1986 Nobel Prizes Development of Scanning Tunnelling Microscopy A. Baratoff, Zürich (IBM Zürich Research Laboratory)

Heinrich Rohrer (left) and Gerd Binnig (also winners of the 1984 EPS Hewlett-Packard Europhysics Prize).

On October 15, 1986 the world learn­ Both instruments rely on electrons ed about the award of the Nobel Prize in and are limited by quantum properties of Physics to Ernst Ruska for his funda­ the latter, but have otherwise little in mental investigations in electron optics common. The present contribution fo­ and the construction of the first elec­ cusses on the development of STM by tron microscope, and to Gerd Binnig and its inventors. A similar article on electron Heinrich Rohrer for designing the scan­ microscopy will appear later. Before ning tunnelling microscope. reviewing the main stages in the evolu­ Ruska, now 79, obtained satisfactory tion of the STM (see also Ref. 2), let me images with his two-stage instrument in describe the physics and operation of 1931 and surpassed the resolution and that conceptually simple instrument. magnification of light microscopes two The state of the art, including complete years later while working at the Techni­ references, is covered in two recent Pro­ cal University in Berlin 1). Since then ceedings 3·4). electron microscopes with their diffe­ rent imaging modes and modifications, Basic Principles have slowly matured and gained accep­ The key to STM is to approach a pro­ tance as indispensable tools in physics, bing tip, made of a refractory hard metal Fig. 2 — Distribution of tunnelling current materials science, microelectronics, bio­ like tungsten, within a few A of the sam­ density in energy (top) and space (bottom). logy and medicine that enable structural ple to be investigated so that a measu­ (Ref. 3), Copyright 1986, IBM. analysis and selective manipulation on a rable current IT (typically 1 nA) flows in scale of few µm down to a few A (under response to a fixed voltage VT ranging in 1960, the new technique relies on tun­ special conditions). It is therefore grati­ from a few mV for metals to several volts nelling through vacuum or an inert me­ fying that the Prize Committee decided for poor conductors. The tip is then dium, like liquid He, thus permitting in to honour Ruska's pioneering work, thus scanned along the surface at a constant situ investigations of local surface chan­ recognizing what The New York Times distance s (to zero order) by means of a ges induced by external means, or by the referred to as a "microscopic oversight". feedback control unit that maintains IT tip itself. Recently STM images of relati­ By contrast, the Scanning Tunnelling constant. As sketched in Fig. 1, both fine vely inert surfaces like gold and graphite Microscope (STM), first successfully approach and scanning are effected by have been obtained in air and water3·4). operated only five years ago, soon prov­ calibrated piezoceramic rods producing To obtain stable images, differing only ed able to resolve atomic-scale features displacements of several A/V. The z-dis- by small drifts, required much ingenuity on the surfaces of even poorly conduc­ placement (proportional to V recorded and perseverance. Mechanical vibra­ ting materials. Although it is only begin­ along successive scans yields a nondes­ tions and noise had to be suppressed. In ning to have an impact beyond funda­ tructive topographic profile (dotted line) the 1970's a few groups tried, but none mental science, the tremendous poten­ provided the tip neither accidentally tou­ could reach the goal of controlled ap­ tial of STM (depending on the context, ches the surface nor induces local fields proach, vacuum tunnelling and scann­ M means either microscope or micro­ causing atoms to jump. ing. The specific designs discussed be­ scopy) has been realized by the Prize In contrast to experiments with solid low all rely on an isolation and damping Committee at a relatively early stage. (e.g. oxide) barriers, initiated by Giaever system against external disturbances and on a coarse approach mechanism that brings the sample within the work­ Fig. 1 — Principle of STM: The ing range (100 A to several pm) of the probing tip (black) is actuated by piezodrives. The influence of internal voltages independently applied vibrations generated by the motion of to three mutually orthogonal piezodrives. While the tip is the tip is avoided by setting the cutoff scanned along the sample by frequency of the feedback well below Vx, Vy, the control unit, CU, the lowest mechanical eigenfrequency records and supplies the voltage of the tip-scanning unit. Vz required to keep the tunnell­ Parameters determining the tunnelling ing current IT (VT) constant. current distribution, and hence the sen- 141 sitivity and resolution of STM, are illu­ strated in Fig. 2. Assume for simplicity that the barrier is vacuum and that both sample and tip are good conductors. The applied voltage VT then appears across the separating potential barrier, and the height of the latter is approximately the average Φ = (Φ1 + Φ2)/2 of the tip and sample work functions. Its thickness s is the distance between the turning points of electrons at a given energy within the window defined by the corresponding Fermi levels EF1, EF2 (assuming rapid equilibration within each electrode and negligible thermal excitation above the barrier). If VT << Φ1, Φ2, the distance dependence of IT is dominated by the exponential decay of the transmission Fig. 3 — Schematic diagram of ''pocket-size" STM. (Ref. 9), Copyright 1986, the American coefficient at EF1 ≈ EF2. For free elec­ Institute of Physics. trons tunnelling between parallel elec­ trodes, the current density would be phy that are difficult to recognize unless ed LEED and Auger surface diagnostics.) j ≈ (e2/hπ) (k/s) VT exp(-2)ks), (1) Φ is simultaneously recorded. As in later designs, including the com­ where h/e2 = 25.8 kΩ and pact one sketched in Fig. 3 9), a "louse" 2k(Å-1) = 1.025V [Φ(eV)]. Experimental Aspects: Early Work consisting of a piezoplate (2) on three For a curved tip IT is approximately ob­ The idea of STM arose in 1978 during metal feet (3) separated from a ground tained by integrating (1) laterally about a visit by Binnig to discuss research plate by high dielectric constant in­ the closest separation s, thus giving an plans with Rohrer while he was writing sulators ensured coarse approach in effective tunnelling area of 2V (r/K) dia­ up his doctoral dissertation. After settl­ steps of 100 A to 1 µm via sequential meter for a tip of radius r at its apex. This ing in Zurich in the fall, Binnig started clamping of the feet in harmony with estimate of the lateral resolution L, e.g. designing and testing various compo­ contractions and elongations of the 50 A for even a smooth tungsten field nents with his uncanny ability to see the body (2). Better procedures for obtain­ emission tip with r = 1000 A and (p = essential. He was helped by Rohrer's for­ ing stable minitips in situ were develo­ 4.5 eV, looked promising enough and mer experienced technician, Christoph ped, e.g. applying up to 100 V for a cer­ motivated Binnig and Rohrer to build Gerber, who has since then instructed tain time. their first STM 5, 6). Their ability to ob­ hundreds of visitors and newcomers to At his first invited talk at the German serve monoatomic steps with apparent the field. Binnig and Rohrer still pursued Physical Society meeting in April 1982 widths < 10 A led them to conjecture their previous research interests, but the Binnig claimed that ≈ 3 A wide features that the ground tungsten wires they situation changed in autumn 1981 when occasionally appearing in adjacent used for tips supported small clusters of a working STM design took shape and scans were single adsorbates. Soon atoms. The extreme sensitivity of IT to s an eager young technician, Edmund afterwards individual close-packed naturally selects the "minitip" closest to Weibel, joined the group. [001] atomic rows on the reconstructed the sample 6): the dream of resolving Initially magnetic levitation above a Au(110) surface could be imaged. This single surface atoms (schematically in­ superconducting lead bowl coated with observation prompted theorists to get to dicated by circles in Fig. 1) suddenly aluminium provided vibration isolation work and provided a basis for testing seemed within reach! Subsequent theo­ and eddy-current damping. A linear estimates of the resolution. ries demonstrated that L is of order 1.6 V decay of log IT vs. s was observed after a But it was the first STM image of the [(r + s)/k], but can also be smaller if, for heuristic cleaning procedure, with

