HIGHLIGHTS tion of benzene consisted in his recognition that the second half of the double bond. unlike the single bond. I S a nomadic entity which cannot always be assigncd a unique position in the molecule. ... The electronic theory of valencq ... shed a flood of light on structural chemistry and revealed the fundamental distinction between 1oc:tlized and non-localized electrons. a ) C . A. Russell, Tlrc H ~ ~ / ~ JI bw l e io i r /I., Humanities. Neu York. 1971; b) F. G. Ariidt. in a letter see: E C'anipaigne. .I C h m . E h c . 1959.36. 336; c) H . C. Longuet-Higgens. Proc. C ' / i ( w i . Soc., London. 1957. 157. 1. H . b m ' t Hot'f, Dir AniiPcs (/mi.\ I'llitroire rlioie T / i k ) r t c . P M . Bazebdi.ik. . Oxford. 1891 Rotterdam. 1887: CIiciiii.s/ri (17 S p a c ~ Clarendon. J. E Marsh, P/ii/u.\. M O R . 1888. 76. 426. See. L. Pauling. 3Ltifin.e of i h ('iieii7icoi ~ h i i d , 3rd cd.. Cornell University Press. Ithnca. NY. 1964. chap. 14. G. Reddelien. J. Pro!,/ Chrrri. 1915, 213. a ) A . A. Lwine. A . G. Cole, J A m . Clion. Sor. 1932.54. 3%: b) similar results w i - e 1;iter published. see: J. P. Wibnut. P. W. Haayman. Siirnce 1941. 94. 49. a ) L Pauliiip. J .Am. C h o i i . Soc. 1931. 1367; b) L. Pauling. ;bid. 1932. 54. 988: c) L Pauling. Q. W. Wheland. J. Clirwi. P/IJ.\. 1933. I . 362. L. E. Sutton. L. Pauling, T,.nii.\. I.irrar/u? Sol. 1935. 31, 939. Thl\ di\tinctioii of single versus double minima is clearly stated by Sutton and Patiling. " I t appe;irs very definitely that benzene is a single molecular species and ... ueciiii say that it IS ii hybrid ofthe two Kekule structures ... and that the angles between pairs of external valencies are all 60 . ... Our procedure here IS different from that of Mills and Nixon. for they assumed that the niolecule had either structure I or structure 11, whereas we consider that initially i t has the symmetrical resonating structure and then calculate how much it 15 distorted when a ring is attached." [I 71 N. Frank. J. S. Siegel. unpublished. 1181 R. Boese. D. Blhser. W. E. Billups, M. M. Haley. A. H. MaulitL, D. L. Mohler. K. P. C. Vollhardt. A n x m ~Clirni. . 1994, 106. 321 -325. Ati,ccu.. C h m Irir. Ed. Eiigl. 1994. 33. 313-17. [I91 A. Stanger, J. A m Clicni. Soc. 1991. I I 3 . 8277. [ X I K. K. Baldridge, J. S. Siegel. J. A i t i . Chmi. Snc 1992. 114. 95x3.  W. Natakul. R . P. Thummel. A. D. Txggart. .lAi7i. C ' / i c i i i . Sot. 1979. I O i . 770. [I21 a) R. Neidlein. D. Christen, V. Poignee. R. Boese. D. Bliser. A Gieren, c'. Ruiz-Perez, T. Hubner. A i i g w . Chwii. 1988. 100. 292-293, Aiigrw. C/I~WI. In/. Ed. Eiigl. 1988. 27, 294-295: h) R. Boese, D. Bliser. ihid. 1988. /Of). 293 -295 and 1988. 27. 304-305.  I n this regard. perfluoro-1 serves as a good model for these strained iiionocqclic annelated bemenes in general. see: R. P. Thummel, J. D. Korp. I . Bernal. R. L Harlow, R. L. Soulen. J. Am. C h i ? .Soc. 1977, YY. 6916.  M. Saunders. M. R. Kates. J A m Clion. Sue. 1980. 102, 6867.  H. C. Longuet-Higgins. C. A. Coulson. Trmi.\. Frtrurlri?, So<.. 1946. 41. 746. Detection and Imaging of Single Molecules by Optical Near-Field Microscopy Thomas BaschC ,.. . . we or niolrcwle. never experiment with .just one electron or aton? In tliocight esperinients u'e sometiiiies tissunie we do: this invariably entails ridiculous consequences. '' Erwin Schrodinger, 1952 The successful storage and spectroscopic characterization of individual atoms in electromagnetic traps and the development of microscopic techniques with atomic resolution such as scanning tunneling microscopy (STM) and atomic force microscopy (AFM) are two examples that make us realize that Schrodinger's notion has been outpaced by reality. One should. however. bear in mind that the observation of single particles, as commonplace as it may seem today, has only become possible within the past 10-15 years, regardless of the technique. This Highlight deals with new developments in the field of optical detection of single molecules: other methods of detection are not covered. The first optical detection of a single molecule in the condensed phase was achieved in 1989 with the spectroscopic characterization of an individual pentacene molecule in an organic mixed crystal at low temperatures ( T = 1.5 K)."' Not much later. and independent of the research on solids. efforts to detect individual dye molecules in a liquid were successful.