSpin by Bruker BioSciences and the imminent formation of a unified Bruker Corporation with 3700 employees. For details, read this press release by Reuters.

To Bruker Users the important question is whether and how will the merger affect them. While Bruker answers to that are totally predictable, I would expect that their top brass will be for a while busy managing themselves and thus have less time for routine matters.

January 22, 2008

Girjesh Govil

Magnetic Resonance of Spermatozoa

There are meetings which I have not listed among primary MR events because their magnetic resonance content appears to me too 'diluted' for an entry. Often, however, such meetings include MR 'splinters' which are of considerable importance and constitute an MR event in their own. This is the case of the G.N.Ramachandran Lecture by Girjesh Govil at the next meeting of IUPAB (International Union for Pure and Applied Biophysics), to be held February 2-6 at Long Beach, California.

The evening lecture (7-8 pm), entitled Magnetic Resonance Studies on Spermatozoa, will be delivered on February 5 and there is little doubt that it will leave a mark in the history of Bio-NMR. Girjesh Govil is a senior scientist affiliated with the Indian Tata Institute of Fundamental Research, member of IUPAB, and recipient of many awards and honors for his life-long dedication to NMR in general (he is a key figure in the history of NMR in India) and NMR of nucleic acids in particular. Most notable is his 1990 NMR and ESR study of intact spermatozoa, a milestone in proteomics, transcriptomics and genomics.

Together with K.V.R.Chary, he is the author of the textbook NMR in Biological Systems: From Molecules to Human (Springer) which is slated to leave the presses in a matter of days (you can already order it).

January 17, 2008

Directory of MR blogs of general interest

I have added a list of Magnetic Resonance blogs as a new category to my MR links page. There are of course an infinity of NMR blogs concerned with the daily routine of this or that NMR center which are of "local" interest and therefore not listed. At this moment I am aware of eight blogs which carry content of general interest (some of them, like this one, are actually closer to e-Zines than to real blogs). Since I am sure that there are others which I have missed, please, let me know of any which you think should be added.

January 15, 2008

Don't miss: 
satellite meeting on
in spotting

(See my Report)

Production cost:
3-30 $US/mine
Mean removal cost:
ca 1000 $US/mine
Current stockpile:
> 250'000'000
> 110'000'000
Current estimate of
the time needed for
complete clean-up:
ca 500 years
> 2'500'000 /year
People killed:
> 26'000 /year
Animals killed:
your guess!

Spin Resonance and Warfare

Spin resonance, be it that of nuclides (NMR, NQR) or of electrons (ESR), luckily plays no role in the development of weapons and offensive systems. There are two related areas, however, where spin resonance does have a role. These have both to do with remote detection of explosives, an evolving technology which is slowly becoming viable to
(1) prevent covert transit of explosive devices through mass-transport hubs like airports, public buildings, and military checkpoints (a defensive measure) and
(2) support clean-up after a war through remote detection of bombs and mines.

The second application [1,2] is of extreme importance since cleaning up mine-infested territories often takes much longer than the war which caused the problem, not mentioning the fact that demining is at present very slow, tedious and incredibly costly (around $1 per square meter, with hundreds of thousands of square kilometers waiting to be demined). At the rate new wars are erupting, there will soon be no mine-free terrain left on this planet.

The problem with todays explosive devices is that they no longer have an easily detectable metal case. The cases are generally plastic and thus impossible to detect by means of classical electro-magnetic induction or X-ray opacity and back-scatter. It is therefore necessary to concentrate on remote recognition of the explosive itself which brings in, among others [3], spectroscopic methods. Of these, the favored ones are those which combine a large penetration depth with pronounced chemical discrimination capabilities. The logical candidates are infrared (IR) and, above all, microwave and radio spectroscopies, such as NMR, ESR and NQR.

Imagine, for example, an airport checkpoint built as a large scanner which can detect both the shape and the chemical nature of whatever passes through, be it an explosive or a drug (the war on drugs has its own warfare, too). Or imagine an interplay between a radio-beam generator which sweeps a terrain and a receiver which analyses the dispersed radiation for specific spectral patterns and draws an explosives dislocation chart. Such Star-Track scenarios are still a fantasy - but it is the kind of fantasy that might come true. Actually, the detection does not need to be that much remote: any functioning device hovering 25-50 cm above the terrain, or placed close to the object to be tested would be extremely useful.

