Hello, my fellow Bench-top NMR enthusiasts! If you’re like me and ever pondered your deeply burning engineering questions concerning NMR on the bench top, then look no further! I have condensed and purified the electrical engineering of NMR to a readily digestible schematic that will assuage the fears that bench-top NMR instrumentation is somehow an invariable “black box” technology. In other words, prepare to be mystified through de-mystification!
My first exposure to NMR spectroscopy was in an advanced elective course taught by my colleague and mentor Dr. Kenneth Metz at Boston College (see how they are enjoying their 3 NMReady 60’s on their YouTube testimonial here). Always with a penchant for hands-on experience in education (just like us here at Nanalysis), Dr. Metz immersed his students in a rigorous world of NMR theory with a strong focus in electrical and mechanical engineering. I’m sure he could still recall my initial reaction when he tasked me to build a 15 MHz NMR right there on my laboratory benchtop with a foil-wrapped box and hodgepodge electrical parts from the local specialty store. Luckily, my recently acquired electrical engineering skills were put to good use that day. What an experience to share with our fellow Nanalysis blog readers!
I’ve generated a similar electrical schematic that I think explains NMR instrumentation quite succinctly. If you’ve taken a look at the schematic already and are a bit worried, don’t be! At most, this will require just a bit of fundamental NMR theory and a basic knowledge of the properties of electromagnetism.
So, if you care to accompany me on this scientific pilgrimage, I’ll guide you through it step-by-step, fueled in part by my newfound Chemistry B.S. confidence and, more importantly, by obscene volumes of coffee.
The rf Generator
Let’s start at the very beginning, which—as my good friend Julie once said to me—is a very good place to start. In this first blog we will focus on the all-important excitation wave, which is vital to exciting our sample such that it generates for us a unique NMR signal, which will be our focus in the second installment of this blog.
We all know the fundamental Larmor Equation:Where ν is the procession frequency of a given nuclei (in this case, hydrogen), which is dependent on that nuclei’s unique gyromagnetic ratio (γ for 1H = 42.576 MHz/T) and the strength of the induced magnetic field in which our sample will be analyzed (Bo). For a common high-field NMR operating at 400 MHz 1H, this corresponds to an induced field of 9.3951 T. But don’t sweat; I was not tasked to build a superconducting magnet with only spare RadioShack parts. The magnet we used was much weaker, and therefore the 1H Larmor frequency required for our set-up is 15 MHz.
Glorious! If the larmor frequency is 15 MHz that is the frequency that the rf generator is set. This will generate a sinusoidal continuous wave with a frequency of 15 MHz and carrying a potential of 10 volts (a property specific to this generator). It is this rf that will serve as the excitation wave that tips the protons from alignment with the external magnetic field, causing them to process, and subsequently allow us to collect the NMR signal.
Attenuator
Our next stop brings us to the attenuator, which will reduce the Volts of our excitation wave from 10 V to 1 V. While this seems menial, it’s actually very important for two reasons: 1) a signal of 10 V will undoubtedly fry the components at the end of our schematic that we need to acquire and process our signal (more on that later); and 2) to allow us to mathematically match the output voltage of our signal that comes out of the tail end of our schematic. In short, it’s very important. It will all be made clear later, but for now…trust me!
*Note here that just after the attenuator, our circuitry splits from the junction into two equally excited (15 MHz and 1 V) signals. The wire in red sets off on it’s own easy journey. We must follow the black wire, which is the more challenging but more exciting path, and we’ll reconnect with the red wire later.
Computer, Gate Driver, and Gate (rf Switch)
So far, our continuous excitation wave is all well and good, but if we excite our sample with it, the protons are going to nutate continuously; this lack of control is of no help to us because the signal that we are attempting to acquire will just be saturated. For simplicity’s sake, it is sufficient to only nutate the nuclei 90° or orthogonal to the induced field vector. This maximizes the induced NMR signal, but requires the continuous wave to excite the sample for a specified period of time, or “pulse width”.
