Home

CONFOCAL MICROSCOPY APPLIED TO THE STUDY OF SINGLE

1. 11 SCANNING Process ee Be Bl BIA ESS 12 VERHINDERTE ay oe is ae 13 Scanning correction oe etre a Be ee ee ee ee we 15 Light scattering set up sss baad eee Fr Be ae A 17 Time correlated single photon counting principle 20 Collimation and alignment of the illumination beam 24 Alignment of the dichroic mirror and microscope objective 26 Huyseman ocular s 2 ke eee re ee GALS Se EE 27 Alignment of the spectrograph a 2 0 ua ea woe tls ana 28 Control panel of the PC user interface 040 29 Sample holder ars Sr ap ter 22 Sch a a Sn Br ws RE Gye Aa Fe 30 Calibration sereen a re N ee Er RN 31 Scanning screen a a oe ah ee te cn eg MS ee en gee ee 32 Line scan screen S wech ck tS sent A ea ee Ea Se A 3 33 Fluorescence lifetime imaging 2 02000 35 Kinetic trace screen a wld rear ale 37 Coordinate system 2ER WERK a oP A Ree a 41 Parameters for the TMA coordinate system 43 Calculation of the electric field near a geometric focus 49 Symmetry considerations in the focal plane 2 2 22 2 2 52 Electric fields along an arbitrary direction 53 Experimental configuration 23 20 83 2 00 Dar en ae nd 58 Reflectivity of the sample system 0 2000 62 Full beam illumination images 25 4 eee yo Darren rt 63 Dye molecules on glass and under annular illumination 64 Different modes of illumination
2. 101 102 103 Evolution from molecular to bulk semiconducting properties J Chem Phys vol 82 no 1 pp 552 559 1985 N Chestnoy T D Harris R Hull and L E Brus Luminescence and photo physics of CdS semiconductor clusters The nature of the emitting electronic state J Phys Chem vol 90 no 15 pp 3393 3399 1986 C B Murray D J Noms and M G Bawendi Synthesis and characterization of nearly monodisperse cde e s se te semiconductor nanocrystallites J Am Chem Soc vol 115 pp 8706 8715 1993 Z A Peng and X Peng Formation of high quality cdte cdse and cds nanocrystals using cdo as precursor J Am Chem Soc vol 123 pp 183 184 2001 X Peng Green chemical approaches toward high quality semiconductor nanocrystals Chem Eur J vol 8 pp 334 339 2002 L Qu Z A Peng and X Peng Alternative routes toward high quality cdse nanocrystals Nano Lett vol 1 pp 333 337 2001 L Qu and X Peng Control of photoluminescence properties of cdse nanocrys tals in growth J AM CHEM SOC vol 124 p 2049 2002 M G Bawendi et al Electronic structure and photoexcited carrier dynamics in nanometer size CdSe clusters Phys Rev Lett vol 65 no 13 pp 1623 1626 1990 E Hilinski and P A Lucas A picosecond bleaching study of quantum con fined cadmium sulfide microcrystallites in a polymer film J Ch
3. Ion net fluorescence intensity figure 6 7 b tmin Minimum allowed on or off time equation 6 4 Tpr photo induced on state lifetime From equations 6 4 and 6 5 the times that the simulated QD spends on the off and on state are calculated as tof f tmin Roe 6 6 dm Tan Min tmin R ey TPI In Ram where Min a b takes the minimum of a and b and Ram 0 1 is a random number The total simulated time T is filled with background photons photon detection times The background photons are separated by times tav off calculated according to equation 5 16 as 1 thv off er In Ram 6 7 Within the on periods extra photons are added with detection times separated by thv on calculated as 1 thv on T7 In Ran 6 8 The Monte Carlo procedure generates photon detection times in the same fash ion as the TCSPC module The simulated data was then analyzed via the trace histogram method see section 5 2 2 in the same way as the experimental data Because T m Ion and Ioff are obtained from the experiments the only param eters left free to reproduce the experimental data are tmin and Tpz The influence of each simulation input parameter on the simulated kinetic traces is explained below T total simulated time The finite length of the experimental kinetic traces sets a higher limit for the detectable on or off period and makes the detection of periods of length compara ble to the length of th
4. 0 0496 Table 2 3 Spectrograph components The angular dispersion of the gratings were measured in the laboratory The other information is presented as provided by the manufacturer and the diaphragm aperture of the photo objective the diameter of the detected beam 8 9mm and the size of the CCD sensor The spectral resolution is directly determined as the ratio of the spectral range and the number of pixels of the CCD sensor For the mentioned objective and CCD camera the maximum detectable spectral range is 260nm with a resolution of 0 19nm with the HVG 650 grating and 340 nm with a resolution of 0 25nm with the HVG 590 grating The position of the detectable spectral range can be shifted simply by moving the CCD camera 2 2 4 Time Correlated Single Photon Counting Time Correlated Single Photon Counting TCSPC 17 has been one of the best ways of measuring fluorescence decay times since the method was conceived in 1961 by Bollinger and Thomas 18 TCSPC is based on the detection of single photons and the computation of their individual detection times Figure 2 9 helps to explain the TCSPC working principle The excitation light should be pulsed at such frequency that the fluorescence is allowed to decay completely in between pulses In addition the detected fluorescence intensity should be low enough so that the probability of detecting two photons in between pulses is negligible The latter condition is naturally fulfille
5. 004 65 Influence of the separation distance to the gold film 66 190 LIST OF FIGURES Aly Qu nchine behavior 4 2 zul 2 ei ce Sa wu 67 4 8 Schematic of the excitation and de excitation rates 69 4 9 Electromagnetic decay rates ooo a a 72 4 10 Detectable fraction of the fluorescence vs spacer thickness 73 4 11 Detectable fraction of the fluorescence vs dipole orientation 74 4 12 Electric field distribution in the samples 76 4 13 Electric field distribution in a sample without the gold film TT 4 14 Modelled FB illumination fluorescence signals 78 4 15 Modelled FL illumination fluorescence signals 79 4 16 Modelled TL illumination fluorescence signals 80 4 17 Modelled and experimental fluorescence signals 82 4 18 Profiles of the experimental and modelled fluorescence signals 83 5 1 Three level description of molecular fluorescence 86 5 2 TCSPC photon detection times os 2 Gas sae a er ek 88 3 3 Autocorrelation 4 4 Su sau lapels ern hs Boo d at a Adele GP os 89 5 4 Effect of the bin width on the kinetic trace histogram 91 5 5 nter photon times histogram a 2 Kr re race 92 5 6 Kinetic trace histogram sob awk ae ae ae 94 5 7 Histograms of the length of the on and off periods 95 5 8 Monte Carlo simulated photon detection times 96 5 9 Analysis of the simulated
6. 4 3 1 Full beam images The experiments presented here show that single molecule fluorescence can in fact be excited and detected through the gold film Figure 4 3 shows a typical full beam illumination image of single fluorescent dye molecules on a sample with a 4 bilayer spacer 24 nm thickness 4 3 Single molecule fluorescence images through a thin gold film 63 The great majority of the detected fluorescence signals have the same character istic spatial distribution which is clearly different from the typical diffraction limited spot that would be observed in the absence of the gold film figure 4 4 a Similar patterns were already observed in single molecule fluorescence images obtained with annular 39 40 figure 4 4 b and radially polarized illumination 59 They are due to the fact that the chromophores have a fixed transition dipole and therefore their fluorescence images reflect the spatial distribution of the electric field intensity along the dipole direction see section 3 3 If the chromophores are randomly oriented and the electric field has similar intensity in all directions but different spatial dis tribution a variety of patterns is observed see figure 4 4 b that provide information about the three dimensional orientation of the transition dipole of the molecules Figure 4 3 Full beam illumination images Typical fluorescence micrograph of single dye molecules separated by a 24nm spacer from the 44nm gold
7. Chemical Physics vol 247 pp 1 9 1999 G Binning H Rohrer C Gerber and E Weibel Surface studies by scanning tunnelling microscopy Phys Rev Lett vol 49 no 1 pp 57 60 1982 G Binning C F Quate and C Gerber Atomic force microscope Phys Rev Lett vol 56 no 9 pp 930 933 1986 D M Kolb An atomistic view of electrochemistry Surf Sci vol 500 pp 772 740 2002 G Decher Fuzzy nanoassemblies Toward layered polymeric multicompos ites Science vol 277 no 5330 pp 1232 1237 1997 A Tronin Y Lvov and C Nicolini Ellipsometry and x ray reflectometry characterization of self assembly process of polystyrenesulfonate and polyally lamine Colloid Polym Sct vol 272 pp 1317 1321 1994 K Vasilev Fluorophores near metal interfaces PhD thesis Martin Luther Universitat Halle Wittenberg Halle Germany 2004 W Knoll Interfaces and thin films as seen by bound electromagnetic waves Ann Rev Phys Chem vol 49 pp 569 638 1998 L Novotny M R Beversluis K S Youngworth and T G Brown Longitu dinal field modes probed by single molecules Phys Rev Lett vol 86 no 23 pp 5251 5254 2001 I Pockrand A Brillante and D M bius Nonradiative decay of excited molecules near a metal surface Chem Phys Lett vol 69 no 3 pp 499 504 1980 H Knobloch H Brunner A Leitner F Aussenegg and W Knoll P
8. Corr_Xmondisplay 2 Pixels 1 XfB 2 Pixels 1 Elseif Corr_Xmondisplay XiF gt Corr_Xmondisplay 2 Pixels 1 j 2 Pixels 1 Do difference Corr_Xmondisplay XiF Corr Xmondisplay j fel While difference gt Pixelsize XfB j 1 Else XfB 2 Pixels 1 j XiF Do difference Corr_Xmondisplay 2 Pixels 1 Corr_Xmondisplay j j 1 While difference gt Pixelsize XiF j 1 Endif If Corr_Xmondisplay Pixels 1 Corr_Xmondisplay XiB XfF Pixels 1 Elseif Corr_Xmondisplay Pixels 1 gt Corr_Xmondisplay XiB j Pixels 1 Do difference Corr_Xmondisplay j Corr_Xmondisplay XiB fel While difference gt Pixelsize XfF j 1 Else should never happen aber j XiB Do 168 Set up control and data acquisition software difference Corr Xmondisplay j Corr_Xmondisplay Pixels 1 jt l While difference gt Pixelsize XiB j 1 XfF Pixels 1 Endif PixelsToDiscard XiF Ydwn Corr_Xmondisplay XiF 1 j XiF Do j l While Xdisplay j gt Ydwn PixelsToCorrectInX j 1 XminusCorr_X_F Sum Xdisplay XiF XfF XfF XiF Sum Corr_Xmondisplay XiF XfF XfF XiF XminusCorr_X_B Sum Xdisplay XiB XfB XfB XiB Sum Corr_Xmondisplay XiB XfB XfB XiB End Function CorrectParameters Variable result Pixels Pixels_display PixelsToDiscard Pixelsize Variable Scanrange Xi Xi_um PixelsToCorrectInXScanrange_um Xi Xi PixelsToCorrectInX Pixelsize Scanrange Scanrange PixelsToDiscard PixelSize Pixels round Scanrange Pixelsize Scanrange_um Scanr
9. LineCtsDisplay transferdatawaveL XfB p q 2 1 Endif ControlInfo W ControlPanel FitGauss If V_value 1 ValDisplay Width disable 0 win LineScan Fit Else ValDisplay Width disable 1 win LineScan Endif AvCounts sum LineCtsDisplay 0 Pixels 1 Pixels AvCPS AvCounts 1000000 Pixeltime DoUpdate End With the function Fit it is possible to fit on line a Gaussian curve to the scanned line and display the FWHM Function Fit Wave LineCtsDisplay W_coef CurveFit N Q W 0 M 0 gauss LineCtsDisplay 0 GoodPixels D A 0 F 0 990000 4 width W_coef 3 End LineScanH scans repetitively a line horizontally on the sample until the user A 2 Igor routines 177 aborts The data from every scanned line is displayed on the on line display Function LineScanH StartLineH ButtonControl String StartLineH SetDataFolder root String graphname message Variable aux finish paramctrl Wave fit_LineCtsDisplay graphname winname 0 1 If cmpstr graphname SurfaceScanF 0 If cmpstr graphname SurfaceScanB 0 message Parameters are taken from the top graph At the moment the top graph is graph name Be sure that one of the surface scan images is the top graph Abort message Endif Endif Xi_um xcsr A Yi_um vesr A Scanrange_um xcsr B xcsr A Pixels Pixels_display If Check Values 0 Abort Scan aborted Parameters out of range Endif CalculateParametersForAdwin Do
10. T31 Furthermore a positive correlation between Ij and kon was found which indicates that the gold film influences I and F21 in a similar way i e the triplet singlet and the singlet singlet transition dipoles have similar orientations The photoluminescence blinking of Zno 42Cdo 535e QDs on glass and ITO sub strates was investigated experimentally as a function of the excitation power P At low P it was observed that the probability of a certain on or off time follows a negative power law with exponent m near to 1 6 As P increased the on time fraction reduced on both substrates whereas the off times did not change A weak residual memory effect between consecutive on times and consecutive off times was observed but not between a given on time and the adjacent off time All of this suggests the presence of two distinct mechanisms governing the lifetimes of the on and off states The photoluminescence of quantum dots presented a variety of on intensity dis tributions in general broader than a Poissonian The probability of the on and 147 off times followed the same power law regardless of the shape of the on intensity distribution The blinking of the QDs was modelled via Monte Carlo simulations The QDs were thought to switch between a dark state and an emitting state with power law probability The simulated kinetic traces showed Poisson distributed off and on intensities demonstrating that the non Poissonian on intensity
11. indices 1 k column number indices 2 0 layer number value 0 dataX i MDSetNumericWavePoint Value transferdatawaveF indices value i i 0 while i lt Pixels indices 0 i row number indices 1 k column number indices 2 1 value 0 dataYfli MDSetNumericWavePoint Value transferdatawaveF indices value i i 0 while i lt Pixels indices 0 i row number indices 1 k column number indices 2 2 value 0 dataCtsf i MDSet Numeric WavePoint Value transferdatawaveF indices value i i 0 while i lt Pixels indices 0 i row number indices 1 k column number indices 2 3 value 0 datat1f i MDSetNumericWavePoint Value transferdatawaveF indices value i i 0 while i lt Pixels indices 0 i row number indices 1 k column number indices 2 4 value 0 datat2f i MDSetNumericWavePoint Value transferdatawaveF indices value i GetlData 11 dataXb 1 Pixels transfers points 1 to Pixels of ADW data_l to dataXb GetlData 12 dataYb 1 Pixels GetlData 13 dataCtsb 1 Pixels GetlData 14 datatlb 1 Pixels GetlData 15 datat2b 1 Pixels i 0 while i lt Pixels indices 0 i row number indices 1 k column number indices 2 0 layer number value 0 dataXb i MDSetNumericWavePoint Value transferdatawaveB indices value i i 0 while i lt Pixels indices 0 i row number indices 1 k column nu
12. ule 19 2 3 Alignment The complete alignment of the set up can be separated in four parts The first one is the in coupling of light provided by any of the light sources to the single mode fiber The second is the collimation of the light provided by the fiber and the alignment of the illumination beam The third one is the alignment of the dichroic mirror or beam splitter and the microscope objective Finally the fourth part is the alignment of the detection optics and detectors In the present section the four alignment procedures are explained 2 3 1 Light coupling into the single mode fiber A fundamental pre requisite for an effective coupling of light into the single mode fiber is that both tips of the fiber have to be be sharply cut For this it is first necessary to remove the polymer cladding and then cut the tips with a diamond fiber cutter RXS Kabelgarnituren GmbH The process of in coupling light into the single mode optical fiber consists of two steps First light provided by any of the sources has to be focused onto the fiber tip Second the fiber position needs to be adjusted in order to maximize the in coupled light The different light sources have different focusing requirements as explained below To couple the focused light into the fiber one should first position the fiber tip approximately in the focus by moving the tip in x y and z directions until light scattered by the fiber tip is observed Then iterat
13. 2 2 Description of the home built confocal microscope 17 Fluorescence measurements For fluorescence measurements in order to remove residual excitation light not filtered by the dichroic mirror suitable long pass Omega Optics Inc and notch Notch Plus Kaiser Optical Systems Inc filters are placed in the detection channel Light scattering measurements For light scattering measurements no filters are used Instead annular illu mination is employed and the outer part of the detected beam which under ring illumination conditions contains mainly reflected light is blocked by a diaphragm so that only scattered light can reach the detectors see figure 2 8 Piezoelectric XYZ stage Microscope objective Anular illumination disc Beam Splitter Diaphragm To the single photon detectors gt or spectograph Mirror Figure 2 8 Light scattering set up configuration The dashed line shows the path followed by the reflected light It is also possible to work with the inverse configuration i e illuminate with a reduced beam and block the inner part of the detection channel with an appropriate disc The single photon detectors The set up counts with two different single photon detectors an Avalanche Photo Diode APD and a Photo Multiplier Tube PMT Both detectors can be used for imaging and time correlated measurements Table 2 2 shows the principal technical characteristics of each detector The APD
14. 44 45 46 L Novotny Light Propagation and light confinement in near field optics PhD thesis Swiss Federal Institute of Technology Zurich 1996 L Novotny Allowed and forbidden light in near field optics i a single dipolar light source J Opt Soc Am A vol 14 pp 91 104 1997 E Wolf Electromagnetic diffraction in optical systems i an integral repre sentation of the image filed Proc Roy Soc London A vol 253 no 1274 pp 349 357 1959 B Richards and E Wolf Electromagnetic diffraction in optical systems ii structure of the image filed in an aplanatic system Proc Roy Soc London A vol 253 no 1274 pp 358 379 1959 E Engel N Huse T A Klar and S W Hell Creating 3 focal holes with a mach zehnder interferometer App Phys B vol 77 pp 11 17 2003 B Sick B Hecht and L Novotny Orientational imaging of single molecules by annular illumination Phys Rev Lett vol 85 no 21 pp 4482 4485 2000 M Kreiter M Prummer B Hecht and U Wild Orientation dependence of fluorescence lifetimes near an interface J Chem Phys vol 117 p 9430 2002 K H Drexhage M Fleck H Kuhn F P and W Sperling Beeinflussung der fluoreszenz eines europiumchelates durch einen spiegel Ber Bunsenges Phys Chem vol 70 no 9 10 p 1179 1966 K Drexhage H Kuhn and F Sch fer Variation of fluorescence decay time of a
