Norman Ramsey - Session III

Pavlish:

Today is Monday, December 4, 2006. I am here to interview Professor Norman F. Ramsey in his office in Lyman Labs. My name is Ursula Pavlish. Professor Ramsey, you would like to tell me about your book, Spectroscopy with Coherent Radiation.

Ramsey:

Yes. This is the nearest that I have to an autobiography. The book is organized in the following way (the publishers do this for some Nobel Prize winners and they did it for me and I’m very happy about it), the publisher asked me to write a brief autobiography, about 20-30 pages long up to the time of when this book was published. It is essentially a scientific autobiography, but there’s some personal biography in it too. Then, the publisher reprinted what I considered to be my 40 or so most important papers. I have written about 500 papers, so I chose the ones that I considered to be the most significant. They reprinted them, with me making a one or two pages of comment about each paper.

Pavlish:

Putting it in context?

Ramsey:

Putting it in context, telling about the history, something or other about each paper. That’s basically it — the book was easy to write because most of it consists of reprints of the papers which I think are the most important. A couple of important ones aren’t included, but it’s has most. That’s the end of the book.

Pavlish:

How did you choose the title? Why is it not, “A Collection of My Most Important Papers with Autobiographical Commentary”?

Ramsey:

Well, in the first place they seemed to prefer a separate title. Secondly, the largest fraction of my research has been in the field of Spectroscopy. When I was working with Rabi we did the first spectroscopy experiment with coherent radiation. We didn’t fully appreciate that at the time because it was done with radio frequencies. If you make a radio oscillator, it’s automatically coherent. So, we never gave the coherency much special thought, whereas, optical experiments often use incoherent radiation. Stars and electric discharge tubes both emit incoherent radiation. I then realized afterwards, that we had actually done the first spectroscopy with coherent radiation and that it’s really quite a big field. I can really lump most of my magnetic resonance experiments as spectroscopy with coherent radiation. It seemed to me that was a title that quite described most of my important researches. Not all of them — it didn’t describe my pure nuclear physics research, but I would say that was the single best title I could get. The book is published by World Scientific.

Pavlish:

To me the term spectroscopy harkens back to the 19th century.

Ramsey:

Yes, that’s right. At that time it was done with light, usually with incoherent radiation. Even if the radiation had some of the virtue of coherence at that time, scientists usually did not recognize it. The first magnetic resonance molecular beam experiments that I did with Rabi were actually magnetic resonance spectroscopy with microwave coherent radiation. We realized it to some extent; when you have an oscillator you know how to make predictions quite accurately of what the shape of the spectra should look like. We did that, but even when we did that and we did it correctly, we didn’t realize that we were taking advantage of the fact that it was coherent. Later, the first experiment that explicitly used the coherency of the radiation was my experiment on the separated oscillatory magnetic fields for which I got my Nobel Prize. Even then it didn’t emphasize the coherency aspect as much as now. I think it’s a very suitable title. It was the first of the explicit uses of spectroscopy with coherent radiation.

Pavlish:

I did notice that at least in one of your early papers you referenced spectroscopy specifically. That made me wonder how much of the molecular beam method was coming out of the methods or the instrumentation of spectroscopy before it.

Ramsey:

Rabi’s initial experiment proposal was just to measure nuclear magnetic moments. This was thought of in terms of a resonance, that the nuclei would precess in an external magnetic field and if one put then an oscillatory magnetic field at the same frequency as it was precessing, he or she would get a transition. We calculated what it should be.

Pavlish:

At the Larmor frequency?

Ramsey:

At the Larmor frequency. To that extent we were doing spectroscopy but we didn’t think of it as spectroscopy. The first paper in which we thought of it as spectroscopy was the paper that Rabi and I wrote on molecular hydrogen. We discovered that we got more than one line. We were not measuring just the magnetic moment of the nucleus. We thought of it as dominantly precession but the precession was in the magnetic field that we put on externally and in the magnetic field that was there due to the other nuclei.

Pavlish:

And rotation?

Ramsey:

And the rotation of the molecule. I was the one who urged that the title of that paper should include the word spectroscopy.

Pavlish:

Was that related to work at Cambridge that you had done in spectroscopy?