Switzerland. rience already acquired, while other during Binnig's sabbatical at Stanford 13) groups started building STM's on their own. Joint projects financed by IBM Eu­ rope and the institutions concerned were established with universities in Experimental Nuclear Physicist Madrid (Autonoma), Marseilles (Lu- There is a vacancy for an experimental physicist to join the miny), Berne and Zürich (ETH). Sessions Nuclear Structure Division at Daresbury Laboratory, an establishment of the Science and Engineering Research Council and symposia on STM started being which operates major national facilities for scientific research. regular features at scientific meetings. The group is involved in carrying out and supporting a research The fruits of the first collaborations programme on the Nuclear Structure Facility (NSF), a large tandem accelerator which is operating at up to 20 MV on terminal obtained with the third STM are sum­ The successful applicant will be required to organise and direct marized in several reviews 11 8). They in­ work necessary for the operation, maintenance and development clude observations of single adsorbed of equipment used primarily in gamma-ray programmes aimed at studies of nuclear collective motion, nuclear reactions and nuclei oxygen atoms on Ni(TIO), of the subtle far from stability. Candidates will be expected to carry out large-scale reconstruction of Au(100), of research in conjunction with university users on the NSF and to play the spiral structure of DNA and of the an active role in initiating and developing new programmes. Other duties will involve the overall coordination and scheduling of the neck of a virus. Somewhat later local scientific programme and the assessment of future needs for tunnelling spectroscopy confirmed the instrumentation and other facilities on the N.S.F. existence of both occupied and unoc­ Applicants should have a good honours degree (or equivalent qualification) in physics and several years postgraduate cupied states (depending on the sign of experience in experimental nuclear research. A Ph.D qualification VT) confined in front of or close to in nuclear physics and a period of relevant post-doctoral various surfaces, while scanning reveal­ experience would be an advantage. The appointment will be made in the grade of Senior Scientific ed that some of the latter were also Officer on a salary scale of £10,568 to £13,762 per annum; the localized laterally, e.g. where unsatura­ starting salary will depend on qualifications and experience. ted bonds are expected between ada­ The superannuation scheme is non-contributory. CLOSING DATE: Friday 13th February, 1987. toms on 7 x 7 Si(111). These results were Application forms may be obtained quoting reported at a workshop in Oberlech in ref DL/974 from - The Personnel Officer, July 1985 where about twenty different Science and Engineering Research Council, Daresbury, Warrington WA4 4AD. groups or investigators working in STM Telephone (0925) 603467. reviewed their progress 3). Subsequent Further information is VT -dependent studies by two groups at available from Dr J S Lilley on IBM Yorktown Heights clarified the (0925) 603558. DARESBURY SCIENCE & ENGINEERING RESEARCH COUNCIL nature of such surface states for the 143 DTH THE TECHNICAL UNIVERSITY OF DENMARK Associate Professor in Experimental Surface Physics