15. H Brivanlou and A Libchaber In vivo imaging of quantum dots encapsulated in phospholipid micelles Science vol 298 pp 1759 1762 2002 K Kalyanasundaram E Borgarello D Duonghong and M Gr tzel Cleav age of water by visible light irradiation of colloidal cds solutions inhibition of photocorrosion by ruo2 Angewandte Chemie International Edition in En glish vol 20 no 11 pp 987 988 1981 D Duonghong J Ramsden and M Gratzel Dynamics of interfacial electron transfer processes in colloidal semiconductor systems J Am Chem Soc vol 104 no 11 pp 2977 2985 1982 L E Brus A simple model for the ionization affinity and aqueous redox potentials of small semiconductor crystallites J Chem Phys vol 79 no 11 pp 5566 5571 1983 L E Brus Electron electron and electron hole interactions in small semicon ductor crystallites The size dependence of the lowest excited state J Chem Phys vol 80 no 9 pp 4403 4409 1984 R Rossetti J L Ellison J M Gibson and L E Brus Size effects in the excited electronic states of small colloidal CdS crystallites J Chem Phys vol 80 no 9 pp 4464 4469 1984 R Rossetti R Hull J M Gibson and L E Brus Excited electronic states and optical spectra of ZnS and CdS crystallites in the amp 15 to 50 a size range BIBLIOGRAPHY 201 91 92 94 95 96 97 98 99 100
16. P kWicm loff cps lon eps Figure 6 7 a Experimental and simulated off intensity b Experimental and simulated net on intensity and experimental overall on intensity All as a function of the excitation intensity P for Zno 42Cdo 5g5e QDs on glass and ITO coated glass substrates 2 07 1 8 E 1 64 Figure 6 8 Exponent of the power law fit to 14 the experimental and simulated glass only 0 Glass off period length histograms of the QDs on 12 N glass and ITO coated glass substrates as a Simulated glass A BE A function of the excitation intensity P 1 04 T T T T 0 0 0 5 1 0 1 5 2 0 P kWiem I performance of the QDs The increase of Ion and Ioff is not perfectly linear with the excitation power This is due principally to the fact that the measured values of intensity are very sensitive to the alignment of the detector Since the active area of the APD has a diameter of 180 um a misalignment of some tens of micrometers can produce a considerable drop in the measured intensity As the off intensity is composed mainly of background intensity its dependence on the detector alignment is less pronounced Power law exponent All the histograms of the off periods length can be satisfactorily fitted with a power law y Ax with an exponent m between 1 and 2 and close to 1 6 As shown in figure 6 8 the power law exponent does not show any evident de pendence on the excit
17. PECON J 0 001 A BAA Ai a zu m m 11a a zu m Te rm ern 0 1 1 10 100 0 1 1 10 100 0 1 1 10 Period length s Period length s Period length s Figure 6 16 Histograms of the length of the on and off periods of the simulated kinetic traces The simulations were performed to reproduce the experimental data of the QDs on ITO for the different excitation intensities P The results are presented in the same fashion as figure 6 4 All the characteristic parameters of the QDs blinking can be reproduced with the model see figures 6 7 to 6 10 except for the non Poissonian on intensity dis tribution figure 6 14 and the residual memory effect of the consecutive on times and consecutive off times Figure 6 17 shows exemplary plots of the on times vs the successive on times the off times vs the successive off times and the on times vs the successive off times obtained from the simulated kinetic traces As expected from the random generated on and off time periods no correlation is observed in this case 130 Photoluminescence blinking of Zno 42Cdo53 Se nano crystals ton S vs next ton S tor S vs next tor S ton S vs next tor S z R 0 02 R 0 01 ji 10 41 0 001 0 1 10 0 001 0 1 10 Figure 6 17 Correlations between adjacent on and off times on logarithmic scales obtained from the trace histogram analysis of simulated traces The photo induced on time lifetime By means of the photo induced lifetim
18. Parr Pe T23 Kopp TV 23 As the on off and on off transitions are single rate processes the length of the on and off periods ton and tory are exponentially distributed Risen ee 5 6 P togp t Ze Vt Toff with average times for each state given by ne L Pate Peine P23 koff I jol 23 5 7 1 1 To z an Kon k31 Then the time averaged intensity of a single molecule is given by Ton Ten Lo To Ionkoff loffken I Noel 5 8 Ton Tof f kon koff From equations 5 8 5 4 and 5 5 it can be seen that for the case of koff gt kon the fluorescence emission is limited to a value independent of the excitation rate given by Parl 31 r 23 Tig 5 9 Liriplet This effect is called triplet saturation or triplet bottleneck 88 Single molecule fluorescence dynamics 5 2 Kinetic traces analysis methods Kinetic traces are recorded with the TCSPC module as explained in section 2 4 3 In order to extract the desired information about triplet blinking dynamics from the recorded photon detection times it is necessary to analyze carefully the raw data obtained with the TCSPC module see section 2 2 4 Figure 5 2 shows as an example the TCSPC mac t times corresponding to 239980 photons detected during 18 seconds from a single DilC1 5 molecule The intensity fluctuations can be observed as changes in the slope of the curve as shown in the small inset of figure 5 2
19. Prioritat 1 Version 1 FastStop 0 AdbasicVersion 2000000 150 Set up control and data acquisition software ATSRAM 0 OPT_LEVEL 1 SAVECOMPIL 0 DIM i check X Y AS INTEGER DIM DATA_1 10000 AS LONG DIM DATA_2 10000 AS LONG DIM DATA_3 10000 AS LONG DIM DATA_4 10000 AS LONG DIM DATA_5 10000 AS LONG DIM DATA_11 10000 AS LONG DIM DATA 12 10000 AS LONG DIM DATA_13 10000 AS LONG DIM DATA_14 10000 AS LONG DIM DATA_15 10000 AS LONG DIM DATA_31 10000 AS LONG DIM DATA 32 10000 AS LONG DIM DATA_21 100000 AS LONG DIM DATA_22 100000 AS LONG DIM DATA_23 100000 AS LONG DEFINE Xi PAR_1 DEFINE Yi PAR_2 DEFINE X PAR_7 DEFINE Y PAR_8 EVENT GLOBALDELAY 1 i 0 X PAR_7 Y PAR_8 IF X lt 32768 THEN X 32768 ENDIF IF Y lt 32768 THEN Y 32768 ENDIF IF X gt 65535 THEN X 65535 ENDIF IF Y gt 65535 THEN Y 65535 ENDIF IF X lt gt Xi THEN IF Xi lt X THEN IF X Xi lt 1900 THEN DO i 0 DEC X A 1 AD Basic routines 151 DO DAC 1 X i i 1 UNTIL i 2000 UNTIL X Xi ELSE DO i 0 DEC X DO DAC 1 X i i 1 UNTIL i 150 UNTIL X Xi ENDIF ELSE IF Xi X lt 1900 THEN DO i 0 INC X DO DAC 1 X i i 1 UNTIL i 2000 UNTIL X Xi ELSE DO i 0 INC X DO DAC 1 X i i 1 UNTIL i 150 UNTIL X Xi ENDIF ENDIF ENDIF i 0 IF Y lt gt Yi THEN IF Yi lt Y THEN IF Y Yi lt 1900 THEN DO i 0 DEC Y DO DAC 2 Y 152 Set up control and data acquisition software i i 1 UNTIL i 2000 UNTIL
20. Variable PxIXi PxIYi cambio n TO String graphname graphname winname 0 1 If cmpstr graphname 0 strlen graphname 2 SurfaceScan 0 Abort The position input is given by the position of round cursor in the top SurfaceScan image At the moment no scan image is the top graph The process will be aborted Endif cambio 0 If Xi_um hcsr A Xi_um hcsr A cambio 1 Endif If Yi_um vesr A Yi_um vesr A cambio 1 Endif Xi Xi_um 32768 80 32768 Yi Yi um 32768 80 32768 DoUpdate PixeltimeForKinetic PixeltimeForK inetic_display 1000 If cambio 1 180 Set up control and data acquisition software Init Endif Kineticduration Kineticduration_display Points Kineticduration 1000000 PixeltimeFor Kinetic If Points gt 50000 Abort Too many data points per cycle The ADWin card Igor comunication process might fail Please set a shorter repetition time or a longer pixel time Endif Make O U N Points 3 transferkineticwave transferkineticwave Nan Make O U N 0 KineticCtsDisplay Make O U N 0 KineticTimeDisplay DoWindow F SurfaceScanF DoWindow F SurfaceScanB DoWindow F PointKinetic SetAxis W PointKinetic A If V_flag 1 Execute PointKinetic Endif n 0 OpenShutter Do ExecuteKinetic InsertPoints n Points Points KineticCtsDisplay InsertPoints n Points Points KineticTimeDisplay KineticCtsDisplay n Points n 1 Points 1 transferkineticwave p n Points 2 1 If n 0 T0 transferkineticwave 0
21. if the illumination beam has rotational symmetry around the z axis and for points r in a plane parallel to the focal plane it is only necessary to calculate one in plane xy component of the electric field along two directions and the z component along one direction In particular E v 2 0 zp Ex p y 0 zp EP 3 29 T EAL 0 y zP Elp T 73 2P ie 3 30 E r 2 0 zp E p p 0 zp EY 3 31 From these components the total field along any radial direction defined by p a in the plane of interest can be calculated To understand this point it is necessary to make some symmetry considerations first for the in plane components Ez and E and then for the E component Due to the mirror symmetry of the problem with respect to the x axis any y component of the fields along the x axis vanishes One way to see this is to imagine that the waves focused through the positive y semi plane of the objective rear lens generate a non zero y component along the x axis Then the plane waves focused through the negative y semi plane would generate a y component along the x axis of equal magnitude and opposite sign Figure 3 4 a shows this effect schematically Due to the inversion symmetry with respect to the y axis imposed by the direc tion of the source field any y component of the fields along the y axis vanishes One way to see this is to imagine the source field generates a non zero y component in a 52 Si
22. t2display Redimension N 2 Pixels Corr_Xmondisplay Xdisplay 0 Xmondisplay 0 Corr_Xmondisplay 0 t2display 0 Make O I U N 2 Pixels 4 calibrationwave Init ActualY Yi Pixelsize 2 XminusXmon 0 j 0 Do ExecuteScanCalibration CalibrationDisplay j 1 While j lt 3 Return 1 End Function FitFinalLinearRange Wave Xdisplay Xmondisplay t2display Wave Corr_Xmondisplay W_coef PrevXdisplay 170 Variable a b Yup Ydwn i j difference Pixels_aux RemoveFromGraph Z fit_Corr XmondisplayF RemoveFromGraph Z fit_Corr_XmondisplayB ControlInfo W ControlPanel KeepPrevFit If V_value 1 Append ToGraph C 10000 65535 10000 PrevXdisplay vs t2display Endif XiF 0 4 Pixels XfF Pixels 1 Cursor P A Corr_Xmondisplay XiF Cursor P B Corr_Xmondisplay XfF CurveFit Q N line Xdisplay 0 Pixels 1 a bx CurveFit Q N H 01 line Corr_Xmondisplay XiF XfF a bx a W_coef 0 b W_coef 1 Make O N Pixels 1 20 fit_Corr_XmondisplayF fit_Corr XmondisplayF a b p AppendToGraph fit_Corr XmondisplayF vs t2display XiB 1 4 Pixels XfB 2 Pixels 1 Cursor P A Corr_Xmondisplay XiB Cursor P B Corr_Xmondisplay XfB CurveFit Q N line Xdisplay Pixels 2 Pixels 1 a bx CurveFit Q N H 01 line Corr_Xmondisplay XiB XfB a W _coef 0 b W_coef 1 Make O N Pixels fit_Corr_XmondisplayB fit_Corr XmondisplayB a b p Pixels Duplicate O R Pixels 2 Pixels 1 t2display t2displayB AppendToGraph fit_Corr_XmondisplayB vs t2displayB j 0 Do dif
23. 0 225 45 675 90 0 225 45 675 9 f Figure 4 14 Modelled FB illumination fluorescence signals Calculated fluorescence signal of an ideal QE 1 fluorophore under FB illumination on the air left and on the spacer right sides of the air spacer interface The orientation of the transition dipole of the fluorophore is defined by and according to figure 4 1 At the bottom the maxima of the signals for 8 0 as a function of 4 4 Modelling the experimental scheme 79 The so modelled signals calculated with the method explained in chapter 3 are proportional to the experimental ones Furthermore as the actual value of the non radiative decay rate of the chromophore is unknown an ideal fluorophore with I 0 was considered in the determination of the detectable fraction of the fluorescence emission This is at first a good approximation because the fluorophores used in the experiments are supposed to have intrinsic quantum yields near unity The changes introduced by a Tt 4 0 are discussed later in this section Figure 4 14 shows the modelled FB illumination fluorescence signals The results on either side of the interface are noticeable different On the air side the molecules with a greater out of plane z component of their transition dipole present the stronger signals and the majority of the signals present the spatial distribution char acteristic of the intensity of the z component of the electric field On the co
24. 01 o 20 40 60 80 100 0 1000 0 0101 1 Time s Counts Time s Figure 6 3 QDs kinetic traces Left Photoluminescence emission vs time traces of different QDs Center histogram of the photons per bin of each trace intensity solid grey and Poisson distributions lines with average equal to the experimental intensities Right histograms of the length of the on and off periods 6 3 QD kinetic traces 115 In comparison to molecular fluorescence kinetic traces there are two evident differences First the QDs are able to remain on and off for much longer time periods than the fluorescent dyes Second the QD traces are more chaotic in the sense that a variety of blinking rates is found even on a single QD Some observations can be made from a direct inspection of the intensity distri butions figure 6 3 center As in the case of molecular fluorescence the blinking of QDs present a Poisson distributed oflf intensity with an average value equal to the background intensity This indicates that the QDs indeed stop emitting during the dark periods In contrast the on intensity is not Poisson distributed The traces shown in figure 6 3 present on populations with different degrees of deviation from a Poisson distribution At this point it is important to mention that even though the algorithm of the trace histogram analysis was designed to find an optimum threshold between two Poisson distributed states section 5 2 2 it also works in this
25. 1 Endif KineticTimeDisplay n Points n 1 Points 1 transferkineticwave p n Points 1 T0 25 1e9 DoUpdate n 1 While 1 End A 3 C routines In this section all C programmed XOP routines are presented The ADWin Boot2 XOP boots the local CPU of the AD DA PC card and loads in it all the ADBasic routines for the operation of the microscope A 3 C routines 181 ADWinBoot2 c include XOPStandardHeaders h Include ANSI headers Mac headers IgorXOP h XOP h and XOPSupport h include ADWinBoot2 h include C ADwin Developer C Microsoft_VisualC Adwin c include C ADwin Developer C Microsoft_VisualC Adwin h All structures are 2 byte aligned if GENERATINGPOWERPC pragma options align mac68k endif ifdef WINDOWS pragma pack 2 endif static void XOPEntry void long Xi Yi firstrun switch GetXOPMessage case CMD CMD is the only message we care about Boot D Doktorarbeit Control ADBasic adwin9 btl 800000 Boot ADwin ADBPrLoad D Doktorarbeit Control CompleteRoutines Scan Acquisition Control ADBasic LP HP LineScan T91 ADBPrLoad D Doktorarbeit Control CompleteRoutines Scan Acquisition Control ADBasic HP Pixel Forth T92 ADBPrLoad D Doktorarbeit Control CompleteRoutines Scan Acquisition Control ADBasic HP Pixel Back T93 ADBPrLoad D Doktorarbeit Control CompleteRoutine
26. 2 in figure 2 5 to a definite extent set by the scanning range and with a definite number of steps set by the number of pixels The sample remains a certain time set by pixel time at each position while it is illuminated and fluorescence or scattered light photons are counted to give the intensity of one pixel of the image Next the sample is moved one step in the y direction in order to start the next x line arrow 3 in figure 2 5 The process is repeated until the scanning of a squared area of the sample is completed The scanning is accomplished by a computer controlled xyz piezoelectric stage TRITOR 101 CAP Jena Piezosystems GmbH operated with a capacitive closed loop feedback to correct the drift of the piezoelectric drivers The piezoelectric stage can be driven via a driving signal with a voltage from 0 to 10 V through a range from 0 to 80 um in each direction with a resolution of 0 5nm The driving signal for the stage is generated with a computer controlled 15 bit Analog to Digital Digital to Analog AD DA converter ADWin light 16 Jager GmbH and then amplified with 3 one for each of the x y and z channels independent low noise amplifiers ENV40CAP Jena Piezosystems GmbH The minimum scanning step is limited by the 15 bit resolution of the AD DA converter to 2 4nm The AD DA converter is 3For details about the different scanning methods see the book by T Wilson 16 2 2 Description of the home built confocal micr
27. AdbasicVersion 2000000 ATSRAM 0 OPT_LEVEL 1 SAVECOMPIL 0 DIM delay timel time2 time Pixeltime AS long DIM Y j Counts AS INTEGER DIM DATA_3 10000 AS LONG Counts DIM DATA _4 10000 AS LONG time0 DIM DATA_5 10000 AS LONG time3 DEFINE Pixeltime PAR_5 DEFINE flag PAR_55 INCLUDE C ADwin ADbasic3 Inc adwgcent inc INIT Globaldelay 1 EVENT Pixeltime PAR_5 Y PAR_7 j PAR_9 flag 0 timel READ_TIMER IF j gt 1 THEN timel DATA_5 j 1 ENDIF DO DAC 2 Y time READ_TIMER delay time timel UNTIL delay gt Pixeltime time2 READ_TIMER Counts CNT_READ 1 CNT_CLEAR 1 DATA_3 j Counts A 1 AD Basic routines 159 DATA_4 j timel DATA_5 j time2 END FINISH flag 1 LP HP LineScan Vert BAS Proze nummer 3 Delay 1 Eventsource 0 Number of Loops 0 Prioritat 1 Version 1 FastStop 0 AdbasicVersion 2000000 ATSRAM 0 OPT_LEVEL 1 SAVECOMPIL 0 DIM test Tpixels j AS INTEGER DIM DATA_1 10000 AS LONG X DIM DATA_2 10000 AS LONG Y DIM DATA_6 10000 AS LONG Monitor X DEFINE Xi PAR_1 DEFINE Yi PAR_2 DEFINE Scanrange PAR_3 DEFINE Pixels PAR_4 DEFINE Pixeltime PAR_5 DEFINE Pixelsize PAR_6 DEFINE Y PAR_7 DEFINE X PAR_8 DEFINE j PAR_9 DEFINE terminated PAR_11 DEFINE flag PAR_55 INCLUDE C ADwin ADbasic3 Inc adwgent inc INIT terminated 0 globaldelay 1 test 0 IF test 1 THEN Xi 32768 Yi 32768 Scanrange 32768 Pixels 5 160 Set up control and data