Ramsey:

I was familiar with optical spectroscopy, which was basically what was done — looking at stars — but this was really quite different in form. The intriguing thing is we calculated it, taking advantage of the coherency but we didn’t realize the fact that that’s what we were using. We just said, we know how to calculate it. If you have an oscillating magnetic field from an oscillator, you can calculate the transitions that are induced. You really have to think of it in terms of spectroscopy when you have more than one interaction. Then, you can’t really think of it too well in terms of just the particles precessing in an external magnetic field. Qualitatively you can say that the nucleus is precessing in the resultant of the external magnetic field and the internal magnetic field of the other proton. Or, if there’s an electric quadrupole moment or the electric force will also result in precession. At that point, it really becomes best to think of it as spectroscopy. I think that was in the name of the second paper we published. We did it as a subtitle. Here’s the first paper, which uses that [Ramsey’s Collected Papers]. My first four papers in these publications talk about measuring the magnetic moments and the electrical quadrupole moment of the deuteron although they don’t say much of how we did it since these papers were just brief abstracts. Then, the next paper says, “The Radiofrequency Spectrum of the HD Molecule in Magnetic Field.” That was the first use of that phrase.

Pavlish:

You were the one in the group who…

Ramsey:

I was the one who pushed that one. In fact, Rabi and I had some arguments. Rabi was not initially very enthusiastic about that title but he agreed that it was accurate. We eventually settled on it and I think eventually he liked it very much.

Pavlish:

What was his initial resistance against it?

Ramsey:

Well, he basically had thought of it in other terms. He had thought of it in the form of using an oscillatory magnetic field to measure the precession. He’d written this very important paper on transitions in a gyrating magnetic field. That, he could calculate. Again, it was only for a single precession frequency.

Pavlish:

Please explain for the layperson, what you mean by coherence.

Ramsey:

Electromagnetic radiation comes in waves moving with the velocity of light. A particular wave has a sinusoidal form and a characteristic direction, amplitude, frequency and phase, where the phase is a measure of the advancement along the sine wave at a particular time and position. If two waves have similar direction, phase and amplitude, they are coherent and can interfere with each other. If the differences in phase vary with time and the waves are incoherent. For photons, coherency is defined in terms of correlation functions.

Pavlish:

What difference does coherence make?

Ramsey:

It makes a number of differences. In the first place, you can calculate things that way and get greater precision. You can take the oscillatory magnetic field and split it into two coherent parts and later have them interfere. Although we had used the term spectroscopy before, our first explicit use of the coherency property was in the case of the separated oscillatory field, and split it into two coherent parts and later have them interfere. What you do is you take the oscillator that you have oscillating. You let that oscillate in two regions. If the oscillating field was not coherent; namely, if it oscillated one way here and one way there, you wouldn’t get any resonance. But, since it is coherent, the radiation is coherent, it’s going up and down in the same way (phase) in the first and second region and you get a sharp resonance. It’s a much sharper resonance because the width of the resonance is determined by the separation of the two oscillatory fields whereas you couldn’t do that with a single oscillatory field. If you’d use a short wavelength and a single oscillatory field, which is Rabi’s method, then you’d be one phase in one part and a different phase in another. When you go from one wavelength to another, one wavelength to the next wavelength in a standing wave, the phases are opposites. One is going up, one is going down. With a separated oscillatory field, you utilized the fact that the phases were coherent, they were limited and they didn’t vary much whereas if you did this with a single oscillatory field the phases were mixed.

Pavlish:

Would you give a little background on the genesis of this idea that you came up with for the separated oscillatory field?

Ramsey:

Yes, I will. As is often the case, it has a slightly peculiar history (It wasn’t directly to use the separated oscillatory field to obtain sharper resonance. It was the following. It was after the war and I’d moved to Harvard. I was setting up my own molecular beam lab. Every former student likes to beat his professor on something and do better still. I decided I wanted to do some of the resonance experiments we had done on H, D, and HD, which were important and I’d like to do them much more accurately. I thought that one way I could do this, was to use slower molecules and to keep it one wavelength long, not to have a mixture and thus lose the resonance, which meant I was limited by the length of the apparatus. But I also had to have the magnetic field uniform in that region. At the radiofrequency it could be a couple of meters long, which would give a good sharp resonance. I knew we had tried earlier with Rabi to do some things with a longer magnetic field, and we didn’t get a sharp resonance. It was worse, not better. I realized that one of the reasons that could be true was if our magnetic field wasn’t uniform enough and I thought that I had a good idea for making the magnetic field more uniform. I was trying to do it, and it didn’t work too well. I was disappointed because that was going to be my big experiment that was going to be better, and it didn’t work. Then, I remembered a thing from my studies at Cambridge. I think it was P. I. Dee taught the course in physical optics. He was later a professor at Glasgow; he was an instructor at Cambridge when I was there. He had a rather dramatic way of describing what’s called the Michelson stellar interferometer. I liked the way Dee described it. To understand the stellar interferometer, you have the following characteristic of optics: if you have a telescope of a given aperture and plenty of light, from the aperture you will get a given sharpness of the pattern.