[21 A single molecule in a solid at low temperature is a very sensitive spectroscopic probe for the properties of its local envi[*I Dr. T Basche lnstitut fur Physikalische Chemie der Universitit Sophimhtrasc 11. D-X0333 Munchen ( F R G ) Telel'ax. Int. code + (89)590-2602 ronment, owing to its extremely narrow optical linewidth. In contrast, for single-molecule detection (SMD) in liquids or, more general, for all techniques that can be performed at room temperature, analytical aspects are as important as the measurement of spectroscopic details. S M D represents the ultimate limit in chemical trace analysis and has been proposed as a realistic method for sequencing the DNA of the human genome within a reasonable period of time. In molecular biology the detection of single fluorescence-labeled species (proteins, viruses) is a foreseeable and increasingly important development. Optical spectroscopy of single molecules has already given rise to a host of new experimental results. including the development of techniques for the optical characterization of individual molecules with ever increasing sensitivity. Ultrasensitive fluorescence detection is ideally suited for the optical observation of single molecules. The experimental challenge here lies in the discrimination of the weak fluorescence of a single molecule from the matrix background light (Rayleigh and Raman scattering, matrix luminescence). Since the signal of the molecule is proportional to the intensity of the incident radiation (poweri area), and since the background light and the concomitant background noise level increase with the power, enhanced sensitivity is achieved by decreasing the area of illumination. Consequently, tight focusing of the exciting laser light is essential. Focusing with lenses, however. invariably involves the fundamental Abbe diffraction limit to resolution. which states that foci of dimensions less than the wavelength of the light used are not possible. Fortunately. the method of near-field scanning optical microscopy (NSOM) developed during the last few years HIGHLIGHTS circumvents this limit and is therefore ideally suited for fluorescence detection of single molecules. Recently, using this new microscopic technique. Betzig et aLi3' and Ambrose et al.i4]independently succeeded in imaging individual dye molecules on the surface of a substrate at room temperature. Before these experiments are described, let us first look at the principles and potential applications of near-field optical microscopy. Conventional optical microscopy has a limited resolution of about i.!2 of the optical wavelength employed. This theoretical resolution limit is a consequence of diffraction effects due to the wave characteristics of light; for green light under ideal conditions this limit is roughly 0.3 pm, and with good optical microscopes it can in fact be reached. (However, even in confocal microscopy resolution better than 2 / 2 can be achieved using special "tricks".) When light is passed through a small aperture. whose diameter is less than the wavelength. the dimensions of the light beam at a distance close to the aperture are nearly unchanged. This area is called the optical near field. Such an aperture can be scanned raster fashion across the surface of a close-lying object. Many recorded points can then be combined to reconstruct an image of the surface. The resolution of this microscopy is not limited by the wavelength of light, but only by the size of the aperture and its distance from the object. Near-field microscopy in the visible regime was first demonstrated independently by Pohl et al.