Magnetic resonance of a single molecular spin

Abstract

Magnetic-resonance spectroscopy on single molecules represents the ultimate limit in sensitivity of electron spin resonance: the detection of a single molecular spin. This is achieved by combining single molecule spectroscopy and optically detected magnetic resonance. Experimental results on pentacene in p-terphenyl both in zero-field and in the presence of a weak magnetic field demonstrate that magnetic-resonance spectroscopy on single molecules adds on to the specificity of single-molecule spectroscopy. It proved possible to identify single molecules which contain ¹³C nuclei in natural abundance and to observe the splitting of the electron spin resonance line resulting from the hyperfine intercation of a single molecular spin with a single ¹³C nuclear spin. From the Zeeman effect the orientation of the symmetry axes of individual molecules with respect to the direction of the external magnetic field is obtained. This allows to compare the orientation of individual molecules with their substitutional-site specific transition frequencies depending on the quality of the host crystal.

Introduction

The development of magnetic-resonance spectroscopy ranks among the most important advances of physics and chemistry of the last 50 years. Only a few other techniques are that widely applicable and offer such a direct and detailed insight on the atomic and nuclear level. Magnetic resonance deals with the magnetic-dipole transitions between states of atoms or molecules which have non-zero total angular momentum and their interaction among each other with or without an applied magnetic field. Depending on the origin of the angular momentae the terms electron spin resonance (ESR, often called EPR – Electron paramagnetic resonance) and nuclear magnetic resonance (NMR) have been established for the experimental methods. However this is a rough classification given the many experimental schemes where electronic and nuclear moments are involved simultaneously.

The underlying ideas of magnetic-resonance spectroscopy can be traced back to the thirties and were strongly connected with the theorist van Vleck [1] and the experimentalist Gorter 2, 3 who both studied the phenomenon of magnetic relaxation. The first successful ESR experiment was reported in 1944 by Zavoisky [4] soon followed by the first NMR experiments carried out by Purcell et al. [5] and Bloch et al. [6] in 1946. Since then a wealth of experimental schedules has been developed which, owing to the specificity of the nuclear moments, turned out to become one of the most powerful analytical tools to chemists today. Magnetic-resonance spectroscopy is utilised to elucidate topics like the structure of crystal defects, molecular relaxation, or the composition of human tissue in such divergent fields as solid state physics, biology and medical diagnostics 7, 8, 9.

Although widespread, magnetic-resonance spectroscopy suffers from two inherent limitations. A minor one is related to the anisotropy in space created by the application of an external magnetic field. This introduces a dependence of the signal on the orientation of the system under study with respect to the magnetic-field direction. For randomly oriented samples like molecules in a fluid or a glass a powder spectrum results featuring very broad magnetic-resonance lineshapes which reflects the averaging over all molecular orientations. The information that is accessible from such a spectrum is limited as compared to that obtained from a well ordered sample like a molecule embedded in a crystalline host matrix. Nevertheless, in the course of time clever experimental set-ups and the utilisation of pulsed microwave sequences have been exploited to obtain detailed knowledge even in random samples 10, 11.

A much more serious handicap of magnetic-resonance spectroscopy is its notoriously low sensitivity. Dealing with transitions in the microwave (ESR) or radiofrequency (NMR) spectral range, the energy of the involved photons is so small that a large number of spins is required for a detectable signal. Typical numbers are about 10¹⁶–10¹⁸ spins for NMR experiments and 10¹⁰–10¹² spins in the case of ESR. A very successful technique to enhance the sensitivity is based on the idea to transfer the detection of a magnetic-resonance transition to the optical domain with an accompanying increase in photon energy. This method is generally called optically detected magnetic resonance (ODMR). First experiments of this type were proposed by Brossel and Kastler in 1949 [12] and carried out by Brossel and Bitter in 1952 on a gas of mercury atoms [13]. They monitored the fluorescence of the mercury atoms and detected the magnetic-resonance transition between the Zeeman components of the 6 ³P₁ multiplet as a change of the polarisation of the emitted light owing to the Δm selection rule for electric-dipole transitions. At the end of the sixties it was realised that optical excitation of an organic molecule embedded in a solid host gives rise to a considerable spin polarisation in the photo-excited triplet state [14] and it was expected that a microwave field in resonance with a triplet transition should alter the phosphorescence intensity [15]. Indeed these changes are observable and the first ODMR experiment on organic molecules in the condensed phase has been carried out by Sharnoff in 1967. He detected the “Δm=2” transition in the lowest triplet state of naphthalene as a change in the phosphorescence intensity [16], soon followed by the detection of the “Δm=1” transitions in phenanthrene by Kwiram [17] and quinoxaline by Schmidt et al. [18]. In the latter paper magnetic-resonance spectroscopy was performed for the first time without an applied magnetic field leading to the “zero-field” spectroscopy. As demonstrated 1973 by van Dorp et al. [19] the change in fluorescence intensity upon the occurrence of a magnetic-resonance transition can also serve as detection channel which enables the study of non-phosphorescent triplet states (fluorescence-detected magnetic resonance, FDMR). For systems which do not radiate at all it is possible to apply the technique of absorption-detected magnetic resonance (ADMR) which relies on the fact that the absorption or emission of a microwave photon leads to a change in the optical absorption intensity. The first ADMR experiment has been carried out by Clarke et al. [20] in 1972. The attractive feature of all these double-resonance techniques is their high sensitivity. In favourable cases as few as 10⁵ spins can be detected [21].