28. Brus Charge polarizability and photoionization of single semiconductor nanocrystals Phys Rev Lett vol 83 no 23 pp 4840 4843 1999 G Schlegel J Bohnenberger I Potapova and A Mews Fluorescence decay time of single semiconductor nanocrystals Phys Rev Lett vol 88 p 137401 2002 B R Fisher H J Eisler N E Stott and M G Bawendi Emission intensity dependence and single exponential behavior in single colloidal quantum dot fluorescence lifetimes J Phys Chem B vol 108 pp 143 148 2004 S Hohng and T Ha Near complete suppression of quantum dot blinking in ambient conditions J Am Chem Soc vol 126 pp 1324 1325 2004 BIBLIOGRAPHY 203 116 117 118 119 120 121 122 123 124 125 126 127 128 X Zhong M Han Z Dong T J White and W Knoll Composition tunable Zn Cd _ Se nanocrystals with luminescence and stability J Am Chem Soc vol 125 no 8 pp 8589 8594 2003 Y Park V Choong Y Gao B R Hsieh and C W Tang Work function of indium tin oxide transparent conductor measured by photoelectron spec troscopy Appl Phys Lett vol 68 pp 2699 2701 1996 H M Wadsworth The Handbook of Statistical Methods for Engineers and Scientists McGraw Hill 2nd ed 1998 M Kuno D P Fromm A Gallagher D J Nesbitt O I Micic and A J Nozik Fluorescence intermittency
29. For a given bin width bw the number of photons per bin of the on or off population is given by equation 5 15 with t bw Then in the histogram of the kinetic trace there are two populations of photons per bin the on and the off populations which have n photons per bin with a probability given by I bw ete n P n 5 23 The iterative process starts with a very small bin width 25 ns and calculates the number of bins corresponding to each state bins T bw Then the total number of bins that are wrongly classified when a threshold T is used to distinguish the on from the off bins can be calculated by T T Wrong bins bins on Pon n bins off gt 20 5 24 n 0 n 0 94 Single molecule fluorescence dynamics Cts 10 15 el yi IH oe IB O 500 1000 Threshold 04H 1 reso UMMA sm N NAR dh sem hd 4 0 42 44 46 48 50 0 5000 Time s Cts Figure 5 6 Kinetic trace histogram Top optimum bin width 0 61 ms histogram of the data shown in figure 5 2 The horizontal line shows the optimum threshold 9 photons to distinguish the on from the off bins Bottom right the intensity histogram photons per bin solid grey and Poisson distributions The vertical scale is the same as in the histogram In the small graph the intensity histogram and the Poisson distributions are plotted up to 1000 counts For a given bin width the total number of wrongly classified bins is calculated f
30. In this model a positively charged QD is a dark QD and a neutral QD is a bright QD and blinking occurs due to ionization and neutralization events Whenever possible the following results will be discussed in the context of this model Off intensity and on intensity From the histogram analysis of the kinetic traces it is possible to count the photons corresponding to on and to off periods Non and Nor and to calculate the off and on intensities Zon and Toff Figure 6 7 a shows as a function of the excitation intensity P the off intensity 118 Photoluminescence blinking of Zny Cd sSe nano crystals 4000 1 P 2 04 kWicm 1000 4 P 1 61 kWicm 1000 ON ON 100 OFF 100 4 OFF 100 fit m 1 63 fit m 1 60 2 10 227 g 10 c ec Cc a a 1 2 1 81 0 1 4 0 14 0 1 0 01 0 01 4 T 108s 0 01 0 001 4 T 0 001 0 01 0 1 1 10 0 1 1 10 100 0 1 1 10 Period length s Period length s Period length s 1000 P 0 74 kWiem 1000 P 0 18 kW cm P 0 026 kWicm ON ON 100 ON 100 OFF 100 4 OFF OFF fit m 1 64 fit m 1 65 10 fit m 1 78 g 10 g 7 2 E c c 2 14 2 14 a 1 re 0 14 014 0 1 4 0 01 j 0 01 4 0 01 4 0 001 1 os d a 0 1 1 10 100 0 1 1 10 100 0 1 1 10 Period length s Period length s Period length s Figure 6 6 Histograms of the length of the on and off periods obtained from the common computation of several individua
31. P21 Figure 5 14 shows the scatter plots of kon vs Ia for the molecules in the samples with and without gold The linear correlation coefficient shown in the corresponding graphs indicate that kon and 2 are positively correlated in the case with gold while they are not in the case without gold The 106 Single molecule fluorescence dynamics wider range of I spanned by the chromophores in the sample with gold is purely an effect of the gold influence Therefore the correlation between kon and 2 in the complete range of I s is a manifestation of the gold film influence on kon This indicates that kon T31 can be enhanced in the same way as F by the gold film To understand this it is important to recall that the transition 3 1 in volves two different processes First a process that is able to pair back the electron spins should occur possibly spin orbit coupling in order to provide the additional angular momentum to break the selection rule Then the molecule can decay to the singlet ground state via the emission of a photon The gold film is able to accelerate the transition 3 1 by influencing the radiative part of the process In addition the positive correlation of 2 and kon implies that the orientation of the transition dipole associated to the 3 1 radiative decay must be oriented close to the singlet transition dipole The latter is not surprising because in order for ISC to occur the excited singlet and triplet sta
32. This is justified by the fact that for usual laser pulses repetition rates 50 MHz the probability of detecting photons in consecutive pulses is negligible detection rates for a single molecule are 0 05 MHz 2 2 Description of the home built confocal microscope 21 pulses which is around 300 ps for both pulsed lasers Second the reproducibility of the rising cant of the detector pulses The latter is important due to the zero crossing method used by the TCSPC unit to determine the temporal position of the signals further details can be found in the TCSPC operation manual 19 2 2 5 Computer control Several functions of the microscope are computer controlled either via commer cially available software or by home made programs developed during the work for the present dissertation This section describes how the computer control of the dif ferent functions is organized Details about the operation can be found in section 2 4 Scanning and confocal data acquisition control The operation of the piezoelectric stage as well as the data acquisition for all types of imaging is accomplished with the AD DA converter The AD DA is com puter controlled with a home made software that consists of three parts The first part consists of routines that operate at the lowest level directly on the local CPU and memory of the AD DA converter and perform the basic operations such as scanning a line or counting photons at one pixel These routin
33. When the molecule is emitting more photons are detected in a unit time and therefore the slope is lower The opposite occurs when the molecule is not emitting For an ideal onSSoff system only two slopes would be observed Detection time s 0 T T T T T T T T T T T T T T T T T T T T T T T 1 0 50 100 150 200 Photon number Figure 5 2 Photon detection times Arrival times TCSPC mac t times of each detected fluores cence photon in chronological order Fluorescence on and off periods are observed as fluctuations in the slope of the curve Two methods were used to extract the onSoff transition rates from the TCSPC data The first one is based on a widely used method for the analysis of the intensity autocorrelation and is explained in section 5 2 1 The second method described in section 5 2 2 consists of finding the optimum bin width to make a histogram of the data and then distinguishing the on and the off periods with a suitable threshold 5 2 1 Autocorrelation analysis The autocorrelation method was first developed for the study of triplet intensity fluctuations of single molecules in solution 67 and then applied to the study of immobilized single molecules at cryogenic 68 69 and room temperatures 70 71 The normalized intensity autocorrelation is defined as the rate of detection of two photons in a time interval 7 relative to that rate if the photon detection would 5 2 Kinetic traces analysis methods 89
34. acquisition software Pixeltime 200 Y 45000 ENDIF EVENT PAR_11 0 Pixelsize Scanrange Pixels Pixeltime Pixeltime 40 usec Y Yi Pixelsize 2 X PAR_8 flag 1 j 0 CNT_ENABLE 1 CNT_CLEAR 1 DAC 1 X DO j j 1 Y Y Pixelsize DATA_1 jJ X DATA_2 jJ Y START_PROCESS 8 DO UNTIL flag 1 DATA_6 jJ ADC 1 UNTIL j Pixels DO j j 1 START PROCESS 8 DO UNTIL fag 1 DATA_1 jJ X DATA_2 jJ Y Y Y Pixelsize DATA_6 jJ ADC 1 UNTIL j 2 Pixels Y Y Pixelsize END FINISH PAR_11 1 HP KineticPixel BAS Proze nummer 6 Delay 1 Eventsource 0 A 1 AD Basic routines 161 Number of Loops 0 Prioritat 0 Version 1 FastStop 0 AdbasicVersion 2000000 ATSRAM 0 OPT_LEVEL 1 SAVECOMPIL 0 DIM DATA_21 100000 AS LONG DIM DATA_22 100000 AS LONG DIM DATA_23 100000 AS LONG DIM time timel time2 delay counts AS LONG DIM test AS INTEGER DEFINE Pixeltime PAR_5 DEFINE j PAR_9 DEFINE flag PAR_55 DEFINE Xi PAR_1 INCLUDE C ADwin ADbasic3 Inc adwgent inc INIT globaldelay 1 test 0 IF test 1 THEN PAR_5 500000 ENDIF EVENT Pixeltime PAR_5 flag 0 timel READ_TIMER j PAR_9 CNT_ENABLE 1 CNT_CLEAR 1 timel READ_TIMER IF j gt 1 THEN timel DATA_22 j 1 ENDIF DO time READ_TIMER delay time time1 UNTIL delay gt Pixeltime time2 READ_TIMER Counts CNT_READ 1 CNT_CLEAR 1 DATA_23 j Counts DATA_21 j timel DATA_22 j time2 j j 1 162 Set up control and data acquisition softwar
35. and detected on and off periods The filled curve shows the simulated on and off periods time resolution lys the black curve is the kinetic trace histogram bin width 0 46 ms and the horizontal line marks the threshold found via the trace histogram method to distinguish the on from the off bins 6 4 3 Simulated blinking Figure 6 14 shows a simulated kinetic trace left the photons per bin histogram center and the histogram of the length of the on and off periods The simulated kinetic trace and its on and off period distributions look very similar to the experimental ones However the intensity distribution does not In contrast to the experiments the simulated on intensity is Poisson distributed This was corroborated by simulations using input parameters spanning the complete ex perimental range and no deviation from the Poisson behavior was observed The hypothesis that the deviations of the on intensity distribution from a Poisson be havior are due to partly detected short on times produced in great quantity by the power law probability 119 can be refuted 128 Photoluminescence blinking of Zno42Cdo53 Se nano crystals 60 40 60 g 20 2 10 40 0 3 01 fo oO 20 0 2000 0 001 0 x 3 E w a os 0 20 40 60 80 100 120 140 O 4000 8000 0 001 01 Time s Counts Time s Figure 6 14 Simulated QD kinetic trace Left Photoluminescence emission vs time traces Center histogram of the photons per bin intensity solid
36. are shown in the graphs In the case of T21 together with the experimental distributions Exp Monte Carlo simulated Sim distributions considering the chromophores on the air side are plotted The inset in the graph of T21 for the case without gold is a Monte Carlo simulated distribution considering the chromophores on the spacer side of the interface 102 Single molecule fluorescence dynamics em em Table 5 3 Calculated normalized to total total de tal electromagnetic de excitation rate of without Au 1 34 2 21 a parallel and a perpendicular dipole in with Au 1 05 5 75 the samples with and without gold on the air side of the spacer air interface 5 4 1 Influence on Ts The influence of the gold can be clearly seen by comparing in the distributions of T a for the molecules in the samples with and without gold The molecules in the sample without the gold film present a narrow distribution with an average value of 0 42 x 10 1 s which width and asymmetry can be explained by the different orientations of the molecules The total electromagnetic decay rate of a chromophore in the samples with gold was calculated in section 4 4 2 by means of the method described in section 3 2 The same method can be used to calculate the decay rate of molecules in the samples without the gold film In chapter 4 it was concluded that the chromophores electrostatically deposited on the surface of the polyelectrolyte spac
37. be uncorrelated and can be expressed as I t I t r nr where t is the intensity as a function of time and the angle brackets denote time average In practice the TCSPC data is divided in time intervals of a given size bw CA r 5 10 and the autocorrelation is calculated as N t N t 7 OO NOP 5 11 where N t are the photons detected in a given interval of length bw centered at a time t and the angle brackets denote time average As an example the autocorre lation of the data shown in figure 5 2 is shown in figure 5 3 Experimental data 25 4 Exponential fit 2 5 Tr 204 gt 2 0 N D 15 1 97 1 0 0 5 10 1 0 T 2 TU 2 3 ios 2 ae N 10 10 10 10 T ms Figure 5 3 Autocorrelation of the data shown figure 5 2 The horizontal axis has a logarithmic scale to easily identify the inflexion point which abscissa corresponds to the characteristic time of the process observed In the inset the same data is plotted with linear scales The experimental data can be satisfactorily fitted with a single exponential decay For uncorrelated events C r 1 and for correlated events C r gt 1 If C r is computed in the right time window for a certain process a decay of C r to unity is observed as 7 increases From this decay it is possible to ex tract information about the characteristic time of the process dynamics For the fluctuations of the fluorescence emission of a single mol
38. brought a new complication the QDs present extremely complicated emission fluctuations that could not be explained until now Even though surface enhancement effects were observed on QDs 6 the lack of knowledge about the blinking mechanism prevents an effective exploitation of the effect It is fundamental to understand the blinking of QDs before trying to improve their performance by other means such as locally enhanced fields The photoluminescence blinking of QDs was experimentally investigated and modelled in order to gain some insight into the underlying physical processes chapter 6 The last experimental tool necessary for the investigation of field enhancements on a single nano structure level is the capability of studying SPRs in individual metallic nano particles In order to fill this need the home built SCOM was adapted for light scattering measurements and its performance was tested on spherical and C shaped gold particles chapter 7 Chapter 2 The fluorescence and light scattering confocal microscope A scanning confocal optical microscope SCOM was designed and constructed to perform local studies of fluorescence and light scattering with diffraction limited spatial resolution and single photon sensitivity This instrument allows working with many different experimental schemes The most distinctive characteristic is the possibility of measuring fluorescence and scattered light from the same focal spot The sections o
39. comparatively even stronger signals than in the case of FB illumination On the spacer side the signals of the in plane along x and out of plane molecules become comparable but much weaker air spacer Belt 0 1 2 Dec 2 y y 00000 8900806 t t TL 45 22 5 T r 1 i t i i i i i x 0 225 45 675 90 0 225 45 675 9 9 Figure 4 16 Modelled TL illumination fluorescence signals Calculated fluorescence signal of an ideal QE 1 fluorophore under TL illumination on the air left and on the spacer right sides of the air spacer interface The orientation of the transition dipole of the fluorophore is defined by and 8 according to figure 4 1 At the bottom the maxima of the signals for 8 0 as a function of Finally the modelled fluorescence signals for transmitted light illumination are shown in figure 4 16 As expected from the electric field calculations the signals on both sides of the interface are very weak in comparison to the FB and FL signals The stronger signals are the ones corresponding to molecules with an important component of their transition dipole along the polarization direction x In fact in comparison to the signal of an x oriented dipole the signals of z and y oriented dipoles are negligible 4 5 Conclusions 81 Fluorescence signal of a fluorophore with T 4 0 Until now an ideal fluorophore with null T QE 1 was considere
40. considered as being formed by plane waves A microscope objective is an optical system that produces aplanatic images i e axially stigmatic and obeying the sine condition 14 null spherical aberration Hence the plane waves of the illumination beam are transformed by the objective in spherical waves with the cen ter at the Gaussian focus The maximum focusing angle y4 is determined by the numerical aperture NA of the objective and the refractive index of the focusing medium Nglass according to NA na arcsin 2 1 Nglass Figure 2 4 Focusing angles Geometric representation of the focal sphere of the microscope objective and the focusing angles for the cases of a Complete illumination of the rear lens of the objective b Annular illumination The radius of the focal sphere R see figure 2 4 a is determined by the radius of the rear lens of the objective R and the refractive index of the focusing medium Nglass Rr Ry Nglass mao sin On 4 7 NA 22 Thus if the rear lens of the objective is illuminated with an annular beam of outer radius R and inner radius Ra the maximum and minimum focusing angles 0 and 63 respectively see figure 2 4 b are given by R Rio NA 6 2 arcsin z arcsin 54 2 3 rl Tlglass The microscope objective used in all the experiments presented in this disserta tion has a NA 1 4 and the refractive index of the glass slides and the matching oil S Ng
41. data 0 00002 eae 98 5 10 kon and koff by histogram and autocorrelation methods 99 5 11 Experimental configuration 2 2 2 2 ana es 100 5 12 Distributions of P1 Kon and koff e aie Dress u 101 5 13 Distribution of hep on Tepe ri et Ks er en a 104 5 14 che asa function Eaj i 4 4 3 ha a ar Dear Dee 105 6 1 Absorption and emission of Zug aaCdossSe QDs 2 2 2 112 6 2 Image of the sample with individual QDs 113 6 3 QDs kinetic traces 4 22 et dk oa Ze ana ee 114 6 4 Histograms of the length of the on and off periods on glass 116 6 5 Histogram of the on and off times of a single kinetic trace 117 6 6 Histograms of the length of the on and off periods on ITO 118 6 7 On and off intensities 25 2 6 o oo a ee BAS OS eS 119 6 8 Power law exponent o Hk aa Oe Oe 119 6 9 Cycles per second 22 8 a ee ea eRe OS BEAD Oe eS 120 6 10 On time fraction ee a ta Ota he th Ss eee 121 6 11 Correlations of adjacent on and ofFtimes 122 6 12 Correlation coefficient vs the excitation intensity 123 6 13 Simulated and detected on and off periods 127 6 14 Simulated QD kinetic trace lt n 2 v3 8 ERG ri Fir 128 6 15 Histograms of simulated on and off times for QDs on glass 128 LIST OF FIGURES 191 6 16 Histograms of simulated on and off times for QDs on ITO 129 6 17 Correlations between simulated adjacent on and off ti