[5' and Lewis et al.'" Since then, several optical near-field probes have been developed. In most cases the light is directed from the light source to the aperture by a some sort of wave guide. To collect images quickly and with a good signal-to-noise ratio. it is of special importance to pass as much light as possible through the miniscule aperture. To this end, the tips of tapered optical fibers and micropipettes with dimensions on the order of 10 nanometers turned out to be ideal apertures. Other aperture geometries have also been proposed and were recently successfully employed, for example, a partially metal-coated tetrahedral tip.''] The aperture can also be used to collect light from the object and to scatter evanescent waves from the near field into propagating modes of the optical wave-guide. We will next describe a schematic setup of a near-field microscope in the so-called illumination (or transmission) mode (Fig. 1 ) . Here the tip of a tapered optical fiber is used as an aperture and scanned raster fashion across the object of interest. The light is coupled into the fiber through an objective. The fiber tapers from the original fiber diameter to a tip having a flat face perpendicular to the fiber axis. The tip is produced by simultaneously heating the fiber with a CO, laser and pulling the ends apart. This process leads to apertures from 20 to 500 nm with high reproducibility. Finally, the tip is vapor-coated with a metal film (Al) to prevent the light from leaking out at the sides. The sample is placed within about 10 nm of the fiber tip and is moved in the X,J direction with a piezoelectric transducer. The signal is detected at each point, and the image is generated pixel by pixel. During the scan the distance between the fiber tip and the object must be kept constant by using an active feedback mechanism. An elegant approach to this is shear force feedback detection (not included in Fig. I), in which the fiber tip is excited at one of its a high-quality mechanical resonance frequencies. The damping of the vibration as the object approaches the tip is then recorded as a distance-dependent detector 0 I/ piezelectric 11 I tranducer I1 /l/::::::::I:-:DA Al coating Optical fiber objective Fig. 1. Schematic setup o f a scanning near-field microscope The inset (middle left) shows the fiher tip with the ncar-tield and far-field regions For details Fee text signal. A lateral resolution (s,~)of about 12 nm (or 4 4 3 ) could be demonstrated with a setup similar to that shown in Figure The best resolution currently possible in NSOM is roughly two to three orders of magnitude less than that of other scanning microscopic techniques like STM and AFM. Yet NSOM offers some highly interesting possibilites: 1) All contrast mechanisms known in conventional optical (far-field) microscopy can be employed. e.g. polarization. refractive index. and fluorescence. 2) NSOM is possible under ambient conditions, and light has only a weak disturbing effect on samples an important prerequisite for the study of biomolecules. 3) Specific interactions in the optical near field (evanescent modes) can, for example, lead to local amplifications of the fluorescence. 4) NSOM can be combined with other optical spectroscopies without loss of its high local resolution. A good review of the latest, in part spectacular research in near-field microscopy, frequently conducted on biological systems. is presented in the conference proceedings'"' of the Second Near-Field Optics Conference. Raleigh. NC, USA, 1993. In addition to NSOM images of blood cells or DNA, images of, for example, quantum layers are shown. First investigations on magnetooptical data storage with NSOM have yielded storage densities of 7 Gbitcm-2 at a local resolution of 30-60 nm.('O1 Figure 2 shows an NSOM image (Betzig et al.) of lipophilic carbocyanine molecules spread with high dilution (about 23 molecules per pm2) over a thin poly(methy1 methacrylate) (PMMA) film.['] High dilution ensures a mean distance between the dye molecules which is larger than the diameter of the aperture- this condition is necessary for the distinction of individual molecules. The setup of the near-field microscope was as shown in Figure 1 . The excited, fluorescing single molecules, which appeared as microscopic points of light. were detected with a lateral resolution of 50 nm and a vertical resolution of 5 nm. The view that the emissions indeed arise from individual molecules is supported by a number of arguments. Most convincing is the fact that photoinduced bleaching, which happens to some of the molecules after longer illumination. does not lead to the gradual decrease of the signals, but to the discrete and total extinction of individual emissions. The very good signal-to- HIGHLIGHTS basically the same NSOM setup.