January 25, 2015
the great guy
who started it
Detection of a single nuclear-spin announced ... and retracted
A micro-story of MRFM nano-imaging, including a recent hiccup.
From the very beginning of Magnetic Resonance Imaging , there emerged an interest in pushing its limits towards the highest possible resolution. It became soon clear that the "classical" MRI set-up had severe resolution limits which would be hard to beat due to limited sensitivity (typically, the smallest number of nuclei an inductive sensor can detect is about 10^12). A different technology was needed.
Starting from a seeding 1991 paper  by John Sidles from Univerity of Washington, Seattle, a group of researchers and engineers at IBM, led by Daniel Rugar, tried to apply to spins a detection method known as MFM (Magnetic Force Microscopy). MFM uses the mechanical resonance of nano-sized, vibrating cantilevers, a nano-technology capable to detect extremely tiny forces (there is nothing that beats it). In their nice 1995 review , Sidles has named the new MFM-based MR technique MRFM (Magnetic Resonance Force Microscopy).
With this new technology, the Sidles - Rugar group started pushing down the minimum number of detectable spins. There was a rapid series of successes, with microscopic images of biological cell structures, viruses, and the like. Finally, in 2004, there came an exciting announcement  of the first detection of a single electron spin. The clever experiment combined cantilever detection with a magnetic field, a pulsed ESR micro-spectrometer, and an ingenious pulse sequence, all designed to make sure that what was detected was indeed a single electron spin. A true jewel of science and engineering!
That achievement started a "race" towards the detection of a single nuclear (not electron) spin. That, of course, is a much bigger challenge. Since electron magnetic moment is over 658 times larger than that or proton, and since the sensitivity of this type of sensors increases roughly with the square of the magnetic moment, a single electron beats in terms of sensitivity more than 400'000 protons! But the fata morgana of nuclear spin imaging at nanometer scale is a strong attractor. If it could be done, structural analysis of molecules would change dramatically: in principle, we could produce 3D images of individual molecules in all their complexity!
Naturally, the experiments of the Sidles - Rugar group were repeated by others. This brings us to the 2005 Thesis of Christian Degen . Though it is oriented towards nano-scale MRMF spectroscopy (implying many spins), it is pertinent also to our topic. It gave rise to a group led by Christian Degen and Kai Ebenhardt at the ETH in Zurich. They did fine job on improving the sensitivity in 2007  and applied their MRFM setup to 1D and 2D spectroscopy in 2008 . In the 2007 experiment they were able to detect a very sharp "ridge" of 19F nuclear spin magnetization, produced by MRI techniques, in a single crystal of CaF2 and "image" it with a resolution better than 10 nm. Not only that: they were actually able to monitor the spreading of the edge due to the spin-diffusion phenomenon and estimate the spin-diffusion rate, which turned out to be in full agreement with previous, indirect methods. A lovely paper which is today carved in stone in the MR Valhalla.
It is not easy to estimate how many 19F spins were involved in the experiment; the linear resolution being about 10 nm, the count boils down to something around 1000, give or take a factor of 5. Some optimistic estimates of detectability I have heard around 2010 [verbal communications] were as low as 50 - 200 spins.
Again, all these achievements were reproduced elsewhere and a lot of impressive results were obtained. See, for example, the top-quality Review of reference .
Then, in October 2014, just 3 months ago, came an announcement  from the ETH group that a single proton spin has been detected. In a very special system (supercool diamond crystal with a nitrogen defect buried just under the surface, close to the proton to be detected) but, nevertheless, a single nuclide! Wow!!!
Unfortunately, something went wrong: the authors themselves, very honestly, pointed out a potential problem  and, a few days ago, retracted  the Science paper. The retraction says that the observed proton signal might have been confused with the 4th harmonic of 13C Larmor frequency (13C nuclides are of course abundant in diamond). The retraction leaves open a chance that, in one out of the three cases described in the paper, a single proton might have been indeed observed.
