Wednesday, November 01, 2006

Work with your hands

A variety of computer programmes is now available and it appears that we are rapidly approaching the day when spectrometer-computer interfacing will permit N.M.R. spectral analysis to be performed in an entirely automatic fashion.
Gerhard Binsch, Mol. Phys. 1968, 15, 469

Binsch was one of the pioneers of NMR software, with a brain 100 times larger than mine. At the time he wrote the sentence above nobody, him included, could have predicted the evolution of the computer industry in the following 38 years. There was no notion of graphic user interfaces and personal computing. He foresaw programs that could have automatically generated lists of chemical shifts and coupling constants and when he stressed the "entirely automatic fashion" he assumed that a computer was more esoteric than an NMR spectrum, therefore he could not have asked the average organic chemist to operate on it. Nor could he foresee the simplification of spectra through higher magnetic fields and multi-dimensional techniques.

The computers that appeared ten years ago allowed us to change the processing work-flow.

  1. It became possible to correct the phase of a 2D spectrum directly, in real time. It was no more necessary to correct a few selected rows and columns and then report the same correction on the whole matrix. We can directly and continuously monitor the final result.
  2. It became possible to perform weighting in the same fashion. We need no more to estimate from the shape of the FID the amount of weighting to apply. We can change it with a mouse movement a monitor the effect, in real time, on the frequency-domain spectrum.
  3. There were the technical basis for human-driven fitting procedures. A peak can be dragged from one position to another inside a synthetic spectrum (a computer simulation), superimposed onto the experimental counterpart. With no background in signal processing and in quantum chemistry, even a child can fit a spectrum, adjusting any kind of parameter (line-width, percentage of gaussian character, coupling constants, etc...)

I don't mean that he/she can do without studying, I mean she can make an instructive experience of the field, which will both open her mind and increase her desire to study.
If you object that the same things should be automatically accomplished by a good NMR software, you are missing an important point. Manipulation is the best form of observation, and NMR spectra are recorded to observe the samples, not to populate the corporate database.

Around the same time (10 years ago) we also witnessed the explosion of the internet and of shareware distributed and sold over the internet. The user could not only interact with the spectrum, but also with the author of the software (in those cases the latter had not been fired by the software house, of course).
In those years, and even more now, we began enjoying larger monitors with more definition. The iMac G5 which I am writing onto has a 20" monitor. With it I can see the spectrum better than when it's printed, therefore I never print. I know that most people are forced to print the spectrum, because the representation they see on the monitor is not clear enough. This is still the common situation on Windows PCs, but things will soon change there too.
Finally new computers come with bottomless discs you where you can keep together the collection of NMR spectra of your entire career, available for immediate inspection.
I have shown how four external causes (faster CPUs, larger monitors, high capacity hard discs, the internet) can alter the NMR processing work-flow, how we manage and store the spectra, how we buy the software. I don't think that all software makers and software users have already realized this.


Post a Comment

<< Home