In this context, NMR is handicapped by the fact that it requires a rather strong magnetic field which is cumbersome to generate and manage. Techniques using very weak magnetic fields (from the Earth's 50 μT to about 1 mT) are being explored but, so far, without much success. ESR also needs a magnetic field but, since its frequencies are much higher, it could be more viable, were it not that most explosives do not contain unpaired electrons.

This leaves NQR, or just QR as they call it in this niche, which can operate at quite high frequencies even in the absence of magnetic field. Moreover, most explosives contain a lot of nitrogen whose most abundant nuclide 14N has a strong quadrupole moment and gives rise to spectral "fingerprints" characteristic of the particular explosive substance [3b, Fig.K2]. The handicap of QR is that it works only in solid substances; in liquids the quadrupole interaction term averages to zero and the NQR transition is not observable. That, however, still leaves a lot of viable substances. Consequently, attention has focused mostly on the NQR of nitrogen (explosives) and, more recently, chlorine (drugs).

The history of the QR niche is quite interesting, though far from clear (to me). The idea was apparently pursued already in the 70's at US Naval Research Laboratory (NRL), but dropped after the Vietnam war. Allen Garroway (still at NRL) dusted it off in early 80's and did a lot of pioneering work on it [4]. At the time - and quite independently - the same idea has been pursued since three decades by Vadim Sergeevich Grechishkin and his group [5,6] at what is now the Kaliningrad State University. In 1993, NRL promoted the formation of Quantum Magnetics [7,8] and licensed the technology to them; this successful venture was later absorbed [8] by General Electric Security (search for explosives on their web site to view what they deem publishable). On the Russion side, Grechishkin's group nearly dissolved, possibly as an aftermath of the failed Afghanistan war, disintegration of Soviet Union, and Vadim's advancing age (he was born in 1933). When a second private venture dedicated to this application, the QR Sciences, was formed in 1995 in Australia, several Grechishkin's ex-coworkers turned up there and there even arose a feud between Grechishkin and QR Sciences regarding conflicting priority claims [6]. Finally, during the 90's, new academic players [14] started entering the picture while, at the same time, detection of explosives partially lost its military character and acquired its present humanitarian [2,9,10] and civil-security connotation (see an example).

On the scientific side, an interesting trend consists in innovative combinations of QR with NMR, even in weak magnetic fields. In one of the most recent articles [13], its authors show how PE-NQR (polarization enhanced NQR) can help to detect week NQR lines in TNT. In plain QR, the low sensitivity to TNT (80-90% of all landmines) is a big problem. TNT resonances are at mere 0.84 MHz and have a long T1, implying a very weak signal (in comparison, RDX frequencies are around 5 MHz, with short T1, and therefore easy to average and detect). By transferring polarization from 1H to 14N, it is possible increase the signal and detect it in a single-shot. Unfortunately, proton T1 is also rather long so one needs about 10s to polarize the protons, which may be too long for some applications.

So NMR, after having been barred at the front entrance, is re-entering through a back door. Moreover, I would not be at all surprized if a sudden breakthrough in SQUIDS technology (superconducting quantum interference devices) permitted to boost up NMR sensitivity in the μT to mT range by several orders of magnitude, in which case low field NMR would almost certainly become a protagonist on the scene. We will see.

Selected References and Links:


For a complete list of all entries since 2005, see the running INDEX

  Visitor # 

Centres & Groups
Journals & Blogs
Free Texts
WWW | Patents & IP
Hebel-Slichter effect
MR Antenna Theorem
NMR Dead Time
One-Page MR Primer
K-space and MRI
S/N Perspectives
SI Units & Dimensions
Physics constants
Science Links

Support this site!
SHOP from here:


This page is

Chemistry software solutions

Dedicated to excellence!

Random offers:

TOP | Running INDEX | PREVIOUS page | NEXT page NMR in Stan's LIBRARY | Stan's HUB
Designed by Stan Sýkora