What can one say? An error, of course, can happen. It may be that the present-day academic pressure to publish fast is just too high. Especially when there are groups tackling the same targets by competing detection technologies such as, in this case, optically detected magnetic resonance (ODMR). A November 2014 ODMR paper , in fact, also claims to have detected a single proton. They use the same diamond surface with nitrogen defects, and similar pulsed RF excitation schemes. Just the detector is different and allows the authors to work at room temperature. Which makes me wonder: might not the reasons behind the retraction of the Degen's group be pertinent also to the ODMR paper?
Anyway, the single-proton MRMF detection milestone is presently lost in a fog, kind of.
Hence, rejoice, because the race is still on! But better do not rejoice for too long, I think.
- Lauterbur P.C., Image Formation by Induced Local Interactions: Examples Employing Nuclear Magnetic Resonance, Nature 242, 190 (1973). DOI 10.1038/242190a0.
- Sidles J.A., Noninductive detection of single-proton magnetic resonance, Appl.Phys.Lett., 58, 2854 (1991). DOI 10.1063/1.104757. This title of this seeding paper expresses the author's dreams, not an actual achievement.
- Sidles J.A., Garbini J.A., Bruland K.J., Rugar D., Züger O., Hoen S., Yannoni C.S., Magnetic Resonance Force Microscopy, Reviews of Modern Physics, 67, p.249-265 (1995). DOI 10.1103/RevModPhys.67.249.
- Rugar D., Budakian R., Mamin H.J., Chui B.W., Single Spin Detection by Magnetic Resonance Force Microscopy, Nature 430, 329-331 (2004). DOI 10.1038/nature02658.
- Degen C.L., Magnetic Resonance Force Microscopy: NMR Spectroscopy at the Micro- and Nano-Scale, Diss. ETH No. 16304 (2005). DOI 10.3929/ethz-a-005128707.
- Eberhardt K.W., Mouaziz S., Boero G., Brugger J., Meier B.H., Direct Observation of Nuclear Spin Diffusion in Real Space, Phys.Rev.Lett. 99, 227603 (2007). DOI 10.1103/PhysRevLett.99.227603.
- Eberhardt K.W, Degen C.L., Hunkeler A., Meier B.H, One- and Two-Dimensional NMR Spectroscopy with a Magnetic-Resonance Force Microscope, Angew. Chem. Int. Ed. 47, 8961 (2008). DOI 10.1002/anie.200802978
- Seppe Kuehn S., Hickman S.A., Marohn J.A., Advances in mechanical detection of magnetic resonance, J.Chem.Phys. 128, 052208 (2008). DOI 10.1063/1.2834737.
- Loretz M., Rosskopf T., Boss J.M., Pezzagna S., Meijer J., Degen C.L., Single-proton spin detection by diamond magnetometry, Science Express, October 16 (2014). DOI 10.1126/science.1259464.
- Loretz M., Boss J.M., Rosskopf T., Mamin H.J., Rugar D., Degen C.L., Spurious harmonic response of multipulse quantum sensing sequences, arXiv:1412.5768 (December 2014).
- Loretz M., Rosskopf T., Boss J.M., Pezzagna S., Meijer J., Degen C.L., Retraction, DOI 10.1126/science.347.6218.139-a.
- Sushkov A.O., Lovchinsky I., Chisholm N., Walsworth R.L., Park H., Lukin M.D., Magnetic Resonance Detection of Individual Proton Spins Using Quantum Reporters, Phys.Rev.Lett. 113, 197601 (November 2014). DOI 10.1103/PhysRevLett.113.197601.
25 January 2015, Stan. Sorry, I forgot a very useful MRFM review:
Suter A., The magnetic resonance force microscope, in Progress in Nuclear Magnetic Resonance Spectroscopy 45, 239-274 (2004). DOI 10.1016/j.pnmrs.2004.06.001.
January 11, 2015
A personal view of the state of NMR at the start of 2015
NMR is presently going through an unusual phase. For a number of reasons, and for the first time in history, traditional NMR spectroscopy is commercially unprofitable. It was totally abandoned by one major manufacturer, while others are anything but enthusiastic about it. At the same time, there are many new commercial ventures aimed at considerably different types of NMR instruments (compact, sub-compact, table-top, cryogen-free, portable, ex situ, ffc, ...).