42. determination of 7 is difficult and can even depend on the alignment of the detection system For this reason the method presented here is set up for the calculation of a theoretical single molecule fluorescence signal proportional to the experimental one The influence of the environment on the fluorescence behavior of a molecule can be theoretically studied in different ways In the approach presented in this chapter the molecule is considered as a point oscillating dipole defined by its transition dipole moment p It is thought to be placed in either side of one of the outer interfaces of a plane non magnetic layered system and interacting classically with the electromagnetic field Figure 3 1 depicts the situation and introduces some of the coordinates relevant for the calculations The present chapter is dedicated to derive expressions for the rates Peze I and Inr of equation 3 1 that permit the calculation of a theoretical single molecule fluo rescence signal Expressions for the electromagnetic de excitation rates T and Prr are derived in section 3 2 The excitation rate I is considered in section 3 3 Interface boundary conditions The model presented here considers a fluorophore in the outermost interface of a plane non magnetic unitary permeability layered system Depending on the 3 2 The emission Al iZ Figure 3 1 Coordinate system The orienta tion of the transition dipole p of the fluorescent dye molec
43. effective for all dipole orientations and almost 7 times more effective for the perpendicular case Chapter 5 Influence of a nearby gold film on single molecule fluorescence dynamics The fluorescence emission of a single molecule is a dynamic process that presents bright and dark periods due to temporary excursions of the excited molecule to the triplet states The international scientific community baptized this phenomenon as fluorescence on off blinking and the traces of fluorescence emission vs time as kinetic traces First in this chapter the accepted theoretical model of molecular fluorescence blinking is presented Then two methods for the analysis of kinetic traces are pre sented and compared via Monte Carlo simulations Finally the influence of a nearby metallic film on the fluorescence blinking of single dye molecules is experimentally investigated 5 1 Single molecule fluorescence Electronic tran sition rates The simplest quantum mechanical picture that one can think of to represent the processes involved in molecular fluorescence and that can explain a dark state is a three level system as depicted in figure 5 1 a ground singlet state S0 an excited singlet state S1 and a triplet state T1 The excited triplet and singlet states share the same common molecular geometry for some energy in order to allow inter system crossing Upon absorption of a photon an electron originally in SO can be excited to S1 with a rate
44. film obtained with FB illumination Intensity 3 2 kW cm scanned area 20 x 20 um 350 x 350 pixels counting time per pixel 1 5ms The small image is a detail 250 x 250 pixels of the region marked with the white dashed line The double arrow in the small image shows the polarization direction The patterns observed through the gold film present different intensities but have all the same geometry two center bright lobes are separated by a dark gap perpendicular to the polarization direction and are accompanied by weaker bright and dark fringes at the sides This is in fact the symmetry that corresponds to the spatial distribution of the intensity of the longitudinal along the optical axis of the objective component of the electric field generated at the focus of a microscope ob jective 39 59 The observed patterns are in great contrast to what is expected for light focused through an non metallic medium no gold film because in that case and under the same illumination conditions linearly polarized light and full beam 64 Single molecule fluorescence through a thin gold film a 5 4 3 b Figure 4 4 Typical fluorescence micrographs of single dye molecules on a glass substrate without gold film a FB illumination round diffraction limited signals are observed b Annular illumi nation a variety of patterns is observed due to the different orientations of the chromophores illumination this longitudinal
45. flection of the glass substrates 6 41 7 the reflectivity increases even more and approaches unity it does not reach unity due to losses produced by the evanescent field at the glass gold interface in the absorbing gold Then the names chosen for the illumination modes are justified The TL illumination mode corresponds to an gles of incidence below e and the FL illumination mode corresponds to angles of incidence mainly above 6 At even higher angles of incidence the component parallel to the interfaces of the wavevector of the incident radiation increases and eventually matches the resonance condition to couple to surface plasmons 58 This resonant coupling to surface plas mons surface plasmon resonance SPR translates in a minimum in the reflectivity and has associated an enhanced evanescent electric field at the metal spacer inter face that decays exponentially into the dielectric medium For the bare gold sample the SPR occurs at an incidence angle of 44 1 and for a sample with a 4 bilayer spacer 24nm at 48 3 Then by comparing these angles to the ranges of incidence 3The curves were calculated by solving Maxwell equations for the layered system with a transfer matriz algorithm see section 3 2 The dielectric constants of the materials for a wavelength of 633 nm are listed in table 4 1 62 Single molecule fluorescence through a thin gold film Eee i gt 06 gt dspacer 24 nm 3 2 04
46. fluorescence blinking and excited state lifetime of single molecules was studied in the presence and in the absence of a nearby gold film in order to investigate the influence of the metal on the electronic transition rates Two methods trace histogram and autocorrelation for the analysis of single molecule fluorescence blinking were presented and evaluated The trace histogram method was improved with respect to previously reported algorithms in order to systematically find the best compromise between time resolution and accuracy in distinguishing between on and off states A Monte Carlo procedure to simulate sin gle molecule fluorescence blinking was set up and used to compare the performance of the two analysis methods The autocorrelation method was found to be more reliable specially for the analysis of short kinetic traces The influence of the nearby gold on the total decay rate Is was clearly observed Comparison to theoretical calculations showed that the observed distribution of T a in the absence of gold can be explained if the chromophores behave optically as on the air side of the interface confirming the results obtained from the intensity distribution of the fluorescence signals No influence of the gold presence on the ISC rate from the excited state to the triplet 23 was observed In contrast the gold presence produced a 2 fold in av erage increase of the transition rate from the triplet to the singlet ground state kon
47. forbidden light FL and transmitted light TL illumination are shown The white bar in the upper left image shows the polarization direction and is a scale bar of 500nm The images are plotted with grey scales in which zero is black and one half of the maximum intensity is white so that the white color saturation in the images show the size at half maximum The maximum intensity value of each component is written in the upper left corner the image To the right of each pattern a line profile is shown For the x and z components the profile corresponds to a vertical centered line For the y components corresponds to a line at 45 passing trough the center of illumination The x components have a central dominant peak and concentric side lobes of lower intensity The intensity of the side lobes is maximum along the polarization direction The y components present two nodal lines one along the polarization direction and the other along the perpendicular direction The maximum intensities lie on directions at 45 with respect to the nodal lines with four main peaks and weaker side lobes The z components present one nodal line perpendicular to the polarization direction and has two main peaks and weaker side lobes at the sides of the nodal line Under FB illumination conditions the strongest component of the electric field is the z one The x component follows and the y component is practically negligible in comparison to the other two Under
48. gave me and the confidence you deposited on me I profited greatly from your advice and enjoyed all the discussions with you Special thanks for your critical proof reading of my dissertation Prof Dr Wolfram Baumann thank you very much for supervising my disserta tion Dr Herr Volker Jacobsen thanks for the nice time in Mainz your help in all the imaginable fields such as science lab tools German translations etc The keine Experimente times are gone for good Dr Krasimir Vasilev thank you for the great work you did in preparing the samples with the polyelectrolyte spacers Will those polymers be so clean again Thank you also for the nice time in the lab and in the office and for stopping kicking me in the football matches Dr Kaloian Koynov thanks for your help during the construction of the confocal microscope Herr Pilz and his Haustechnik team thank you for helping me to adapt so quickly the laboratory for the experiments of this dissertation Herr Richter and his Elektroniklabor team thank you for you support with the electronic problems Thanks to Jennifer Shumaker and Heiko Rochholz for the scattering samples C shaped nanoparticles Andreas Scheller thank you for the computer support Dr Christian H bner thank you You were a very valuable scientific partner during these years I want to thank Tatiana Dimitrova Volker Jacobsen Kaloian Koynov Katy Lovejoy Joe and Gale Robertson Heiko Rochholz
49. grey and Poisson distributions lines with average equal to the experimental on and off intensities Right histograms of the length of the on and off periods Reproduction of the experimental data Simulations were carried out to try to reproduce the experimental results To simulate the data obtained with the different excitation intensities the correspond ing experimental parameters T Ion Ioff were used as input for the Monte Carlo algorithm For each excitation intensity a number of traces equal to the number of experimental traces was simulated and then analyzed with the trace histogram method All the traces were simulated with exponents m between 1 72 and 1 74 1000 2 2 f 2 P 2 04 kWicm 1000 P 1 61 kWicm 1000 P 0 95 kWicm 100 4 m 1 63 100 m 1 63 100 4 m 1 59 10 4 a a 104 a 104 P 3 14 gt 1 4 8 A 8 8 Pe 1 4 4 01 4 4 0 01 4 0 01 0 01 4 T 150s 0 001 saad 001 01 1 10 001 01 1 10 0 01 01 1 10 Period length s Period length s Period length s 1000 P 0 74 kWicm P 0 18 kWicm 1000 P 0 026 kWicm 1000 100 4 m 1 60 100 m 1 66 109 m 1 75 10 a 10 g 10 e c c c 14 3 14 5 1 8 3 3 01 4 0 1 0 1 0 01 aos 0 01 4 T 150s 0001 T 185s 0 001 0 001 Terre FT Te Tg reg 0 01 01 1 10 001 01 1 10 100 01 1 10 100 Period length s Period length s Period length s Figure 6 15 Histograms of the length of the on and off periods of the simulated kine
50. half maximum M4Care should be taken to scan through the center of a feature because otherwise the signal might seem to become narrow only because a part of a higher order Airy disc is being detected 2 4 Operation 33 Scan Parameters f 1m S al xd WO Shae Figure 2 18 Line scan screen Screen shot of the user interface software during a line scan FWHM is displayed A line scan is also useful to adjust the position of the APD in order to maximize the signal Data format The obtained data is stored in the form of a third order tensor The rows and columns are determined by the number of pixels of the complete image forward and backward Then each layer third dimension of the data tensor has different information stored for every pixel The first and second layers store the x and y position respectively of each pixel in units of the AD DA digits AD which can be translated into wm by _ 80 X AD Klum aeg ee The third layer stores the counts i e the number of photons detected in the pixel The fourth and fifth layers store the initial and final time respectively of each pixel in ns with a resolution of 25 ns The sixth layer contains the capacitive monitor signal for the x position of each pixel again in AD DA digits 34 The fluorescence and light scattering confocal microscope Saving Data To save the complete data the buttons Save Data in the division Save Data of the control panel shou
51. in dendronboxes Superior chemical photochemical and thermal stability J Am Chem Soc vol 125 pp 3901 3909 2003 D S English L E Pell Z Yu P F Barbara and B A Korgel Size tun able visible luminescence from individual organic monolayer stabilized silicon nanocrystal quantum dots Nano Lett vol 2 pp 681 685 2002 M Kuno D P Fromm H F Hamann A Gallagher and D J Nesbitt Nonexponential blincking kinetics of single CdSe quantum dots A uni versal power law behavior J Chem Phys vol 112 no 7 pp 3117 3120 2000 R G Neuhauser K T Shimizu W K Woo S A Empedocles and B M G Correlation between fluorescence intermittency and spectral diffusion in sin gle semiconductor quantum dots Phys Rev Lett vol 85 no 15 pp 3301 3304 2000 K T Shimizu R G Neuhauser C A Leatherdale S A Empedocles W K Woo and M G Bawendi Blincking statistics in single semiconductor nanocrystal quantum dots Phys Rev B vol 63 no 205316 2001 M Kuno D P Fromm H F Hamann A Gallagher and D J Nesbitt On Off fluorescence intermittency of single semiconductor quantum dots J Chem Phys vol 115 no 2 pp 1028 1040 2001 M Kuno D P Fromm S T Johnson A Gallagher and D J Nesbitt Modeling distributed kinetics in isolated semiconductor quantum dots Phys Rev B vol 67 p 125304 2003 T D Krauss and L E
52. in medium m kmar Km Further integration between k and infinity accounts for the contribution of high wavevector evanescent components of the dipole field which can eventually propa gate as waveguide modes in some of the layers but do not contribute to the detectable 44 Single molecule fluorescence through a layered system far field emission Taking into account that kp km sin m it is possible to calculate the emission rate in a given angular range by integrating between the appropriate limits of k To calculate the theoretical fluorescence signal the radiation emitted into the collection solid angle of the microscope objective needs to be computed 3 2 2 Total electromagnetic decay rate The normalized total electromagnetic decay rate for a dipole with dipole moment u located in a reference medium of dielectric constant er and in front of a plane layered system can be calculated by application of Poynting s Theorem as follows 30 34 fy 1 ER Ten By 1 3 7 H KR where o is the permittivity of vacuum er is the dielectric constant of reference medium and kp is the wavevector of the emitted radiation in the reference medium Ep is the back reacted electric field which is defined as the difference at the dipole position between the field in the presence of the interface and the field in the free dielectric reference medium A detailed description of the calculation of E can be found in 34 35 In the ca
53. in single InP quantum dots Nano Lett vol 1 pp 557 564 2001 E J B H Xu M Kall and L B rjesson Spectroscopy of single hemoglobin molecules by surface enhanced raman scattering Phys Rev Lett vol 83 pp 4357 4360 1999 J Jiang K Bosnick M Maillard and L Brus Single molecule raman spec troscopy at the junctions of large ag nanocrystals J Phys Chem B vol 107 pp 9964 9972 2003 P Kambhampati C M Child M C Foster and A Campion On the chemi cal mechanism of surface enhanced raman scattering experiment and theory J Chem Phys vol 108 pp 5013 5026 1998 K Bromann C F lix H Brune W Harbich R Monot J Buttet and K Kern Controlled deposition of size selected silver nanoclusters Science vol 274 pp 956 958 1996 D M Kolb R Ullmann and T Will Nanofabrication of small copper clus ters on gold 111 electrodes by a scanning tunneling microscope Science vol 275 pp 1097 1099 1997 Y Y Yu S S Chang C L Lee and C R C Wang Gold nanorods Electrochemical synthesis and optical properties J Phys Chem B Vol 101 No 84 1997 vol 101 pp 6661 6664 1997 J Bosbach D Martin F Stietz T Wenzel and F Trager Laser based method for fabricating monodisperse metallic nanoparticles Appl Phys Lett vol 74 pp 2605 2607 1999 R Jin Y Cao C A Mirkin K L Kelly G C Schatz and J G Zhen
54. in the previous paragraph figure 2 1 a In addition Minsky also recognized the possibility of employing a reflected light scheme as shown in the FIG 3 of figure 2 1 c and reproduced in figure 2 1 b In this case only a single lens on one side of the specimen is used and a half silvered mirror separates the entering and exiting rays This arrangement is equivalent to 2 1 The confocal principle 7 the symmetric one in terms of resolution but much simpler in terms of alignment and operation The price to pay for the simplicity is the fourfold brightness loss due to the beam splitter However in fluorescence measurements the efficiencies of both configurations are similar because a dichroic filter can be used to spectrally separate the excitation and fluorescence light e Illumination Detection Image pin hole pin hole Illumination pin hole Beam l splitter Light source V Illumination Detection Ta Detection pin hole Detector Figure 2 2 Confocal image formation The diffraction patterns of both illumination and detection aperture pinholes are multiplied to form the image of a point of the specimen A confocal microscope behaves as a coherent optical system in which the diffrac tion patterns of both illumination and detection pinhole apertures are multiplied to form the image of a point of the specimen 9 This gives rise to a sharpened central peak and weak outer rings with the consequent in