In the field of optical spectroscopy individual atoms and molecules have attracted considerable attention during the last decade and fascinating experiments – commonly referred to as “Gedanken experiment” in textbooks about quantum mechanics – have become possible. The first experiments on single atoms have been performed as early as 1977 by Kimble et al. [22]. They sent a beam of sodium atoms through a small aperture and collected, after laser excitation, the induced fluorescence with a microscope objective from a volume restricted to about $\text{10}{}^6\text{μ}$m³. The incident atomic current was adjusted such that only one atom at a time was present in the sample volume. This way they were able to observe photon-antibunching usually concealed when averaging over the emission of many atoms. However, their signal was the signal of many molecules – studied one by one. The observation of photon-antibunching and quantum jumps of an individual quantum system were made possible by storing single atomic ions in an electromagnetic trap 23, 24, 25, 26, 27.

In liquid media it proved possible to detect individual molecules by focusing the output of a laser onto a capillary and confining the excitation light to a tiny volume of a few $\text{10}{}^3\text{μ}$m³. Similarly as in the experiments of Kimble et al. the flow in the capillary was adjusted such that only one molecule was present at a time in the excitation volume 28, 29, 30, 31, 32, 33. This type of spectroscopy has become an important tool mainly in biology related studies like DNA sequencing 34, 35 and biochemical analysis 36, 37. Based on microscopic techniques it became possible to observe single DNA molecules tagged with fluorescent dye [38], and even the diffusion of individual lipid molecules through a phospholipid membrane has been followed 39, 40.

The pioneering work of Moerner and Orrit during the last couple of years formed a landmark to obtain information about the solid state in a completely new way by studying, at cryogenic temperatures, single impurity molecules embedded in a host material 41, 42. This was achieved by confining the excitation light to a small spot and taking advantage of the fact that the absorption frequencies of the guest molecules are spread out in frequency due to slight differences in their microscopic environment. Using a laser with a spectral bandwidth much narrower than this distribution of absorption frequencies they observed, independently, the emission of individual impurities. This enabled the investigation of host–guest interactions on a truly local scale, where the averaging over many centres is removed 43, 44, 45, 46, 47, 48.

In parallel a completely different type of spectroscopy was developing intended to achieve superior spatial resolution – scanning near-field optical microscopy commonly abbreviated as SNOM (Europe) or NSOM (USA). It is based on the idea to use an aperture as light source with spatial dimensions smaller than the wavelength of the used radiation and to overcome the (far-field) diffraction-limited resolution, a method suggested as early as 1928 [49]. A first experiment which used this principle was carried out in the microwave region by Ash et al. [50]. In the optical domain SNOM achieved a breakthrough in 1993 when Betzig et al. used glass fibres tapered at their end to diameters of 50–100 nm [51]. To make use of the small aperture as light source the sample has to be very close, in the near-field, of the tip which was realised by a special distant regulation mechanism – shear force detection 52, 53. With near-field spectroscopy it became possible to perform studies like observing single molecules on a substrate 51, 54, detecting the emission of the quantum constituents of a semiconductor quantum well [55], or to study a single human chromosome [56]. Reviews of the near-field work can be found in 57, 58, 59, 60, 61.

Single molecule spectroscopy in general has become a flourishing field of interest and it is impossible to keep track of the literature conscientiously. I have to admit that the examples given above are chosen subjectively to illustrate the potential of these newly developing techniques. The selection of topics is far off from a complete overview and, as can be expected from the rapidness of the progress in this field, it is not up to date when this is printed. A much more detailed bibliography is available under http:⧹⧹www.chem.ethz.ch/sms/html/biblio.html on the Internet.

A question that naturally arises when considering the success of both magnetic resonance and single-molecule spectroscopy is whether it is possible to make the methods of magnetic resonance accessible for single-molecule spectroscopy. This demands that single-molecule spectroscopy will lead to a single-spin spectroscopy. In other words: Will it be possible to reach the ultimate limit in sensitivity of ESR – the detection of a single molecular spin? The answer is affirmative. Two independently working co-operations performed this experiment in 1993 62, 63 on single pentacene molecules embedded in a p-terphenyl host crystal thereby demonstrating that the sensitivity of (optically detected) magnetic resonance is limited only by the photophysical properties of the system under study. The single chromophores were isolated as in the work of Moerner and Orrit and the magnetic-resonance transition of the single molecular spin was detected by FDMR. Magnetic-resonance spectroscopy allows new information to be obtained in a way which is not available by pure optical means. From the point of view of magnetic resonance single-spin spectroscopy offers, besides the enhanced sensitivity, the possibility to study molecules in random samples as if they were in a single crystal. Studying the molecules one by one removes the broadening of the spectra common to powder spectra which results from averaging over many differently oriented molecules.

This contribution details the efforts to detect the magnetic-resonance signal of a single molecular spin. Apart from the very first experiments¹ the work described has been carried out at the “Centre for the Study of Excited States of Molecules” at the University of Leiden in the Netherlands. After a general introduction of the methods of magnetic-resonance spectroscopy and especially optically detected magnetic resonance the topics of optical single-molecule spectroscopy, necessary for the selection of a single molecule, are described. Details of the experimental set-up are given in a separate section subsequently followed by the experimental results on single pentacene molecules embedded in a p-terphenyl host crystal. A “spin-off” section describes results on small ensembles made possible “for free” by the high sensitivity. Finally an “outlook” section gives the space for speculations.