The worrisome situation is limited to high-field NMR spectroscopy (HFNMRS) which was for many decades the staple of just three manufacturers: Varian (later Agilent), Bruker, and Jeol. It does not regard MRI, whose market keeps growing at a steep yearly rate of 11-12%. Sincerely, even in MRI, I do not think that the growth is a merit of innovation. My feeling is that, compared to three years ago, there is less talk about fNMR, spectroscopic MRI, molecular MRI, interventional MRI, diffusion and susceptibility imaging, or even just novel pulse sequences. Rather, the steady growth of MRI is driven by the fact that the majority of people on this planet still have only a very limited access to this important diagnostic tool. Even with limited innovation efforts, MRI is therefore enjoying a vast potential market. With current production capacity, any saturation of that market is out of question for 100 more years.
HFNMRS, on the opposite, faces a saturated market, mountains of used but recyclable hardware, and disenchanted manufacturers who have nearly forgotten the word "innovation" (indeed, I suspect that they might be dreading it).
Recent market reports show that between 2012 and 2014 the market for HFNMRS has slumped. The drop does not correlate much with the overall economy (it is true that Europe is in an economic quagmire, but USA are running strong, and some parts of the World are doing even better). Some explanations, apart from the European weakness, might be the following:
- In 2010 - 2011, many big Pharmas reduced their research facilities. Since, excepting academic NMR Centres, Pharmas represent the principle HFNMRS users, their crisis has dealt a sizable blow to that market.
- Many Pharmas pushed their downsizing as far as actually dismantling existing facilities (some with tens of NMR spectrometers per facility), sending lots of NMR hardware into the second-hand instrumentation market to compete with newly manufactured units.
- NMR as a structural-analysis chemical tool is no longer as prominent and unique as it used to be. Even though my friends among the chemists assure me that NMR is as highly considered as ever, it is evident that other analytical methods have made impressive advances. HFNMRS looks today as a mature technique, developing at a substantially slower rate than in the 80's when its innovation rate was peaking.
While the NMR market was slowing down, the principal manufacturers of HFNMRS equipment were (and some still are) facing other, more specific, challenges:
On October 14, 2014, Agilent Technologies have abruptly terminated all their NMR activities. The brusque, almost offensive, announcement came as a surprise even to most Agilent employees. Well, not all, perhaps. The eminent scientist Erics Kupce, for example, must have figured it out already by the end of 2012 because he resigned from a nominally excellent position at Agilent, effective January 1, 2013. But Agilent NMR guys in Italy, for example, were in the dark until the announcement.
Agilent entered NMR on July 27, 2009, by acquiring Varian Inc, the company born in 1999 from a split of Varian Associates. The latter, in turn, used to be a much studied textbook example of Silicon Valley industrial innovation capabilities in the post-WWII period (roughly up to 1975 - then it became more complicated). The Agilent's decision to drop NMR de-facto guillotined the last vestiges of a once glorious Chapter in NMR history.
It is ironic to think that in 2009 Agilent entered NMR at the sound of fanfares: they spent 1500 million dollars to acquire Varian, including its magnets production facility in Yarnton, UK, (ex Magnex), only to drop it all at a huge loss mere five years later! I was at the public grand open-house kick-off of new Agilent NMR premises (Santa Clara, April 8, 2011, prior to ENC). At the meeting, where historic personalities like Weston Anderson were present, American Chemical Society conferred to the would-be continuators of Varian traditions a plaque stating that the mythical A60 NMR spectrometer was declared a National Historic Chemical Landmark. There were bright talks, great buffet, and a lot of hype. However, the two questions that I fielded (how would Agilent handle the maintenance of the installed base of legacy Varian instruments, and how would Agilent react to the forthcoming helium crisis) were answered in a way that did not convince me.