There’s basically an interference of the light that goes on one side of the telescope lens and light from the other side of the telescope, and those interfere to give the sharpness to the resonance. To get that sharp resonance, they have to be coherent between the two places. Dee said, if you have a situation with this big aperture telescope and you have plenty of light but you can’t quite tell whether you’re looking at a single star or a double star, it’s just below the resolution. If you take a can of black paint, and paint over all the middle of the telescope except for two strips at opposite edges, then the sharpness goes up by a factor of about two. It’s because it’s the greater distance between the two slits in contrast to the distance between the middle of one half the aperture and the middle of the other. I realized that the sharpness of the image could not be destroyed by the quality of the glass painted over and there might be some similar way I could eliminate the effects of the non-uniform magnetic field.

Pavlish:

The poor magnetic field was the equivalent of the unpainted inside of the telescope?

Ramsey:

The poor magnetic field was the equivalent of having bad glass in this form of Michelson stellar interferometer, which clearly wouldn’t do any harm if it is painted over with black. That’s why I said there must be an analogy here. It wasn’t a direct analogy. It took two or three days of thinking. Then I realized that yes it would overcome this non-uniformity problem by having the oscillatory fields only at the beginning and end of the magnetic field region so the non uniformity in between would cause no trouble. That’s what I was aiming for, not to get it sharp necessarily, but just to overcome the non-uniformity. Then, I realized as I was writing up the paper for publication, I realized that it also meant that you could make the space bigger than a wavelength long. At that time I was still thinking about Rabi’s experiment with a single oscillatory field, which meant that the frequency had to be quite low to be able to have a long field. If you have a long field you have more time between the two oscillatory fields so you get a sharper resonance. I realized that the separated oscillatory field this also overcame the problem of having a high frequency. You could have the separation between the two oscillating fields much greater than one wavelength long if you just made sure that the two were coherent. If they were coherent oscillations they went up and down together. You have the advantage of a greater time between the two, to get a much sharper resonance. You still had a good resonance because the radiation was coherent in whatever region the atom was exposed to. Well, it turns out there are several other advantages. The original advantage that I was looking for also worked. It did average the magnetic field. But, it had an even greater advantage in other respects. The most important was the fact that you were not limited by the wavelength. This method in fact also applies to optics and microwave fields. It’s a very valuable technique.

Pavlish:

In masers and atomic clocks?

Ramsey:

Yes. In my work with Rabi, we speculated some about an atomic clock but we felt it was not going to be very good because to get accuracy you had to have a long separation to get a big phase differences in between. But you also had to avoid phase differences along the way. You were stuck at that point with the Rabi method. Until the time I invented the separated oscillatory field, or the Ramsey method, nobody tried seriously to make an atomic clock because it wasn’t any more accurate than a pendulum clock. There was no point of using a new technique when the new one was thought to be no better.

Pavlish:

So the idea existed to make atomic clocks but because of the techniques available, it didn’t make much sense?

Ramsey:

That’s right.

Pavlish:

Then, your technique, which you weren’t doing in motivation to make atomic clocks, you were doing it independently…

Ramsey:

Then I realized that it would make atomic clocks.

Pavlish:

Do you remember when in the course of your research you realized that application? That’s another thing I was wondering. In radar work, of course you were thinking about applications but in your basic research before and after the war were you thinking about applications?

Ramsey:

There are fundamental things you can do with atomic clocks. Testing the theory of relativity, for example, the best tests of the theory of relativity, both general and special, are probably done with atomic clocks. The accuracy of the atomic clocks, until the separated oscillatory field method, was not sufficient to be better than any other clock. You can get very broad resonances and then extremely sharp resonances, a thousand sharp resonances inside the broad reference due to the separated oscillatory field. You can measure one of those very accurately and determine what it is. Then it gives you a thousand or more times greater accuracy.

Pavlish:

Within that broad resonance, wow, that’s amazing.

Ramsey:

It’s not the broad resonance that you are measuring it’s the sharp resonance. Soon after I invented the separated oscillatory field, other people began working on applications. Particularly one of my former collaborators, Zacharias: who by then had gone to MIT. This was after the war. There was quite an interim between the first magnetic resonance experiment and the separated oscillatory field method, due in part to the war.