55. in those cases the on and off times show a very clear power law probability density It can be concluded then that the observed power law is not a product of the analysis of a non Poissonian on intensity The power law reflects the 6 5 Conclusions 133 statistics of an underlying real process A Monte Carlo procedure was set up to simulate the blinking of the QDs The model behind considers a random switch between an on and an off state both with respective Poisson distributed intensities and an additional independent pathway from the on state to the off state to account for the photo induced shortening of the on times This independent on off transition introduces an exponential probability for the on times with a consequent characteristic lifetime for the on state The simulated kinetic traces do not show any correlation between adjacent times as it is expected because the time periods are randomly generated This gives evidence that the on and off times are not completely random and that the physical processes behind the blinking needs to introduce a slight memory effect It was suggested that the deviations of the on intensity from the Poisson behavior was due to the fact that very short on times which are highly probable due to the power law probability were partly detected 119 The simulated kinetic traces show a Poisson distributed on intensities in all cases This allows to refute the hypothesis of the partly detected on times a
56. its domain by means of an arbitrary minimum time tmin Given that for very low excitation intensities the on time fraction is close to 0 5 figure 6 10 and that P t and P torr become practically indistinguishable the same tmin was used for the ton and the t probabilities Then the time t spent by the QD in state j on or off has a probability density P t given by m 1 _ P t m i RE Ge 6 4 In addition to the random power law distributed on off blinking an extra in dependent pathway from the on to the off state is included in the simulations to explain the experimentally observed photo induced PT reduction of the on time fraction This transition is considered to occur at a constant rate and is therefore exponentially distributed with a characteristic time Tpz Pp1 ton etler 6 5 PI ton TPI It is important to note that although it is not possible to assign an average on time because of the power law distribution which normalization depends on the arbitrary tmin the extra exponentially distributed pathway allows to assign an av erage photo induced on time Tpz 6 4 2 Monte Carlo procedure Based on the general assumptions described above a Monte Carlo procedure was set up to simulate a blinking QD with the following input parameters T total simulated time m exponent of the power law probability density I background intensity figure 6 7 a 6 4 Modelling the QDs blinking 125
57. of interest generated by 50 Single molecule fluorescence through a layered system a such a wave it is necessary to multiply the normalized expressions 3 21 by the corresponding p or s component of the source wave Fou E 0 di i 21s ER E 0 di i 215 cos Y Ej Eoy E 0 di i 21s Erc Ej 9 di i 215 sind 3 22 Foz ES 0 di i 21s EB E 0 di i Zis cos Y This is the electric field written in the primed coordinates system generated at the point of interest of the layered system by a source plane wave focused by a point of the rear lens of the objective defined by y and 6 Thus in order to obtain Eo Ej has to be rotated to the original xyz coordinates system cosy sind 0 Eo cosy Eoy sin Y Eo Ey siny cosy 0 Eox siny Eoy cosw 3 23 0 0 1 Foz Finally the x y and z components of E can be written as a function of 6 and w as Eve E 8 cos Y E 5 0 sin Y Eoy EE 0 E5 0 cos w sin Y 3 24 Eo EE 0 cos y where the dependencies on d e and zj were not explicitly written Next to calculate the integral 3 19 in an arbitrary point near the geometric focus an expression for k r is necessary According to figure 3 3 one can write pcos Y cos Ysin 0 r psing k k sinwsin 3 25 Zp cos 0 and the scalar product leads to k r k psin 0 cos w p zp cos 0 3 26 The microscope objective is an aplanatic system of revolution
58. of the excitation light is opened and data is collected as explained in the previous section However in this case the detected light should be directed to the VPH grating either by a mirror or a beam splitter depending on the experimental requirements The detected light is spectrally dispersed by the grating and detected by the CCD camera which has its own control and data acquisition software provided by the manufacturer 22 The images acquired with the CCD camera contain the spectral 38 The fluorescence and light scattering confocal microscope information and can be analyzed with another home made software which is not discussed in this text If the VPH grating and the CCD camera are properly aligned and focused see section 2 3 4 the spectral response is linear through the whole range of the grating After every realignment of the instrument it is necessary to check the calibration of the spectrograph This can be accomplished simply by directing monochromatic light of different wavelengths to the spectrograph measuring the position of those wavelengths on the CCD sensor and obtaining a calibration curve After a sample change or small alignment corrections it is recommended to verify the calibration curve with one wavelength This can be easily done by measuring the position of one laser line on the CCD sensor 19The file is called CCDAnalysis pxp and is attached in a CD ROM to the copy of this dissertation in the MPI P lib
59. of the QDs is not a product of the underlying power law probability of the on and off times and that the blinking of QDs occurs between a non emitting off state and a distribution of emitting on states with different intensities In order to account for the photo induced shortening of the on times an inde pendent single rate transition from the on to the off state was introduced in the simulations Exclusively by means of this transition it was possible to reproduce the experimentally observed photo induced effects shortening of the on time fraction and maximum on time and increase of the cycles per second and to assign a char acteristic photo induced lifetime tp to the on state of QDs on glass and on ITO The QDs on glass presented a Tp proportional to P suggesting the presence of a one photon process In the case of the QDs on ITO the dependence of tp on P could be reasonably fitted by an single exponential Light scattering images and spectra of colloidal and C shaped gold nano particles were acquired From the studies on colloidal gold nano particles it was possible to determine that the minimum size of a metallic scatterer detectable with the SCOM lies around 20 nm The C shaped gold nano particles were studied under excitation with two polar ization states Some particles showed two distinct resonances for the different polar ization states and some particles did not These experiments showed the potential of the SCOM to investig
60. of triplet blinking is explained in section 5 1 40 Single molecule fluorescence through a layered system radiative and non radiative de excitation rates of the molecule and n is the overall detection efficiency including the collection efficiency of the optical system and the quantum efficiency of the detectors The ratio nz represents the radiative thus detectable fraction of the de excitation It is important to note that for the problem considered here Fnr has two different components First every fluorophore has an intrinsic non radiative de excitation rate T due to the internal relaxation pathways of the molecule which might depend on experimental conditions such as the surrounding medium for example solvent and pH and the temperature Second the nearby layered system can introduce additional non radiative decay channels such as excitation of evanescent modes or direct energy transfer This leads to an electromagnetic non radiative de excitation rate I With these considerations equation 3 1 writes i oo TD The rer 3 2 The intrinsic non radiative decay rate does not depend on the interactions of the fluorophore with the electromagnetic field Therefore the model presented here can only account for it by introducing an intrinsic fluorescence quantum efficiency as an additional parameter The electromagnetic non radiative decay rate can be calculated with the model as presented in the next section Accurate
61. or photon bunching Based on this model under constant excitation and far from optical saturation Tai gt Tig the fluorescence intensity J emitted by a single molecule while it cycles in the singlet subspace is given by the product of the radiative decay rate and the fractional instant population of the singlet excited state Ty2 Bere T23 1 Voie 5 1 And the fluorescence intensity while the molecule resides in the triplet subspace is zero However in a real experiment only a fraction of the emitted photons are detected and there is always a background intensity J Then the experimental intensities while the molecule cycles in the singlet or the triplet subspace called Zon and Ioff respectively are Dio In T2 HI 5 2 ane Popol gig la gt Torf Tig 5 3 where T is the fraction of the radiative decay rate emitted into the detector 5 1 Electronic transition rates 87 It is also possible to derive the transitions rates from the on state to the off state and viceversa Following the nomenclature proposed in 66 the rate for the transition from the off to the on state will be called kon the transition rate from the on to the off state will be called koff Then kon is simply the relaxation rate of the triplet state 3 and koyp is the product of the fraction of time that the molecule spends in the singlet excited state S1 and the probability to jump to the triplet state T1 kon L31 5 4 Tja
62. presence of a one photon photo induced process responsible for the shortening of the on times Within the frame of the ionization model this 6 4 Modelling the QDs blinking 131 would mean that on glass a one photon photo ionization process dominates the shortening of the on periods it P ITO a gooxio 7 E a opi 10 eee amp Glass 500 hg double exp fit 400 10 7 A 300 200 100 0 00 O5 10 15 20 P kWiem P kW em c 10 riss 4 ER I 2g Figure 6 18 Photo induced lifetime of the ee aR i on state Tpz used to reproduce the glass and ITO experimental data with the Monte Carlo ER model for each excitation power P a Linear plot b Log linear plot c Log log plot 10 4 power law fit m 1 3 456 2 3 456 2 0 1 1 P kWicm No trace of a P dependence of Tp is found for the QDs on ITO indicating that a completely different mechanism is ruling the photo induced shortening of the on times The approximately exponential behavior of Tp with P leads to a photo induced rate koff PI X a c constant 6 11 No physical picture was found for this dependence For example a thermally activated process cannot account for it assuming that koff pr corresponds to a AE kT and therefore a linear heating thermally activated process then koff pr x e of the QD or the ITO substrate by the excitation intensity would not explain the observed exponential behavior 132 Ph
63. spheres Optimization of the ion milling process 8 minutes J 0 1mA V 100V is important for removing the gold film while preserving the structure of the polystyrene sphere i e avoiding sphere melting Finally the PS spheres were removed lift off from both sides of the coverslips by using adhesive tape Scotch Magic Tape 3M Inc The diameter of the colloidal spheres controls the diameter of the metallic C shaped nano structure and the evaporated metal film thickness determines the nanopar ticle thickness The rotation angle between the two evaporations determines the opening angle of the C structure Measurement Light scattering images and spectra were acquired with the home built SCOM described in 2 set up for light scattering measurements A blocking disc with a diameter of 6mm was used to produce annular illumination section 2 2 1 and the pinhole size was adjusted in order to optimize the signal to background The images and the spectra were acquired under white light illumination Xe arc lamp A polarizing film was used to linearly polarized the illumination beam 142 Light scattering from single metallic nano structures 7 3 2 Images and spectra of C shaped gold nanoparticles Figure 7 5 shows an SEM image of the gold C shaped nano structures top right and a light scattering image bottom right In contrast to the spherical and much smaller gold nanoparticles the signals from the individual C shaped nano particl
64. the cases without and with Au respectively The T 3 of an arbitrarily oriented molecule was calculated by equation 3 5 For the case of chromophores on the AIR side of the interface the values of I2 and I g1 were obtained by multiplying the theoretical values shown in table 5 3 by a factor equal to the ratio of the MINIMUM detected T21 in the sample without gold and the theoretical value for the parallel dipole F21 0 35 1 34 0 26 For the case of chromophores in the spacer side of the interface the theoretical values were multiplied by a factor equal to the ratio of the MAXIMUM detected T in the sample without gold and the theoretical value for the parallel dipole T 0 63 1 34 0 47 5 4 Influence of a nearby thin Au film on the electronic transition ratk 3 the conclusion made in chapter 4 that the chromophores behave optically as being in the air side of the spacer air interface If the chromophores were in the polymer side due to the electromagnetic boundary conditions at the interface see section 3 1 the parallel molecules would have the same ea and the perpendicular ones would have a Tia 0 41 This predicts a range of 2 two times broader than the observed experimentally and a distribution with the opposite asymmetry i e richer in higher values of Ta A simulated distribution of this situation is shown in figure 5 12 inset in the graph of I for the case without gold The molecules in the sample with th
65. the electronic transition rates 101 5 4 1 Influence on Da 2 22mm 102 9 4 2 Influence on Agee ogo Sa ans Be DER SE 103 34 3 Influence On kon lt ca oe el A ee aE GSS Ee 105 KO GONE USNS e th ae ys ra AR oe OS Gi ae BER BES 106 6 Photoluminescence blinking of Zn Cd Se nano crystals 109 6 1 Brief Introduction and current status 00 110 62 Experimental 2 2 02 2122 a a ws ara kr art hat 112 6 2 1 Sample preparation lt a vn vu Doe eo oe 112 6 2 2 Measurement aa st a Near 113 bio QD Kinetietracess ag au a ei Bra rar Ra Hate 114 6 3 1 General characteristics a bcm Wer Veen Bl doe Gh 114 6 3 2 Effect of the excitation intensity 115 6 4 Modelling the QDs blinking 2 22 nu 2 258 Kanada wed es 123 6 4 1 Blinking model 2 a a be ee ee beh a eh ee a 124 6 4 2 Monte Carlo procedure 6 2 eg ee a Dee ES eS 124 CONTENTS ili 6 4 3 Simulated blinking nek 2 Shoes ee kee Ge ce wk 6 5 Conclusions 2 2 mom 7 Light scattering from single metallic nano structures fl An TOUCH es A a ay Ste ye ee we 7 2 Light scattering of individual colloidal gold nanoparticles 112 1 IESPERTIeNntel earen iy Brae OIG et ts hl Bae Stes Gh Ad Bp este 7 2 2 Images of colloidal gold nanoparticles 2 2 7 2 3 Spectra of colloidal gold nanoparticles 7 3 Light scattering of individual C shaped gold nanoparticles 7 3 1 Experimental 4 4 2 0 hu a ee Oe a ae ee te 7 3
66. the fluorescence of single molecules through a thin metallic film If that is the case how close to the gold film can the molecules be before their emission becomes undetectable How does this experimental scheme compare to the detection from the air side and to the case without the gold film 4 2 Experimental The sample geometry is depicted in figure 4 1 together with the illumination modes employed in the experiments In this section details about the sample prepa ration and the measurement procedures are given 4 2 1 Sample preparation The sample preparation consists of five steps cleaning of the glass substrates deposition of the gold film functionalization of the gold surface layer by layer deposition of the polyelectrolyte spacer and finally deposition of the fluorescent dye molecules Thin 0 13 0 16 mm glass coverslips N 1 Menzel Gl ser were cleaned suc cessively with Hellmanex 2 Milli Q water and ethanol gt 98 Riedel de Ha n In addition to remove any rest of organic material the coverslips were heated in air for two hours at 500 C Next a 44nm gold Agar Scientific Ltd film was thermally evaporated onto the substrates thermal evaporator Edwards FL400 The freshly prepared gold surfaces were functionalized with a self assembled monolayer SAM of 3 mercaptopropionic acid MPA Aldrich as the substrates were immersed in a 0 03 M Milli Q water solution of MPA for one hour After that the substrates
67. the other excitation intensities both for QDs on glass and on ITO substrates No evident correlation is observed when the graphs are plotted on a linear scale upper two rows in figure 6 11 Nevertheless when the graphs are plotted in a logarithmic 122 Photoluminescence blinking of Zno 42Cdo53 Se nano crystals ton S vs next ton S toft S VS next tor S ton S vs next tor S 1 24 08 0 10 20 30 40 0 10 20 30 40 Figure 6 11 Correlations of adjacent on and off times plotted on linear first and second rows and logarithmic scales The linear correlation coefficient of the logarithm of the times is shown in the corresponding graphs The scatter plot of tof vs the next ton not shown is similar to the one of ton vs the next tof also in that case no correlation was found scale a noticeable correlation is observed between consecutive on times and between consecutive off times Remarkably no correlation is observed between adjacent on and off times In order to quantify the linear correlation between the logarithm of adjacent times the Pearson correlation coefficient R was calculated as 118 Denen R 6 3 ER _ N po Y R measures the strength of the linear relationship between two variables by taking values between 1 and 1 inclusive This limiting cases indicate that all the x y points of a scatter plot can be connected by a straight line which slope is 1 or 1 respectivel