[41An interesting result of their investigation is the observation of reversible photobleaching of individual molecular chromophores. The authors attribute this effect to hopping motions that change the molecular orientation and consequently the excitation probability of the molecules. In a sense. this implies that a photoinduced reaction of a single molecule ;it room temperature has been obscrvcd. We can expect that near-field microscopy will lead to many tkscinating experiments with individual molecules in the near futurc. These include investigations on the conditions for lightinduced photobleaching of individual chromophores. an unwanted effect that limits all sensitive fluorescence measurements at room temperature. Time-resolved measurements after excitation with picosecond pulses or by analysis of the fluorescence autocorrelation function are also just being tackled.l"] New perspectives of special interest are offered by investigations of single molecules at low temperatures with a combination of NSOM and optical spectroscopy. Here. an extremely high spectral resolution can be used additional to the high local resolution in NSOM. Gcrm'iii \ersioii: I'ig 2 Ncar-lield inicroscopic pic1 tirc of Iluorescsnt \inple carhocq.iniiie d > c inoleculer on the \urf;ice o f a P M M . 4 suhstrate (Trom rel: [ 3 ] ) noise ratio in Figure 2 is attributed to the low background signal in the NSOM technique; a s already discussed. only ii very small sample volume is illuminated with high intensity. It is not yet clear if fliioresceiice amplification by evanescent fields in the optical near field contributes to the signal. At closer inspection. different geometric shapes of the structures in Figure 2 are apparent. They result from different orientations of the molecular transition dipoles. From an analysis of these shapes, the authors could determine the arrangement of the individual molecular transition dipoles, in other words of the individual molecules. on the surface. Moreover, by rcverse reasoning, Betzig et al. were also able to use the molecular point dipole to map the optical near field of their fiber tip. Ambrose et al. succeeded in imaging the fluorescence of single rhodamine 6G molecules on a silica surface using ,-lii~y(,ii. C/iciii 1994. 106. 1x05 [ l ] W E Moernei-. L. Kador. /'/I> 5. R<,.r..Leri. 1989. 62. 2 5 3 5 : M Orrit. J. Bernard. i h i d 1990. 65. 2716: W E. Moerner. T.Bascht. . A n , q w . C/XWI. 1993. f05. 537: A i i , z r i ~ Chmii. l i i r . Ed. Eii,ql. 1993. 32. 457. [ 7 ] E . H. Shera. N . K . Seitringer. 1.M . Daviz, R. A. Keller. S. A. Soper. C i i t w Phi \ Lcrr. 1990, 174.553. R Riglcr. .I.Widsngren. U M r ~ in s F/iiorwm[c, . S p i ~ i r o \ u i / .~ Vcii j ,Mc2//rod\ rriii/ . 4 j ~ p / i c ~ i i o(Ed.. ii\ 0. S. N'ollbeis). Sprinzer. Berlin. 1993. p I 3 [ 3 ] E Betrig. R. J. Chichcstcr. S<.ic.ii<.c1993. 3 2 . 1422 . C. Marlin. R . A . Kellcr. f ' h >. Rev. Lei!.  U P.Amhi-ose. P M. G u o d ~ i i i J. 1994, 72. 160. 151 L). M.: Pohl. W. Dcnk. M. Limy. ,-1pp/ P / i v Leii 1984. 44. 651. . Haroo~tiiiian,M . Muray. L ; / i , . i ~ i , i i ~ , v , \ [ i , / 1984, ~~' 13. [h] A Lewis. M I s a a c ~ m A. 777.  U. C . f-ischer i n ,Liwi. Pic/</O/i/ic\ (Eds.: D. W Pohl. D. Courjon). Kluwer. Dordrecht. 1993. p. 255. [XI t Betrig. J. K . Tratitmnii. T D Harris, J. S. Weiner. R. L. Kostelak. S<.ir.ii[i, 1991. 3 1 . 136X [')I See k i ~ i i i r / S e i i r - ~ i i ~ / ~5 /~O' ~~ ii ii ii /~~ ~ r(Releizh. < ~ i i c e NC. October 1993). special issuc i ' / / r i i i 1 7 i [ i - o \ i [ ~ / ) l1994. ' 111 prezs. [lo] E Betrig. J K . Tr'iutinan. R. Wolfe. E. M .Gyoi-gq. P L Finn. 41. tI Krqder. C.-H. Chang. A p p ~ .Phi.\. LPII.1992. 6 1 , 142. [ I 1 1 See. f o r example. M P.Aiiibrose. P. M . Goodwin. 1 C. Martin. K. A . Keller. S(irI7r~l~ 1994. Xi. 364.