At the Technical University of Denmark a new Surface Science Research Group is being created. When fully staffed it is planned to Fig. 5 — Real-time STM image of cleaved consist of 3-4 faculty positions. The first of these, at the associate graphite. Minima are only 2.42 A apart. professor (lecturer) level is now open at the Laboratory of Applied Physics II. and the atomic resolution studies of cleaved graphite performed before his Applicants are expected to have several years of post doctoral departure with the STM shown in Fig. 3. research experience with surface spectroscopies, preferably also A stack of stainless-steel plates separa­ with some applications to industrial problems. ted by Viton dampers provided enough The applicant must be able to undertake teaching in elementary vibration isolation for stable operation in and advanced physics courses. air. When placed into the UHV chamber The successful applicant is expected to play a major role in the build of a scanning electron microscope, this up of the surface physics groups and the acquisition of equipment. compact instrument could produce Further information can be obtained from Dr. Steen MΦrup, Labo­ traces of excellent clarity 9) and the cor­ ratory of Applied Physics II, Building 307, The Technical University responding image clearly shows minima of Denmark, DK-2800 Lyngby, Denmark. Tel. (45) (2) 88 24 88, with the expected spacing. Although only every second surrounding carbon ext. 2321. atom appears as a slightly shifted maxi­ Curriculum vitae including relevant publications, if possible in three mum in that picture, Binnig and cowor­ copies, and an account of teaching experiences are to be directed kers very recently managed to image all to: six, albeit in a narrow range of current, Faculty of Basic Sciences, Building 101 as shown in Fig. 5. This photograph was Technical University of Denmark, taken in a few seconds direct from an DK-2800 Lyngby, Denmark oscilloscope driven by a fast-scanning STM operating in liquid Fie similar to one not later than 6 January, 1987. that was built in one day in Stanford 4).