68. the x and z components of the electric field along the x axis and the x component along the y axis 7 E and EY equations 3 29 3 31 and 3 30 respectively by E p p Zp E3 p cos p E p sin p E r 9 9 2P Ey p 9 2p Erle E2 e sin y cosp 3 37 E p p zp E7 p cos p In order to calculate the fluorescence excitation rate of a single molecule it is necessary to write an expression for the projection of this electric field on the direction of the transition dipole of the molecule u equation 3 18 If the orientation of u is defined by an azimuth angle amp and a polar angle 3 as shown in figure 3 1 the projection of the excitation field on the dipole direction is E r E r sindcos 8 E r sin sin 8 E r cos 3 38 Finally recalling equation 3 18 the fluorescence excitation rate for a molecule positioned at r and with a transition dipole oriented according to amp and is Deac t 8 x E r sin cos 6 E r sin sin 8 E r cos 3 39 3 4 Single molecule fluorescence signal At this point it is possible to recall equation 3 1 and combine the results obtained for the excitation and the emission in order to write an expression for the theoretical fluorescence signal of a fluorophore in the outermost interface of a layered system Is r 6 8 Teac r Q 8 Tac 6 3 40 where Perc is calculated by equation 3 39 and Paet by equation 3 17 Because of the uncertai
69. were thoroughly rinsed with Milli Q water in order to remove all unbound MPA The so prepared surfaces contain free carboxylic groups which hydrolyze in water to render the surface negatively charged The next step is the layer by layer deposition of the polyelectrolyte spacer which was accomplished based on the procedure published by Decher 55 Successive polyelectrolyte layers of positively charged poly allylamine hydrochloride PAH MW 70 000 Aldrich and negatively charged poly styrene sulfonate PSS MW 70000 Aldrich were deposited onto the functionalized gold Both polyelectrolytes were deposited from 0 02 monomol l water solutions Salts were added to the solutions to adjust the ionic strength MnCl 99 Merck 0 5 M was used for the PSS solution and NaBr 99 Aldrich 2M for the PAH 56 60 Single molecule fluorescence through a thin gold film The acidity of both polyelectrolyte solutions was adjusted to pH 2 by adding HCl To perform the deposition the samples were alternatively immersed in the poly electrolyte solutions for 20 minutes starting with PAH In between immersions the samples were thoroughly rinsed with water and dried with nitrogen Samples with 1 2 3 and 4 PAH PSS bilayers were prepared corresponding to spacer thicknesses of 4 5 10 15 24nm respectively as determined in earlier studies 46 57 Before use both polyelectrolytes were purified in order to minimize the fluorescence background PAH was puri
70. with the x axis along the polarization direction of E and the z axis along the optical axis as shown in figure 3 3 a For simplicity a source field with unitary intensity is considered Esre 1 0 0 Plane waves that are focused by a certain point of the rear lens of the objective strike the layered system with a given incidence angle and defined p and s polarized components see figure 3 3 The radial position of the focusing point R defines the incidence angle of the focused wave as explained in section 2 2 1 The angular position of the focusing point Y defines the p and s polarized components of the focused wave The local electric field in any point of the layered system produced by a plane 3 3 The excitation 49 a Focusing Lens Focal plane b p yt E N ExE E F ab E 1 0 0 SLAN 2 dy x Object space Image space JB Figure 3 3 Electric field near a geometric focus a Geometry of the problem of the calculation of the electric field near a geometric focus The source field Esrc is a linearly polarized monochro matic plane wave front k 27 X A Cartesian coordinate system is taken such that the source field lies along the x axis and the optical axis along the z direction Focusing is accomplished by an aplanatic system Rays focused at different radial positions R of the focusing lens are focused with a different angle 6 b Rays focused at different angular positions of the lens
71. y are focused with different p and s polarized components EP and E respectively wave incident with angle 0 with respect to z and with given p and s components of its electric field can be calculated with the transfer matrix algorithm TMA see section 3 2 The TMA calculates the fields in a primed coordinate system rotated around the z axis in order to place the y axis along the s polarized component as shown in figure 3 3 b and 3 2 In this rotated coordinate system the p and s polarized components of the source field are EP cosy E siny 3 20 It is convenient first to use the TMA to obtain expressions for the electric field components at the point of interest generated by an incident plane wave with both p and s components of unitary intensity Let those field components in the rotated P pP 2 g s coordinate system be E and E For a given wavelength the expressions of these normalized field components are a function exclusively of the layered system properties thickness d and dielectric constant e of each layer and the z position in the layered system z and the angle of incidence 6 E E5 0 di Ei Zis yi E 0 di i 21s 3 21 oe EE 0 di i Zis Bue A wave of E focused by a given point of the rear lens at an angular position defined by w has a p and an s polarized components given by equations 3 20 Then to obtain the components of the electric field at the point
72. 000 V L Colvin M C Schlamp and A P Alivisatos Light emitting diodes made from cadmium selenide nanocrystals and a semiconducting polymer Nature vol 370 pp 354 357 1994 200 BIBLIOGRAPHY 78 79 80 88 89 N Tessler V Medvedev M Kazes S Kan and U Banin2 Efficient near infrared polymer nanocrystal light emitting diodes Science vol 295 pp 1506 1508 2002 V I Klimov A A Mikhailovsky S Xu A Malko J A Hollingsworth C A Leatherdale H J Eisler and M G Bawendi Optical gain and stimulated emission in nanocrystal quantum dots Science vol 290 pp 314 317 2000 M V Artemyev U Woggon R Wannemacher H Jaschinski and W Lang bein Light trapped in a photonic dot Microspheres act as a cavity for quan tum dot emission Nano Lett vol 1 no 6 pp 309 314 2001 M Bruchez M Moronne P Gin S Weiss and A P Alivisatos Light emitting diodes made from cadmium selenide nanocrystals and a semicon ducting polymer Science vol 281 pp 2013 2016 1998 W C W Chan and S Nie Quantum dot bioconjugates for ultrasensitive nonisotopic detection Science vol 281 pp 2016 2018 1998 M Y Han X H Gao J Z Su and S M Nie Quantum dot tagged mi crobeads for multiplexed optical coding of biomolecules Nat Biotechnology vol 19 pp 631 635 2001 B Dubertret P Skourides D J Norris V Noireaux A
73. 1 s loff 1 s Table 5 2 Comparison of the ton s analysis results of the simulated toff s data obtained by the autocor relation and the trace histogram Kon 1 s methods koff 1 s Toff ms Ton ms 5 9 d presents the histogram of the length of the simulated on and off periods The results obtained from the analysis of the simulated data with the autocorre lation and trace histogram methods are presented in table 5 2 Again both methods provide values for kon and korr similar to the real ones but still different To test the reliability of the two methods to retrieve the values of kon and koff kinetic traces of different total times 10 for each time were simulated with the input parameters shown in figure 5 8 and analyzed with both methods The average values obtained for kon and korr are plotted in figure 5 10 the error bars denote plus minus one standard deviation The longer the kinetic traces the higher the number of on off cycles and therefore the results become more accurate with both methods The trace histogram method retrieves systematically slightly lower values of koff Even though the statistical errors are similar for both methods the average values of kon and korr provided by the autocorrelation method are closer to the actual values note that the first point of korr obtained by the TH method is out of the range of the plot The application of the autocorrelation method as presented here is li
74. 2 Images and spectra of C shaped gold nanoparticles Bi CHR STON ne Yes dg ase se ee ae op es 8 Summary A Set up control and data acquisition software AT AD Basic routines Asa cian AO tae ge fee a ae he he ve ey SS a Bock AZ SCOP TOUR AR EB AA Ka BBY A CFE Toms ae rare List of tables List of figures Abbreviations Bibliography Acknowledgements Curriculum Vitae 127 132 135 135 136 136 137 139 140 140 142 143 145 149 149 163 180 187 191 193 195 205 207 Chapter 1 Introduction In 1974 Fleischmann and co workers 1 wanted to perform experiments on pyri dine by combining electrochemistry and Raman spectroscopy In order to increase the Raman signal they deposited the pyridine onto a roughened silver electrode The idea was to increase the surface area of the electrode and therefore the amount of adsorbate on the sample It worked the Raman signal was indeed greatly in creased Three years later Jeanmaire and van Duyne 2 as well as Albrecht and Creighton 3 recognized independently that the large intensities observed could not be accounted for simply by the increase in the number of scatterers present They proposed that an enhancement of the scattered intensity occurred in the ad sorbed state Already at that time a surface plasmon enhancement mechanism of the scattered intensity was proposed 3 4 Since then this effect was called surface enhanced Raman scattering SERS and ca
75. 2 c include XOPStandardHeaders h Include ANSI headers Mac headers IgorXOP h XOP h and XOPSupport h include ADWinlnit2 h include C ADwin Developer C Microsoft_VisualC Adwin c include C ADwin Developer C Microsoft_VisualC Adwin h All structures are 2 byte aligned if GENERATINGPOWERPC pragma options align mac68k endif ifdef WINDOWS pragma pack 2 endif static void XOPEntry void long Xi Yi char varname switch GetXOPMessage case CMD GetLong amp Xi Initial X GetLong amp Yi Initial Y Set Par 1 Xi Set Par 2 Yi ADBStart 4 break CMD is the only message I care about main ioRecHandle This is the initial entry point at which the host application calls XOP The message sent by the host must be INIT main does any necessary initialization and then sets the XOPEntry field of the ioRecHandle to the address to be called for future messages HOST_IMPORT void main IORecHandle ioRecHandle A 3 C routines 183 ifdef XOP_GLOBALS_ARE_A4_BASED ifdef _MWERKS__ For CodeWarrior 68K XOPs SetCurrentA4 Set up correct A4 This allows globals to work SendXOPA4Tolgor ioRecHandle GetA4 And communicate it to Igor endif endif XOPlInit ioRecHandle do standard XOP initialization SetXOPEntry XOPEntry set entry point for future calls SetXOPResult OL All structures are 2 byte aligned if GENERATINGPOWE
76. 2004 Wavemetrics inc IGOR v 4 02 http www wavemetrics com Last ac cessed in Jan 2004 PCO imaging GmbH http www pco de Last accessed in Jan 2004 Melles Griot inc Shear plate collimation tester theory http beammeasurement mellesgriot com tut_ shear_ plate test asp Last accessed in Jan 2004 P I GmbH Tittle Wavemetrics inc Igor Pro v4 Programming and reference manual 2000 A Sommerfeld ber die ausbreitung der wellen in der drahtlosen telegra phie Annalen der Physik vol 28 no 4 pp 665 736 1909 R R Chance A Prock and R Silbey Molecular fluorescence and energy transfer near interfaces Adv Chem Phys vol 37 pp 1 65 1978 H Weyl Ausbreitung elektromagnetischer wellen ber einem ebenem leiter Annalen der Physik vol 60 no 22 pp 481 500 1919 G W Ford and W H Weber Electromagnetic interactions of molecules with metal surfaces Phys Rep vol 113 no 4 pp 195 287 1984 W Lukosz Theory of optical environment dependent spontaneous emission rates for emitters in thin layers Phys Rev B vol 22 no 6 pp 3030 3038 1980 R E Collins Field Theory of Guided Waves John Wiley and Sons 1990 R J Potton Reciprocity in optics Rep Prog Phys vol 67 pp 717 754 2004 33 P Yeh Optical waves in layered media John Wiley and Sons 1st ed 1988 BIBLIOGRAPHY 197 34 39 36 43
77. 3 Detection un au ods ee re ale hy be Se i 16 2 2 4 Time Correlated Single Photon Counting 19 2 2 5 Computer controls sas arena trat 21 23 ANGE oo Sur area ip re ae PS Ae SESE OS 22 2 3 1 Light coupling into the single mode fiber 22 2 3 2 Collimation of the illumination beam 23 2 3 3 Alignment of the dichroic mirror and the microscope objective 25 2 3 4 Alignments in the detection ooo a 0004 26 24 VOPETati n Era ar GR cael ee ad ae A a A 29 2 4 1 Sample requirements and mounting 30 A Oy SMV VIN Ao nn eo u a be ahd eh a an see AS rn 30 2 4 3 Time correlated measurements 0 4 34 2 4 4 Spectra measurements 0 0 0002 eee 37 3 Single molecule fluorescence through a layered system 39 3 1 Description of the problem 24206 lt p Pe ewe ana 39 3 2 Ihe EMISSION Ay are A a a Hees BRR RS Ee ee re ar 41 3 2 1 Radiative decay rate to different regions of space 42 3 2 2 Total electromagnetic decay rate 2 2 a 44 3 2 3 Non radiative electromagnetic de excitation rate 2 46 3 2 4 Detectable fraction of the de excitation rate 2 46 3 3 The excitation Aw Ba a So Nomi ad ch oe Wek AS Unt dens 47 3 3 1 Electric field distribution near a geometric focus in a layered SVS UCI eae ee ee a ee et 47 3 4 Single molecule fluorescence signal 2 2 222mm nennen 54 3 5 Conclusions as ed he gt Pe ie e ne Se ne eee it G
78. 60 80 100 120 140 On period length ms On period length ms Figure 5 9 Analysis of the simulated data Autocorrelation and trace histogram analysis of the simulated data shown in figure 5 8 a Autocorrelation b Inter photon times histogram c histogram of the kinetic trace with optimum bin width 0 66 ms and poisson on and off population of number of photons per bin The optimum threshold 9 photons is marked by the horizontal line d Histograms of the length of the on and off periods of the simulated kinetic trace 5 3 Experimental 99 TH AC 300 300 200 koff 1 s roh rq rh eal ro koff AC 1 s N 8 es ee ro rota noo m H 100 100 gt q g 80 g o 80 s 7 a lt A 5 fa 1 60 60 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Simulated time s Simulated time s Figure 5 10 Average values of kon and koff retrieved by the trace histogram TH and autocor relation AC methods from the analysis of simulated kinetic traces of different length simulated time only difference between the samples is the presence or the absence of the nearby gold film In both cases the fluorescence emission as a function of time as well as the ex cited state lifetime were recorded from single molecules in order to gain information about the influence of the nearby gold film on the electronic transitions rates 5 3 1 Sample preparation The samples with gold film were prepared with a 4 bi
79. CONFOCAL MICROSCOPY APPLIED TO THE STUDY OF SINGLE ENTITY FLUORESCENCE AND LIGHT SCATTERING Dissertation zur Erlangung des akademischen Grades des Dr rer nat im Fachbereich Chemie der Johannes Gutenberg Universitat Mainz vorgelegt von Fernando D Stefani geboren in Buenos Aires Argentinien Mainz Juli 2004 Dekan Prof Dr R Zentel 1 Gutachter Prof Dr W Knoll 2 Gutachter Prof Dr W Baumann 3 Gutachter Prof Dr B Hecht Universitat Basel Miindliche Prifung am 12 November 2004 Diese Arbeit wurde in der Zeit von September 2001 bis Juni 2004 unter der Betreuung von Prof Dr W Knoll und Dr M Kreiter am Max Planck Institut f r Polymerforschung in Mainz angefertigt Finanzielle Forderung Bundesministerium fiir Bildung und Forschung Nano Nachwuchsgruppe Nr 03N8702 The work for this dissertation was carried out between September 2001 and June 2004 under the direction of Prof Wolfgang Knoll and Dr Maximilian Kreiter at the Max Planck Institute for Polymer Research in Mainz Germany Financial support Bundesministerium f r Bildung und Froschung Nano Nachwuchsgruppe No 03N8702 Contents 1 Introduction 1 2 The fluorescence and light scattering confocal microscope 5 2 1 The confocal principle 6 2 22 32 02 8 28 ra 5 2 2 Description of the home built confocal microscope 8 2 2 1 Light sources and illumination 44 0434 Ga ee Ga 8 Pie CAMINO ed De Det 12 22
80. Ds 92 96 Although the large number of investigations dedicated to these colloidal QDs the nature of the emitting state is still controversial In comparison to CdSe and CdS bulk exciton electron hole pair recombination the emission from colloidal QDs is much red shifted and has an extremely long radiative lifetime 1s at 10 K compared to Ins for bulk 91 97 Parabolic band theory cannot explain this data in terms of recombination of internal states and it was proposed that band edge emission of II VI semiconducting colloidal QDs arises from the recombination of weakly overlapping surface or defect localized carriers 91 97 98 Nevertheless more careful band structure calculations for I VI semiconducting nano particles of different shapes show that the exciton ground state is not optically active it has angular momentum projection 2 This can explain the experimental observations without introducing surface or defects states Higher energy states of the exciton are produced upon absorption of one photon which then thermalize to the optically forbidden ground state from which radiative recombination has a long decay time 99 Experiments supporting the existence of the optically passive exciton state have been reported 100 101 With the advent of photoluminescence studies on a single QD level severe fluc tuations in the photoluminescence emission were found 102 This luminescence fluctuations also called on off blinking depend
81. FL illumination conditions the main peaks of all components are slightly sharpened their intensity is reduced and the intensity of the side lobes increased Because the y and z components are mainly generated by waves focused with great angles of incidence the intensity of the x component is the most reduced This effect of the annular illumination is well known and was used to equalize the intensities of the in plane and out of plane components of the fields in order to facilitate the determination of the three dimensional orientation of single molecules 39 Under TL illumination conditions the results are naturally 4 4 Modelling the experimental scheme 77 X 1 16 250 600 2 200 oo 150 FB 100 200 4 50 0 0 0 00 05 10 15 20 00 05 10 15 20 00 05 10 15 20 um um um 2504 am 200 4 12 150 150 4 8 100 FL 100 4 s04 50 0 00 05 10 15 20 00 05 10 15 20 um um 0 16 12 0 12 8 0 08 TL 4 0 04 0 00 o 00 05 10 15 20 00 05 10 15 20 00 05 10 15 20 um um um Figure 4 13 Electric field distribution in a sample without the gold film Theoretical calculations of the square modulus of the x y and z component of the electric field on the air side of the spacer air interface of a layered system like the one of the samples but without the gold film The results are presented in the same fashion as in figure 4 12 with the difference that the profiles of the x components correspond here to a horizontal centered line complementary