My guess about why Agilent totally dropped NMR only 5.5 years after the merger:
- They declared to SEC, and therefore to their investors, that the 'merger' would cost 1500 M$, and that they 'expected a 20% return on it in 4-5 years'. After 5 years, instead, there was a large loss. Hence, since the takeover was no bleeding NMR heart operation, it was scrapped and written off as a deductible loss.
- The problem emerged just when Agilent was going through another internal crisis. The following announcement appeared on all their websites, starting at the same time as the NMR dismissal (it is still on most Agilent pages):
"As of November 1, 2014, Keysight Technologies, the division of Agilent Technologies that develops and sells electronic measurement (EM) products, will de-merge from the Agilent Technologies group into a separate publicly traded company, ..."
Now, Agilent's EM products were highly valued by electronic engineers worldwide. To thousands of electronics professionals, the term Agilent was simply equivalent to top-quality electronics-lab-instruments. I have realized myself only after I have seen the above announcements that this regarded just an acquired subdivision.
And Agilent was about to lose it!
- In those 5 years Agilent did not contribute any innovation worth mentioning to the ex-Varian products. They redesigned the closet on the DDR console, so that they could put the new logo on it and call it DDR2. Internally, they changed only one board, as far as I know (the filters). They did development on probes, but nothing revolutionary. And that's it. Everybody expected that the masters of RF and MW would redesign it all (the DDR console was an ageing heredity of the ex Chemagnetics takeover by Varian Inc). But those masters (Keysight) were in another department and, as it turned out, ready for a divorce.
- To top the mess, Agilent decided to relocate the production of the DDR2 console from New Mexico to Malaysia, while, at the same time, software development was being moved from India back to California! Can you imagine!
- World-wide maintenance of the ex-Varian legacy base must have been a nightmare. At the grand opening, they said they would rely on a software package to handle it! And already then they, themselves, anticipated problems due to lack of RoHS compliance of the old stuff. It is illegal nowadays to sell or even just to ship around old tin-and-lead containing electronic boards. So what you do if that is the only kind of spare parts you will ever have? With no plans for spare-parts production you are stuck to what you can cannibalize from discarded units, and you can't ship it anywhere without breaking the law. It did not make much sense to me.
Anyway, all this is History today.
And Agilent should officially return the A-60 landmark plaque to ACS, I think.
Bruker Corporation, of Swiss-German origin, is now US-based, with an impressive (excessive?) range of activities. NMR represented most of their business from 70's to mid 90's, but it is nowadays in minority, though not yet marginal. In the last decade, investments into innovation targeted mostly branches other than NMR.
My very personal opinions:
- The overall quality of Bruker NMR is high, but it is based largely on the heredity of the past (except for high-field magnets - they did a good job on them).
- Bruker business structure is geared towards commercially heavy items. In the past they went into cryoprobes. More recently, they went into DNP, adopting its most costly variant. To address the helium supply problem, they have chosen online helium re-liquefaction - again the most costly option. Their 'normal' probes are good, but outrageously costly. Etc. It is not a technical problem, I think. They probably cannot make profit on anything costing less than ~70 K$.
- On the applications side, it is 7 years I hear the same talk about orange juice ...
- They remind me of IBM when the PC appeared on the horizon. There are impending NMR revolutions, but I suspect that Bruker is dreading any of them that might mean lower prices!
Bruker hard facts gleaned from public documents (study the link, particularly the 'COMPANY' section on the left, including the SEC filings):
Bruker is today a public Company with 'insiders' still detaining a sizable part of its shares, but not a majority (about 35%). Presently it declares 6200 employees, compared with 6700 in January 2014. Among major shareholders there are banks of the like of J. P. Morgan. In the third quarter of 2014 they declared a loss. Commercially speaking, they estimate that NMR will perform poorly in 2015 - 2016 (at best in the 'lower single digits'). A year ago, to tranquillize their shareholders, they announced ongoing drastic measures: reduction of operating divisions (from 10 to 3 just in USA), reorganization of all businesses, continuing 'headcounts' (HC) reduction, continued outsourcing, relocation, and dismissals ... (source: Bruker JPMorgan 2014 Healthcare Conference presentation, pages 14,15,18,20).