This is done by way of the computer, which generates the chosen pulse width (imagine the square wave shown in the schematic). This is first sent through a Gate Driver to increase the current generated by the computer (to be compatible with the instruments circuitry). The computer is programmed by the user to dictate when the square wave will be generated, at which point it joins up with the 15 MHz continuous wave to be compounded by the Gate. This is also called the rf switch, and it is essentially a double balance mixer (see more advanced electrical circuitry blogs for a tutorial on that). The outcome is that the Gate generates a 15 MHz wave of a fixed pulse width that is specific to nutating our 1H nuclei.
rf Amplifier
Now that we have trimmed and tailored our excitation wave, it needs to be amplified. There are a few reasons for this: 1) Perhaps most importantly, we need a strong enough signal to nutate the bulk magnetization of the sample out of alignment (recall that even in high-field, high power environments, an NMR will only nutate or “flip” a small number of nuclei in excess of those that remain aligned with B0, and this small excess is usually on the order of a few ppm); 2) a very close second in terms of importance – amplifying our desired wave will improve the separation of this signal from the baseline noise, which we will need to cleanly separate for an accurate acquisition; and 3) we must also differential this big excitation signal from the small Free Induction Decay (FID) signal we hope to acquire from our sample, and separate it from this, too…..more on this later…..
First Crossed Diode
If any of you are rusty on the properties of crossed diodes, I must say they are pretty clever: they act as an intrinsic switchboard of sorts for allowing large voltages to pass while simultaneously hindering small voltages. This hindering understandably causes a buildup of energy that needs to be released, and we’ve learned to harness that property in devices like Light Emitting Diodes. They can also act as easy-to-replace safety measures, or lightening rods to attract renegade energy to fry the cheap diodes as opposed to the other, more expensive components. But I digress…
For our application here, we have a diode that allows large voltages to pass (i.e. our signal), but does not allow small signals (noise) to pass. Only one is drawn here, but it is easy to visualize that running many of these in series will protect the instrument from burning out, optimize the signal-to-noise in the circuitry, and prevent the holy-grail FID signal from passing backward.
Now that our excitation signal is optimized and stripped of noise, it is ready to do its job. Following the schematic, the excitation signal takes a quick turn at the junction to charge the Matching Capacitor.
Matching Capacitor
First, I’ll remind you of a theorem in electronics, which is namely that in order for a maximum (complete) energy transfer across two systems, they must be of equal impedance; and impedance is of course an intrinsic property of a conductor (wire) that opposes flow when any voltage is applied to it. In case you plan to impress your friends at trivia in the near future, this is also known as the Maximum Power Transfer Theorem, or Jacobi’s Law. The job of this matching capacitor, then, is to ensure maximum energy transfer to the probe circuit by matching the impedance values of the probe circuit to the rest of the circuitry in the schematic. If we didn’t ensure maximum transfer of energy of the excitation signal, then we wouldn’t excite the sample sufficiently to ensure a maximum FID, and we certainly wouldn’t have a maximized FID. This matching capacitor is usually engineered to be variable by a spinning dial on high-field instrument probes (labeled “M”).
Finally, ultimately, and at last, the excitation wave has reached the inductor in the probe, induced a field B1 for a given pulse width to nutate the protons for a short time, and then is ceased. Of course, this excitation wave is so strong that it will saturate the remaining circuitry (namely the Pre-Amplifier; see later), and so we must make sure we have a system in place to ensure that the large voltage takes that “right turn” I mentioned earlier, and doesn’t continue beyond that to obscure the FID signal that we need to propagate to the end of the schematic. This brings us to:
Second Crossed Diode, Grounded
Just like the First Crossed Diode, one can imagine that the circuit is “on” for large voltage signals and “off” for small voltage signals. This crossed diode is also subsequently grounded, which in turn, causes the whole crossed diode to have no resistance up to the crossed diode’s junction with the rest of the circuit. Since at this junction point Ω=0, this also means that at a distance λ/4 from this junction (λ being the wavelength of the 15 MHz signal), Ω=∞. Of course, this is only true if the crossed diode is “on”, which is in turn only true if the signal in question has a large voltage. If not, and there is a low voltage signal (like our FID signal), then the switch is “off” and the FID signal continues to the Pre-Amp unhindered. In this way, the second crossed diode acts as the true and definitive traffic officer that directs the high voltage excitation wave only to the probe to excite the sample, but allows the freshly acquired FID signal to continue on past it to the Pre-Amp.
Since we have followed the journey of the excitation wave from its generation to its arrival at the resonance circuit (more on this later), now is as good a time as any to take a quick break to separate the concepts of the excitation wave and the resulting NMR signal. The latter is what we will follow in the second installment of this instrumentation blog entry.
So stretch, grab some water, and tell your colleagues about your epic quest on this blog. Act 2 is about to commence!
[1] Metz, K.R.; Experiments with a Bench-Top NMR Spectrometer. 2011, 1-17
[2] Zwolinski, N. M.; Laboratory Notebook. 2011, 20-27
[3] Crews, P.; Rodriguez, J.; Jaspars, M.; Organic Structure Analysis. 2nd Edition. 2010. 27-30