82. I Py Then several processes can occur The excited molecule can decay directly to the singlet ground state both radiatively and non radiatively with rates P21 and P1 nr respectively The molecule can perform many cycles like this 86 Single molecule fluorescence dynamics 3 T1 1 S0 Figure 5 1 Three level description of molecular fluorescence SO and S1 are the ground and excited singlet states respectively and T1 is the triplet The levels are located vertically in a schematic energy E axis The thin lines represent the vibrational states of each electronic state in the singlet subspace and emit a number of photons i e the molecule is bright In addition if there is a mechanism to unpair two electron spins such as spin orbit coupling and the excited singlet state SO shares for some energy the same molecular geometry with the triplet T1 the excited molecule may with a probability given by T a3 undergo inter system crossing ISC to the lower energy triplet state T1 Due to spin selection rules the singlet triplet transitions are radiative forbidden 65 The molecule is then trapped in the triplet state until ISC occurs again During this time in the triplet subspace the molecule remains dark If the triplet lifetime is long enough slow T 3 these bright and dark periods can be observed when the fluorescence of a single molecule is followed in time This effect is known as fluorescence or triplet blinking
83. Interacting with a metal Pere is the excitation rate from the singlet ground state SO to the singlet excited state Sl T is the radiative decay rate Ti the intrinsic non radiative decay Near rate and T r represents the electro metal magnetic non radiative decay channels introduced by the metal affect both the excitation and the de excitation rates as schematically represented in figure 4 8 The excitation rate is modified via changes increase or reduction in the local strength of the electric field For example under surface plasmon resonance con ditions 58 the electric field intensity at the metal dielectric interface can reach enhancements of several orders of magnitude This enhancement strongly depends on the experimental geometry the dielectric constants of the materials and the frequency of the radiation 61 The emission of a molecule can be characterized by its radiative and non radiative decay rates The presence of a nearby metallic surface can influence both The radia tive decay rate I can be enhanced or suppressed depending on the separation dis tance to the metal and on the orientation of the transition dipole of the molecule An enhancement of the radiative decay rate can lead to an increase of the observed quan tum yield and shortening of the observed lifetime 62 The non radiative decay rate is modified because the metal introduces additional electromagnetic non radiative decay channels which involv
84. Jennifer Shumaker and Krasimir Vasilev for proof reading parts of my thesis 206 Acknowledgements Dr Rodolfo Acosta thank you for the mates in between experiments I want to thank all the AK Knoll members for the nice working atmosphere at the MPI P and the unforgettable moments in Mainz I want to thank in particu lar Steffen Berg Tatiana Dimitrova Stuart Fraser Lisa Henke Volker Jacobsen Thomas Jakob Gleb Jakubov Toby Jenkins Ralf Ktigler Alessandro Manni Bern hard Menges Thomas Neumann Kirstin Petersen Rashmi Sahoo Marco Stemmler and Angela Vogt Special thanks to my father Daniel my mother In s my brothers Gustavo and Julian and my sister Leticia Thank you for your love and being there for me always Finally but above all thank you Silke for your support and love and for being my partner in everything Curriculum Vitae PERSONAL INFORMATION Name FERNANDO DANIEL STEFANI Gender Male Date of birth 19th of November 1975 Place of birth Buenos Aires Argentina Citizenship Argentine and Italian EDUCATION 1989 1994 1995 1997 1997 2000 2000 2004 High school studies at Instituto Tecnol gico Philips Argentina Awards silver medal 1992 and honours 1994 in the Olimipiada Argentina de Quimica Title Electromechanic Technician Chemical Engineering studies at Universidad Tecnol gica Nacional Buenos Aires Average grade 7 7 scale from 0 to 10 Materials Engineering studies at In
85. Pixels Yf gt PAR_4 SetPar 5 Pixeltime Scanrate gt PAR_5 SetPar 8 Y Y gt PAR_8 check 1 switch scantype case 0 ADBStart 1 check GetPar 11 while check 0 check GetPar 11 DataTransfer transferdatawaveF transferdatawaveB Pixels k break case 1 ADBStart 1 check GetPar 11 while check 0 check GetPar 11 DataTransferLine transferdatawaveF Pixels break static void XOPEntry void long Xi Yi Scanrange Pixels Pixeltime Y scantype k waveHndl transferdatawaveF transferdatawaveB switch GetXOPMessage case CMD GetLong amp Xi Initial X 186 Set up control and data acquisition software GetLong amp Yi Initial Y GetLong amp Scanrange GetLong amp Pixels GetLong amp Pixeltime GetLong amp Y GetLong amp scantype GetLong amp k transferdatawaveF GetWave transferdatawaveB Get Wave Scan Xi Yi Scanrange Pixels Pixeltime Y scantype k transferdatawaveF transferdatawaveB break CMD is the only message I care about main ioRecHandle This is the initial entry point at which the host application calls XOP The message sent by the host must be INIT main does any necessary initialization and then sets the XOPEntry field of the ioRecHandle to the address to be called for future messages HOST_IMPORT void main IORecHandle ioRecHandle ifdef XOP_GLOBALS_ARE_A4_BASED ifdef _MWERKS_ For CodeWarrior 68K
86. RPC pragma options align reset endif ifdef WINDOWS pragma pack endif All types of scanning are controlled by the XOP Sac2 c The parameter scantype defines which kind of scan will be performed scantype 0 for image scan scantype 1 for horizontal line scan and scantype 2 for a vertical line scan Sac2 c include XOPStandardHeaders h Include ANSI headers Mac headers IgorXOP h XOP h and XOPSup port h include Sac2 h include C ADwin Developer C Microsoft_VisualC Adwin c include C ADwin Developer C Microsoft VisualC Adwin h All structures are 2 byte aligned if GENERATINGPOWERPC pragma options align mac68k endif ifdef WINDOWS_ pragma pack 2 endif static void DataTransfer waveHndl transferdatawaveF waveHndl transferdatawaveB long Pixels long k int i long dataXf 10000 long dataYf 10000 long dataCtsf 10000 long datat1f 10000 long datat2f 10000 long dataXb 10000 long data Yb 10000 long dataCtsb 10000 long datat1b 10000 long datat2b 10000 long indices 3 dimensions of the wave double value 1 dimensions of the value GetlData 1 dataXf 1 Pixels transfers points 1 to Pixels of ADW data_1 to dataXf Get Data 2 dataYf 1 Pixels GetlData 3 dataCtsf 1 Pixels GetlData 4 datat1f 1 Pixels GetlData 5 datat2f 1 Pixels i 0 184 Set up control and data acquisition software while i lt Pixels indices 0 i row number
87. Recall that the TCSPC module is started before the confocal image therefore the first part of the TCSPC data does not correspond to the confocal data 36 The fluorescence and light scattering confocal microscope pixel of the confocal image in order to find the pixel to which the photon corre sponds At the end an image is constructed in which for every photon of each pixel the TCSPC mic t time is known i e a fluorescence lifetime image Then as shown in figures 2 19 b and 2 19 c from the fluorescence lifetime image the fluorescence excited state lifetime of different molecules or regions of the sample can be studied Fluorescence vs time measurements Fluorescence vs time measurements also called kinetic traces consist of record ing the fluorescence emission of a certain molecule or region of the sample as a function of time The procedure to carry out a fluorescence vs time measurement is as follows First an image of the region of interest of the sample is acquired and the TCSPC must be ready to start in FIFO mode see 19 for details Next the round A cursor on any of the confocal images is positioned on the point of the sample that is going to be studied Then pressing the Start button in the Kinetic division of the control panel moves the sample to the position marked by the cursor A opens the shutter of the excitation light starts recording the photons detected by the AD DA counter as a function of time and displ
88. TA_5 j time2 END FINISH flag 1 LP HP LineScan BAS Proze nummer 1 Delay 1 Eventsource 0 Number of Loops 0 Prioritat 1 Version 1 FastStop 0 AdbasicVersion 2000000 ATSRAM 0 OPT_LEVEL 1 SAVECOMPIL 0 DIM test Tpixels j AS INTEGER DIM DATA_1 10000J AS LONG X DIM DATA_2 10000J AS LONG Y DIM DATA_6 10000 AS LONG Monitor X DEFINE Xi PAR_1 DEFINE Yi PAR_2 DEFINE Scanrange PAR_3 A 1 AD Basic routines 155 DEFINE Pixels PAR_4 DEFINE Pixeltime PAR_5 DEFINE Pixelsize PAR_6 DEFINE X PAR_7 DEFINE Y PAR_8 DEFINE j PAR_9 DEFINE terminated PAR_11 DEFINE flag PAR_55 INCLUDE C ADwin ADbasic3 Inc adwgent inc INIT terminated 0 globaldelay 1 test 0 IF test 1 THEN Xi 32768 Yi 32768 Scanrange 32768 Pixels 5 Pixeltime 200 Y 45000 ENDIF EVENT PAR_11 0 Pixelsize Scanrange Pixels Pixeltime Pixeltime 40 usec X Xi Pixelsize 2 Y PAR_8 flag 1 j 0 CNT_ENABLE 1 CNT_CLEAR 1 DAC 2 Y DO j j 1 X X Pixelsize DATA_1 j X DATA_2 jJ Y START_PROCESS 2 DO X PAR_7 UNTIL flag 1 DATA_6 jJ ADC 1 UNTIL j Pixels DO j j START_PROCESS 2 156 Set up control and data acquisition software DO X PAR_7 UNTIL fag 1 DATA_1 j X DATA_2 jJ Y X X Pixelsize DATA_6 jJ ADC 1 UNTIL j 2 Pixels X X Pixelsize END FINISH PAR_11 1 LP HP LineScan_Calibration BAS Proze nummer 7 Delay 1 Eventsource 0 Number of Loops 0 Priori
89. TEoER a 42 Single molecule fluorescence through a layered system because the normalized rates do not depend on the actual value of the transition dipole moment w is the frequency and kr is the wavevector in the reference medium of refractive index np of the emitted radiation o and er are the dielectric constants of vacuum and the reference medium respectively First the radiative de excitation rates to different regions of the space are cal culated via the application of the Lorentz reciprocity theorem Then following the back reacted field approach the total electromagnetic de excitation rate is calculated 3 2 1 Radiative decay rate to different regions of space By application of the Lorentz reciprocity theorem 31 32 the radiation emitted by an oscillating dipole in a certain direction can be calculated as the local electric field at the dipole position generated by a plane wave incident from that direction Therefore the calculation of the radiative decay rate to a given region of space is analogous to the determination of the electric field intensity generated at the dipole position by plane waves incident from all possible directions in the region of interest The electric field produced by a plane wave in any point of a layered system can be calculated by solving Maxwell s equations via a transfer matrix algorithm TMA 33 In the TMA the layers see figure 3 2 are defined by their thicknesses d and complex diel
90. Window B PointKinetic DoWindow F SurfaceScanF DoWindow F SurfaceScanB DoWindow F LineScan Xi Xi PixelsToCorrectInX Pixelsize Correction using previous calibration parameters Scanrange Scanrange PixelsToDiscard Pixelsize Correction using prev calibration parameters Pixels round Scanrange Pixelsize Make O I U N 2 Pixels 2 Pixels 5 transferdatawaveL transferdatawaveL 0 Make O I U N Pixels LineCtsDisplay LineCtsDisplay 0 fit_LineCtsDisplay 0 SetScale x Xi_um Xf_um um LineCtsDisplay SetScale x Xi_um Xf_um um fit_LineCtsDisplay DoUpdate ActualY Yi Scantype 1 Init OpenShutter Do ExecuteSAC2ForLine DataForLineDisplay graphname 178 Set up control and data acquisition software While 1 End LineScanV scans repetitively a line vertically on the sample until the user aborts The data from every scanned line is displayed on the on line display Function LineScanV StartLineV ButtonControl String StartLineV SetDataFolder root String graphname message Variable aux finish paramctrl Wave fit_LineCtsDisplay graphname winname 0 1 If cmpstr graphname SurfaceScanF 0 If cmpstr graphname SurfaceScanB 0 message Parameters are taken from the top graph At the moment the top graph is graph name Be sure that one of the surface scan images is the top graph Abort message Endif Endif Xi_um xcsr A Yi_um vesr A Scanrange_um vesr B vcsr A Pixels Pixe
91. XOPs SetCurrentA4 Set up correct A4 This allows globals to work SendXOPA4Tolgor ioRecHandle GetA4 And communicate it to Igor endif endif XOPInit ioRecHandle do standard XOP initialization SetXOPEntry XOPEntry set entry point for future calls SetXOPResult OL All structures are 2 byte aligned if GENERATINGPOWERPC pragma options align reset endif ifdef WINDOWS pragma pack endif List of Tables 2l 2 2 2 3 4 1 4 2 5 1 5 2 5 3 5 4 Lieh SOUrcES oru ce ee oe e eee ee eee Sy Se a 10 Single photon counting detectors 2 000004 ee 18 Spectrograph components 2 0 00 eee es 19 Properties of the sample layers o 4 eee we DE a See ew di 70 Influence of the gold film on the decay rates 75 Trace histogram and autocorrelation analysis of TCSPC data 96 Comparison of the trace histogram and autocorrelation methods 97 Total EM decay rate in the samples with and without gold 102 Detectable decay rate in the samples with and without gold 105 List of Figures 2 1 2 2 2 3 2 4 2 5 2 6 2 7 2 8 2 9 2 10 2 11 Zeke 2 13 2 14 2 15 2 16 2 17 2 18 2 19 2 20 3 1 3 2 3 3 3 4 3 5 4 1 4 2 4 3 4 4 4 5 4 6 Contocal principles t sreang pied 2 Bee teilt 6 Confocal image formation nr ee ae ee Bh eee 7 Schematic of the home built confocal microscope 9 Focusing angles in the microscope objective
92. Y Yi ELSE DO i 0 DEC Y i 0 DO DAC 2 Y i i 1 UNTIL i 150 UNTIL Y Yi ENDIF ELSE i 0 IF Yi Y lt 1900 THEN DO i 0 INC Y DO DAC 2 Y isi 1 UNTIL i 2000 UNTIL Y Yi ELSE i 0 DO INC Y DO DAC 2 Y isi 1 UNTIL i 150 i 0 UNTIL Y Yi ENDIF ENDIF ENDIF PAR_21 0 PAR_22 0 PAR_23 0 i l DO DATA_1 iJ 0 DATA_2 i 0 DATA_3 i 0 A 1 AD Basic routines 153 DATA_4 i 0 DATA gt 5 i 0 DATA_11 i 0 DATA_12 i 0 DATA_13 i 0 DATA_14 i 0 DATA_15 i 0 DATA_31 i 0 DATA_32 i 0 INC i UNTIL i 10000 i 1 DO DATA_21 i 0 DATA_22 i 0 DATA_23 i 0 INC i UNTIL i 100000 check 1 END HP Pixel BAS Proze nummer 2 Delay 1 Eventsource 0 Number of Loops 0 Priorit t 0 Version 1 FastStop 0 AdbasicVersion 2000000 ATSRAM 0 OPT_LEVEL 1 SAVECOMPIL 0 DIM delay timel time2 time Pixeltime AS long DIM X j Counts AS INTEGER DIM DATA_3 10000 AS LONG Counts DIM DATA_4 10000 AS LONG time0 DIM DATA_5 10000 AS LONG time3 DEFINE Pixeltime PAR_5 DEFINE flag PAR_55 INCLUDE C ADwin ADbasic3 Inc adwgent inc INIT Globaldelay 1 154 Set up control and data acquisition software EVENT Pixeltime PAR_5 X PAR_7 j PAR_9 flag 0 timel READ_TIMER IF gt 1 THEN timel DATA_5 j 1 ENDIF DO DAC 1 X time READ_TIMER delay time time1 UNTIL delay gt Pixeltime time2 READ_TIMER Counts CNT READ 1 CNT_CLEAR 1 DATA_3 j Counts DATA_4 j timel DA
93. a sample with and without the 44nm gold film To 0 28 1 08 2 67 1 87 perform before undergoing irreversible photo bleaching equation 4 2 Then the number of detectable photons np emitted by a single molecule which transition dipole has an out of plane orientation defined by is lee Nn Ne 9 x Ton 4 4 total From the values displayed in table 4 2 it can be seen that the present detection scheme through the gold film is less effective for molecules with its transition dipole parallel to the interface but it is more effective in the case of molecules with its transition dipole perpendicular to the interface The number of detectable photons emitted by a parallel molecule in the sample with the gold film is 26 of the photons that the same molecule would emit in the sample without the gold film For the case of a perpendicular molecule this percentage rises to 140 4 4 3 Excitation field at the chromophores position Following the method presented in section 3 3 the electric field distribution at the molecules position was calculated for the different modes of illumination The calculation parameters were set in order to represent the experimental conditions A linearly polarized A 633 nm plane wave front was considered to be focused at the spacer air interface of the samples with a gold film thickness of 44nm and a spacer thickness of 24nm The dielectric constants of the materials for a wavelength of 633nm are li
94. al range of interest and to place the CCD camera in the right position The photo objective and the CCD camera can be moved and tilted in all direc tions To maximize the detectable spectral range the photo objective needs to be as close as possible to the grating and its diaphragm has to be completely open Then the position of the camera and the focus of the photo objective need to be adjusted in order to produce a sharp image of the spectra in the desired position of the CCD sensor For this it is convenient to use white light illumination and set the camera on live mode 24 because this allows to monitor on line how light of different wavelengths is focused on the CCD sensor The light dispersed by the VPH grating should be focused by the photo objective to a straight line on the CCD sensor It is advisable to accommodate the camera position in such a way that the light is focused on a horizontal pixel line of the CCD sensor because in that case the data analysis is simpler Alignment of the detection for light scattering measurements In this case annular illumination is used and the diameter of the detected beam should be reduced with a diaphragm in order to block the reflected light see fig ure 2 8 Because of the annular illumination the reflected light beam used for the 2 4 Operation 29 alignment looks like an annulus The diaphragm should be placed centered to this annulus and closed to a point in which no light can pass th
95. an be effectively acquired All the observed scatterers present 138 Light scattering from single metallic nano structures a diffraction limited signal with enhanced Airy discs because of the annular illumi nation The signals present different intensities spanning around an order of magnitude Figure 7 2 right also shows representative maximum and minimum signals ob tained from the vertical white lines in the main image the ratio of the maximum to minimum signal is 1940 120 280 120 12 9 The scattering cross section of the particles is proportional to the sixth power of their diameter 132 From the size distribution of the particles figure 7 1 the minimum particle size is approximately 17 5nm and the maximum 23 5nm This size range can account only for a ratio of maximum to minimum signal of 5 8 The signals with intensities beyond this range belong probably to small conglomerates of particles or more than one particle lying close to each other within the diffraction limited focal spot um f Counts 500 4 per pixel IN Nr 0 5 10 15 20 00 05 10 15 um um Figure 7 2 Light scattering of individual Au nanoparticles In the center a scattered light image of single gold nanoparticles On the left a detail of the main image corresponding to the dash line square On the right profiles of the strongest bottom and the weakest top signals of the main image marked by the white vertical lines Figur