REFERENCES 5. Binnig G., Rohrer Fl., Gerber Ch. and 10. Binnig G., Rohrer Fl., Gerber Ch. and Weibel E., App/. Phys. Lett. 40 (1982) 178. Weibel E., Phys. Rev. Lett. 50 (1983) 120. 1. Ruska E., The Early Development of Elec­ 6. Binnig G., Rohrer H., Gerber Ch. and 11. Binnig G. and Rohrer Fl., Physica 127B tron Lenses and Electron Microscopy (S. Weibel E., Phys. Rev. Lett. 49 (1982) 57. (1984) 37, and in Trends in Physics 1984 (J. Hirzel Verlag, Stuttgart) 1980. 7. Binnig G. and Rohrer H., Surf. Sci. 126 Janta and J. Pantoflicek, eds. EPS) 1985, p. 2. Quate C.F., Phys. Today 39 (August 1986) (1983) 236. 38. 26; Salvan F., La Recherche 181 (1986) 8. Binnig G. and Rohrer H., Sci. Am. 253 12. Becker R.S., Golovchenko J., Higashi 1202. (August 1985) 50. G.S. and Swartzentruber B.S., Phys. Rev. 3. IBMJ. Res. Develop. 30 (1986) 355, 430. 9. Gerber Ch., Binnig G., Fuchs H., Marti 0. Lett. 57 (1986) 1020 and references therein. 4. Proceeding of the STM'86 Conference, and Rohrer H., Rev. Sci. Instrum. 57 (1986) 13. Binnig G., Quate C.F. and Gerber Ch., Surf. Sci.. to be published. 221. Phys. Rev. Lett. 56 (1986) 930.

EPS Divisions. Sections and Group Europhysics News is the official journal of the European Editor: E.N. Shaw Astronomy and Astrophysics Division Physical Society which comprises 29 National Socie­ Solar Physics Section ties, Academies and Group, about 4000 Individual Editorial Board: Atomic and Molecular Physics Division Members and 70 Associate Members. Governing bo­ K. Appert, A. Baratoff, F. James. Atomic Spectroscopy Section dies of EPS are the General Meeting, Council and an M. Lehmann, M. Mayor, J. Muller Chemical Physics elected Executive Committee responsible for detailed Electronic and Atomic Collisions policy. EPS promotes the collaboration of physicists Molecular Physics throughout Europe, organising and harmonising confe­ Editorial and Advertising Office at the EPS Computational Physics Group Secretariat Condensed Matter Division rences and publications, improving physics education, Liquids Section encouraging physics applications, awarding scholar­ Low Temperature Physics Section ships to sponsored schools in Erice. EPS publishes in Address : EUROPEAN PHYSICAL SOCIETY Macromolecular Physics addition to Europhys. News, Europhysics Letters (in P.0. Box 69, Magnetism partnership with national societies), European Journal CH-1213 Petit-Lancy 2 Metal Physics of Physics (in partnership with The UK Inst, of Phys.), Switzerland Semiconductors and Insulators Europhys. Conf. Abs., Europhys. Ed. News. Individual Surfaces and Interfaces Members receive Europhys. News free of charge (price Telephone: Geneva (22) 93 11 30 High Energy & Particle Physics Division to insts.: Sw.Fr. 90.-/a), Europhys. Lett, at Sw.Fr. 70.-/a Telex : 428 024 eps ch Interdiv. Group on Exptl. Phys. Control System: (insts. 595.-), rebates on many other publications and Interdiv. Group on Physics for Development Cables : europhys genève Nuclear Physics Division on conference fees. Annual EPS membership fee for In­ Optics Division dividual Members belonging to an EPS member society Plasma Physics Division is: Sw.Fr. 44.-; independent members: Sw.Fr. 132.-; Printed by: Pfirter frères sa Quantum Electronics Division members of a Collaborating Society: Sw.Fr. 55- ($ 33). CH-1213 Petit-Lancy/Switzerland

144