As I wrote on a LinkedIn forum, I do not enjoy all that at all. I feel sympathy for Bruker and I would much prefer to see them snap back into shape. But right now, they are going through serious troubles and they view NMR more as a culprit than as a solution.
I cannot find out much about Jeol (the Japanese 'SEC' is not as open-access as its US counterpart). They were keeping low profile recently, mostly focusing on probes (MAS microrotors, cryo-cooled MAS), in collaboration with Universities. Somewhere I read that they had kept NMR on a stand-by for three years. On the other hand, they have recently installed a 600 MHz instrument in my birth town, Olomouc (Moravia, Czech Republic). I was amazed because Olomouc is a small place, and Czech Republic is in the heart of Bruker territory, jealously guarded by Scientific Instruments Brno. As always, Jeol's presence is a patchwork: significant in some Countries (UK, Scandinavia) and weak or null in others (I don't know about USA). Overall, they do not appear to be a formidable Bruker competitor, though on the technical side they do have all the prerogatives. Their magnets supplier is Jastec Superconductor, a very respectable Company indeed.
Wrapping it up:
Something worrisome is happening to HFNMRS instruments manufacturing and R&D. The market is depressed and there are dangerously few manufacturers. One manufacturer has closed shop, and the remaining two are anything but enthusiastic about NMR. At the same time, there are several new players, but they are still far behind in both volume and technology.
There is a lot of used hardware, so the NMR community can go on for 2-3 years regardless of manufacturers. Competent tinkerers and maintenance engineers will be much needed, of course, but there are lots of them. They are being made 'redundant' even while you read these lines. Btw, the capability of the community to go on is the only thing that can keep Bruker from exaggerating with prices. The vicious price wars of the last decade are over, of course, and the net tabs for new equipment will go up, no doubt about that. But, maybe, at least the list prices will not change.
Might it be that something big is looming on the horizon?
For would-be players of tomorrow, this is probably the best time to step in! It could be one of the smaller ventures (there are several candidates), or it could be Jeol. Or, more likely, one of the 'official' competitors of Bruker, which are Thermo Fischer and Danaher, both about ten times bigger. Though, have a look at this.
Oh well, so many things could happen ...
COMMENTS are welcome (especially from the Companies in question):
12 January 2015, Carlos Cobas, Mestrelab Research
Hi Stan. Just to let you know: Ray Freeman and Garret A. Morris have recently published an article entitled The Varian Story which you will no doubt enjoy.
12 January 2015, Stan
Thanks, Carlos. I must also acknowledge that you kept me posted with Agilent news during last three months. I liked and commented the article Agilent Draws Academics Ire on C&EN, and I have printed and saved your email containing a copy of the Agilent Open Letter which appeared on Agilent's Spinsights Blog. Btw, I can no longer access Spinsights - I guess it is dead now.
13 January 2015, Craigh Butts, Bristol University
Hi Stan, Just a couple of comments on your 'State of NMR' blog:
Firstly, I think that is a nice overview, balanced and pretty fair given the scope of the situation!
Secondly there are a couple of relatively minor (but perhaps important) points I'd like to add:
(1) I agree that from 2009-2013, Agilent had very little impact on the hardware but they did have a massive impact on the software - the latest versions of VNMRJ are much more powerful than anything I've seen in the past (I just installed a new probe, and the software shimmed it from zero shims and had it at lineshape spec, with full multinuclear calibration and signal:noise spec within 13 hours all under full automation (I did nothing after I had loaded the calibration samples onto the robot and pushed the 'Go' button). My research chemists can set-up and run a dozen different selective multi-dimensional NMR experiments in less than a minute (if they are being slow) and the gradient shimming on our routine CDCl3 samples takes less than 12 seconds. That just wasn't possible 5 years ago.....
(2) In 2014 Agilent released some absolutely gorgeous new hardware - in particular their new coldprobe which whipped the performance of the existing cryo/cold probe technologies in terms of sensitivity (the 600MHz coldprobe had better sensitivity than 800MHz cryo/coldprobe). And they had a new console and probe ready to release (it would be out by now if they had not closed up shop) which had a pretty large number of modifications and improvements.