Pavlish:

You mentioned that after the war a lot of the instruments used in the Radiation Lab were given to physicists. Were any of your instruments a result of that?

Ramsey:

Some, but I tended to concentrate, maybe even excessively concentrate, on radio frequencies. I like to put my reasons for doing so in this way: Radio frequency spectroscopy is a little like measuring the width of the street by measuring the distance from one side of the street to the other. Whereas, doing the same thing with higher frequencies is a little like measuring the distance from one side of the street to 42nd street, and measuring the other side of the street to 42nd, and taking the difference. In fact, that meant that I was able to make a number of important discoveries. Originally, the laser didn’t exist. After the invention of the laser you actually could make a very accurate measurement analogous to going from one side of the street to Times Square and the other side of the street to Times Square.

Pavlish:

One of the themes I’m interested in within the discipline of History of Science is the theory developing about scientific objects. What is a scientific object? How does one think of the atom as a scientific object? Instruments you use are maybe not scientific objects, but maybe they’re technological objects. Do you remember specifically which components of your apparatus were given to you from the lab?

Ramsey:

I would say most of them were not, actually.

Pavlish:

You bought them?

Ramsey:

I was doing most of my work at radio frequencies. That equipment, you could buy at Radio Shack. They were just vacuum tube oscillators. I did use some. I did eventually do some things at higher frequencies. For example, a different atomic clock, invented by Dan Kleppner and myself, the so-called atomic hydrogen maser. For that one we did use techniques that came from the Radiation Lab including some of which I had invented there.

Pavlish:

Techniques as well as instruments?

Ramsey:

They go together. Most were instruments we built ourselves but some instruments we got from military surplus.

Pavlish:

That reminds me, as I was looking at your papers from before the war, it looks like when you were in D.C. as a Carnegie fellow before the war (you were working on different things from what you had done at Columbia) it seems like, on pg 72 of the first volume of your Collected Papers, on pg 227 of the paper, you say that you used the same apparatus as you had used before. I was wondering whether you were going back and forth between D.C. and New York to do some magnetic resonance on the side?

Ramsey:

No, it was due to a time lag on the publication of papers.

Pavlish:

Oh, I see.

Ramsey:

That paper says written by me from Columbia University, although it says, now at the University of Illinois and was published more than a year after I had done this research at Columbia, so it was basically a Columbia paper, and yes we had used one the old apparatuses. One thing that was very pleasant at that time is that we essentially had no competition and therefore was no rush to publish the paper. We tended to be a little bit slow and we had several papers to write up. This paper was my Ph.D. thesis.

Pavlish:

Out of curiosity, when do you decide to publish one of these little notes and when does it become a paper?

Ramsey:

In principle you publish both. Usually, we did. They were abstracts of what was to be a lecture to be given in a scientific meeting. It gave people warning of what was coming up. It also got your claim out; it was very useful. You were supposed to follow it up with a more detailed paper.

Pavlish:

Earlier here, you have a page with four bits like that, all from the Rabi group, on one page.

Ramsey:

At that time, when there was a scientific meeting, they published short abstracts for the meeting.

Pavlish:

You might have given talks on those.

Ramsey:

I gave talks about them. They were also genuine publications. The history of that has changed. At that time, these abstracts for the meetings also were reproduced in the Physical Review, later the Physical Society introduced the so-called Bulletin of the American Physical Society in addition to the Physical Review.

Pavlish:

That’s where those kinds of abstracts were later published?

Ramsey:

Yes.

Pavlish:

Did you give talks before 1940? Do you remember giving any talks on your work? Would that have been for Rabi to do?

Ramsey:

No, it wasn’t just Rabi. We took turns. My first big invited talk was in 1940 at the Seattle Meeting of the Physical Society. Invited talks were longer than contributed talks. I also gave some of these contributed talks fairly early, yes.

Pavlish:

I wonder how much other people who got into magnetic resonance later might have been influenced by your presentations, and how much your papers influenced them.

Ramsey:

They were influenced by both.

Pavlish:

Was it impressive showing a resonance curve on a slide?

Ramsey:

Yes, that’s right.

Pavlish:

Was it the theoretical…

Ramsey:

There were three early things that could be published. One was the abstract for a meeting. Then, you can write a letter to the editor, which had to be very important and very brief — now they tend to be longer, in fact now there’s a separate journal for them. This one, for example [flips Collected Papers Volume 1]…

Pavlish:

Letters. You would also give talks, and the final detailed paper would be the culmination.

Ramsey:

Yes.