(3) JEOL have actually had a number of hardware developments in the past 12-24 months that are directly competitive with Bruker in the HFNMRS field e.g. 13C-cryogenic (helium) probe, a liquid-N2 cryogenic probe (both of those cryoprobes are at 500-600MHz, so limited field availabilities at the moment) and ultra compact (400MHz) spectrometers (and a few less interesting ones like He-recycling magnets and new sample-changer). Their software needs a lot of work, but their hardware is actually pretty competitive.
13 January 2015, Stan
On your first 2 points: You are right that I left out the period of Agilent NMR enthusiasm following the Varian takeover. Agilent had retained almost all Varian STEM employees, and since they were all unhappy with the old Varian management, they were truly delighted. They would say "finally things will start moving" and in fact they did, including R&D support. This lasted 4 years. The first layoffs started in Autumn 2013 when probably the nasty decision was taken (at the investors level, I would guess). So, there were four years of motivated and competent work! That must leave a trace, and your comment illustrates that it did. Were the results insufficient, or were they too late? Or were the initial expectations too high? Or did the market really drop so hopelessly that developing for one more year was out of question? The 2014 period is the most puzzling one: rumors about great new hardware were persistent, but I am 99% sure that the decision to stop was already taken. I can only speculate on what was going on there between October 2013 and October 2014. One thing is sure: the destinies of HFNMRS are no longer in the hands of people who know what a spin is.
14 January 2015, Vanni Piccinotti, ex VPNMR (Italy), ex Varian
Ciao Stan. From this situation there MUST emerge in Europe at least 3-4 persons willing to start an independent technical service for the Varian/Agilent installed instrumentation, as well as sales of used, reconditioned units. Should this not happen, it would be - in my opinion - a terrible signal of total lack of entrepreneurship and of willingness to accept any risk. As you know, I have done it many years ago, it worked very well for me, and if I just had 10 years less, I would go into it again.
15 January 2015, Stan
Hi, Vanni, what a pleasure to hear from you. You are totally right, and I am afraid that EU entrepreneurship is, indeed, in a woeful shape. But the situation is not totally bleak. The French company RS2D is well positioned for refurbished Varian (and Bruker) instruments market (look for RMN/Reconditionné on their website). The German Spin-Doc, run by Bert Heise (an ex Varian/Agilent guy whom you certainly know well), is available for maintenance, as well as whole units recycling. Another existing German company, NMR Service is well positioned. They are, among other things, a Tecmag console dealer, a feasible option for ageing ex Varian systems, at some point.
But you are right, given the volume, and the geographic spread of the installed ex Varian/Agilent 'park', much more is needed. For example, what are the Italian Agilent NMR engineers waiting for? Or, with your competence as a resource, you could start a youngster on this road; an independent one, or maybe someone hired for the specific purpose by an Italian NMR company such as Stelar and/or SpinLab.
19 January 2015, Vanni Piccinotti
** Point 1, existing activities: I knew of most of them. Very good, but they all work independently, when the most obvious thing is to co-operate with each other to supply good service and products over Europe. To compete with each other would be total madness, there is plenty of room and work for everybody.
** Point 2, ex-Agilent people: They are still recovering from the bombshell, I guess. Give them time, something will hopefully come out.
** Point 3, training a youngster: I tried years ago, but could not find anybody interested in such a job. Reasons: too much to study and too much work, travelling around and often far from home, too many working hours, etc. This is the situation, at least in Italy (no comment). For what I know, big Companies also have problems to find willing (and capable, on top of it) servicemen for complex instrumentation like NMR.
20 January 2015, Stan
Vanni, your comments could start many new discussions!
For the moment, I hope others (like ex-Agilent NMR guys) will contribute. I have checked on LinkedIn; in Europe, most of them are apparently still with Agilent which means that they might eventually keep their jobs, but drop out from NMR.
On another front: JEOL seems to be waking up and moving. At least in Italy where they have even translated into Italian (an overkill!) their promotional materials for the new line of JNM-ECZ instruments.