96. ange 80 32768 Xi_um Xi 32768 80 32768 result 1 If Xi um lt 0 DoAlert 0 X is out of range after calibration result 0 Endif If Scanrange_um gt 80 DoAlert 0 Scan range is too big after calibration result 0 Endif Return result End Function RestoreDisplay Values RestoreDisplay Values ButtonControl String RestoreDisplay Values Xi_um Xi_um_display Yi_um Yi_um_display Scanrange_um Scanrange_um_display Pixels Pixels_display Xi Xi_um 32768 80 32768 Piezo table in closed loop moves 80um Yi Yi_um 32768 80 32768 for the 0 10V 32768 ADWin units range Scanrange Scanrange_um 32768 80 A 2 Igor routines 169 Xf Xi Scanrange Yf Yi Scanrange Xf_um Xi_um Scanrange_um Yf_um Yi_um Scanrange_um End Once the corrected parameters are calculated a new set of 3 scans is performed by ScanFinalCalib with the corrected parameters Then through FitFinalLinear Range it is verified that with the new parameters an image with the desired number of pixels and position can be constructed within the linear range of the piezo stage trajectory Function ScanFinalCalib Variable paramctrl j Offset Wave Xdisplay t2display Xmondisplay Corr_Xmondisplay DoWindow B PointKinetic DoWindow F Calibration ControlInfo W ControlPanel KeepPrevFit If V_value 1 Duplicate O Xdisplay PrevXdisplay Duplicate O t2display Prevt2display Endif Redimension N 2 Pixels Xdisplay Redimension N 2 Pixels Xmondisplay Redimension N 2 Pixels
97. ate surface plasmon resonances in metallic nano particles and the importance of studying them on a single particle basis In conclusion experimental and theoretical contributions were made for the quantitative understanding of the influence of locally enhanced electromagnetic fields on single molecule fluorescence Appendix A Set up control and data acquisition software In this appendix the programmed code for the control of the home built SCOM and the data acquisition is presented The routines are commented and some im portant characteristics of the programming languages are explained All the computer controlled functions of the home built SCOM are driven by signals provided by the Analog Digital Digita Analog AD DA PC Card Opera tions to be performed by the AD DA card can be controlled by routines loaded in its local CPU These routines are programmed in AD Basic and can be found in the section A 1 The user interface was programmed with Igor in order to use Igor s built in capabilities for data treatment The code of the user interface functions is presented in section A 2 Since the AD Basic language has no drivers for Igor intermediate routines in C were needed to complete the communication flow between the user and the local CPU of the AD DA card The corresponding C routines can be found in the section A 3 A 1 AD Basic routines ScanInit BAS Proze nummer 4 Delay 1000 Eventsource 0 Number of Loops 0
98. ation power The weak tendency to higher values as the power 120 Photoluminescence blinking of Zno 42Cdo ss5e nano crystals increases is an artefact of the trace histogram analysis method All the simulated points shown in figure 6 8 and the error bars were obtained from Monte Carlo simulated data see next section with exponents between of 1 72 and 1 74 Detected cycles per second Another relevant parameter for the characterization of the blinking process is the frequency with which a QD jumps from the on to the off state and viceversa The number of on off cycles per second was computed and is shown in figure 6 9 a for the QDs on glass and ITO coated glass substrates as a function of the excitation intensity a b 3 604 Glass x 6 Glass E 50 0 ITO N gt ITO 3 Simulated 54 Simulated YA ei ae 8 44 d E 5 31 gr E Dt 5 pap BY pe 10 4 9 1 4 04 O o r 4 T T I T T T 0 5 1 0 1 5 2 0 0 0 0 5 1 0 15 2 0 P kWicm P kWicm Figure 6 9 a Experimental and simulated on off cycles per second for the QDs on glass and ITO coated glass substrates as a function of the excitation intensity P b Experimental cycles per second normalized by the signal to background ratio SBR It can be observed that the number of cycles per second follows a trend similar to the net on intensity Jon ner This is due to the fact that as the signal to background ratio SBR increases shorter on period
99. ative decay rates to the complete space According to figure 3 2 Lir 8 Tiota lP Tol Tn 9 3 13 where Io and Iy are the radiative decay rates to the semi spaces above and below the thin layered system 3 2 4 Detectable fraction of the de excitation rate The detectable fraction of the de excitation rate is the fraction corresponding to radiation emitted into the collection solid angle of the objective For an ideal fluorophore with a rf 0 and a transition dipole with an out of plane orientation according to this fraction writes o Tod _ sin Q Toz cos Topj Taal ee nl FT ee Poz is the radiative de excitation rate emitted into the collection solid angle of the microscope objective calculated by the integrals 3 6 between the appropriate limits Tg is the total de excitation rate calculated by equations 3 11 Both Ton and I are normalized to the total emission of a dipole in the same reference medium Modifications for the case of a fluorophore with T 4 0 The intrinsic quantum efficiency QE of a fluorophore in a given medium is defined as a function of its intrinsic radiative and non radiative de excitation rates in that medium T QE r Tr 3 15 So far an ideal fluorophore with T 0 or equivalently with QE 1 was considered If a fluorophore has T 0 its quantum efficiency is smaller than 1 and the amount of radiation emitted in a time unit is reduced Taet is comp
100. ays the kinetic trace on the top right graph Figure 2 20 shows a screen shot of the PC user interface during a fluorescence vs time measurement of the feature marked by the cursor A on the bottom left backward image The stage needs around 10 seconds to move to the selected position and reach stability This time can be used to start the TCSPC module in FIFO mode to record the kinetic trace If this time is not sufficient the TCSPC module can also be started earlier Like this the fluorescence vs time trace is recorded indepen dently by the AD DA counter and the TCSPC unit However these two data sets are different The data collected by the AD DA counter and displayed on the PC user interface is intrinsically binned i e it is not the detection time of each photon what is recorded but the number of photons detected within a time interval of a size determined by the Pizel time parameter in the Kinetic division of the control panel This is not convenient because the interesting time scale of the processes in volved in the fluorescence fluctuations need to be known a priori in order to choose the appropriate bin width In addition the data presents gaps due to the fact that the AD DA card cannot store and deliver data at the same time So the photons detected during the time necessary to transfer the data to the computer memory are not recorded The Time parameter in the Kinetic division of the control panel determines how often this data trans
101. beam is determined by the x y position of the fiber tip Modifying the direction of the fiber theta and phi changes the intensity profile of the beam because the illumi nation over the surface of the collimating lens changes The aim is to illuminate the collimating lens as uniformly as possible The x y position and the direction 6 amp have to be adjusted iteratively until the beam is directed along the optical axis with a uniform intensity distribution At the end of this procedure the collimation should be checked again with the retarded interference plate and if necessary the distance d should be corrected 2 3 Alignment 25 Alignment for annular illumination Annular illumination requires the positioning of a blocking disc in the center of the illumination beam There are two ways to do this The blocking disc can be glued onto a glass slide or it can be held with very thin cords The disc is mounted in a positioning stage that allows it to be moved it in three dimensions The position of the disc can be controlled by closing the diaphragms that define the optical axis If the disc is in the center of the illumination beam the diaphragms should close concentrically with the disc 2 3 3 Alignment of the dichroic mirror and the microscope objective In fluorescence measurements the illumination beam is directed to the micro scope objective by a dichroic mirror that also separates the illumination from flu orescence ligh
102. bjective in the samples with and without gold The average avg value is calculated as 1 3xTL 2 3xT 5 4 3 Influence on kon In the case of kon the gold film produces a noticeable effect The distribution of kon for the case with gold is more than two times broader and presents an average value two times higher than the case without gold It should be noted that the experiments with gold have in average a lower signal to background ratio This can produce a slight broadening of the distribution due to a higher uncertainty in the autocorrelation analysis However a statistical analysis with simulated data shows that this is not enough to account for the experimental broadness Another effect of the lower signal to background ratio is to make the detection of short on periods more difficult This leads to the fact that the measured average length of the off periods is higher than the real one leading to a smaller values of kon The experiments show however higher values of kon so that the real effect could be stronger than observed 4x10 No Au R 0 05 kon 1 s kon 1 s 0 35 0 40 0 45 0 50 0 55 04 06 08 10 T4 1s I54 1 8 Figure 5 14 kon as a function of 2 for the chromophores in the sample with and without gold The corresponding linear Pearson correlation coefficient is shown in the graphs Further evidence of the influence of the nearby gold film on kon is found in the correlation between kon and
103. case and finds reasonable thresholds The thresholds found in the case of the QDs do not always separate the two states but they do exclude the off state The histograms of the length of the on and off periods on and off times figure 6 3 right show the power law probability density that was already observed in QDs of different compositions 109 111 It is remarkable that similar power law proba bility densities are found for on and off period lengths regardless the distribution of the on intensity Even the on and off periods of the first trace in figure 6 3 which shows an on intensity only slightly broader than a Poisson distribution have clear power law probability densities The deviation of the on intensity from a Poisson distribution has been attributed to very short on periods very probable given the power law probability density which are partly detected because of the finite time resolution of the experimental methods 110 The Monte Carlo simulations presented later in this chapter will show that this is in fact not true 6 3 2 Effect of the excitation intensity Kinetic traces of QDs were recorded at different excitation intensities from QDs on both glass and ITO coated glass substrates In figure 6 4 the on and off periods recorded from many 12 for each intensity QDs on glass substrates are computed together in a common histogram for each excitation intensity P For every excitation power the same time T shown in eac
104. ccurs at a constant rate from the excited state 47 Therefore an enhancement of the total de excitation rate is accompanied by an increase in the average number of cycles nc that a molecule can perform between the ground and excited states before it photo bleaches Then it is possible to write nex ins E 4 2 For a given set of excitation and de excitation rates a higher number of excitation de excitation cycles nc yields a higher number of photons emitted by a single molecule Spacer air interface spacer or air side The fluorescent dye molecules are placed by the electrostatic deposition method on the spacer air interface of the samples but it remains a priori uncertain on which side of the interface As it will be shown in the next two sections because of the electromagnetic boundary conditions for the electric field both the excitation and emission rates are considerably different for molecules on either side of the interface Equations 3 3 write for the present case El aap E spacer 4 3 Eair El air Espacer Piisspader In the modelling both possibilities are considered and all calculations are per formed for the cases of molecules on the air and on the spacer side of the interface Later by comparing the theoretical results to the experiments it will be possible to conclude how the molecules optically behave as being on the air or on the spacer side 4 4 Modelling the experimental scheme 71 4 4 2 Det
105. chromatic 0 lenses Mirror Dichroic Mirror Diaphragm Figure 2 3 Schematic of the home built confocal microscope types of fibers were used standard single mode fibers for 635 or 515 nm Thor Labs Inc or photonic crystal fibers Endless Single Mode Blaze Photonics Ltd the latter has the characteristic of acting as a single mode fiber in a wide range of wavelengths 12 13 The pure Gaussian mode TEM00 light coming out of the fiber is collimated with a 150mm focal length achromatic lens OWIS GmbH to form the illumination beam The illumination beam is aligned with the main optical axis of the microscope for alignment details please refer to section 2 3 and its diameter is adjusted with a diaphragm If the experiment requires annular illumination a blocking disc can be introduced in the axis of the illumination beam The diaphragm and the blocking disc should be placed as close as possible to the microscope objective in order to minimize the optical path to the focus and therefore the diffraction effects from the borders of the pin hole or the disc Furthermore it is advisable to minimize the number of optical elements between the fiber tip and the microscope objective because small distortions of the illumination beam can produce important effects in the focus A 100x 1 4 NA oil immersed microscope objective Plan Apo Nikon GmbH is 10 The fluorescence and light scattering confocal microscope Pulses Wavelength Manu
106. component of the electric field would be considerably weaker than the in plane component along the polarization direction 4 3 2 Different illumination modes In order to discriminate the contributions of transmitted light and forbidden light for the excitation of the chromophores the samples with 4 bilayer spacers were studied with the three different modes of illumination described in section 4 2 In figure 4 5 the images on the left column were recorded with full beam FB illumination the ones in the center with transmitted light TL and the ones on the right with forbidden light FL illumination The images obtained with FL illumination are very similar to the ones obtained with FB illumination described above in section 4 3 1 In general the background in FL illumination is reduced approximately to one half due to the considerable smaller amount of reflected light see figure 1 2 The images obtained with TL illumination are strikingly different The characteristic patterns are no longer observed Instead very weak round signals are detected with a size larger than the diffraction limit 500 nm By comparing the images obtained with the different illumination modes it can be seen that these round signals obtained with TL illumination do not always overlap with the characteristic patterns obtained with the FB or FL illumination modes The figures in the lower row are intended to further clarify this effect The scanned region is s
107. consequence that the probability of detecting a photon at a given time is independent of when the last photon was detected i e the process of de tecting photons is memoryless which can be seen by verifying that for 0 lt s lt t Px gt t x gt s Pix gt t s Pix gt s Pex gt s The time elapsed between two consecutive detected photons inter photon times is a probe of Pix gt n As there are two populations the on and the off with two different average intensities Jon and Ioff a double exponential equation 5 16 curve is found in a histogram of the inter photon times Post Aone t Aygpe tert 5 19 Experimental data 3 Double exponential fit Figure 5 5 Inter photon times his togram Histogram of inter photon times of the data shown in figure 5 2 For the double exponential fit one of the exponents is fixed to the value of the experimental average background intensity Ios 0 0 0 5 1 0 1 5 2 0 2 5 Inter photon time ms Figure 5 5 shows the inter photon times histogram of the data shown in figure 5 2 A double exponential fit allows to obtain the experimental values for Ion and Ioff from the exponents In fact there is no need to extract ofp from the fit it is possible to obtain it directly from the data as the average background intensity and use this value as a constraint in the fit to obtain Ion But there is more information to be extracted from the two pre factors Aon and Aoff The normalized expon
108. control and data acquisition software XminusCorr_X_F Sum Xdisplay XiF XfF XfF XiF Sum Corr_Xmondisplay XiF XfF XfF XiF XminusCorr X_B Sum Xdisplay XiB XfB XfB XiB Sum Corr_Xmondisplay XiB XfB XfB XiB End Confocal Imaging All the routines employed for imaging are grouped in the Confocal Imaging Proce dure The function Check Values checks that the input values from the ControlPanel are proper i e the initial position scanning range the scanning speed and number of pixels are whitin the possible limits pragma rtGlobals 0 Function Check Values Variable result Result 1 If Xi_um lt 0 Xi_um gt 80 DoAlert 0 Xi must be between 0 and 80 um Result 0 Endif If Yi_um lt 0 Yi um gt 80 DoAlert 0 Yi must be between 0 and 80 um Result 0 Endif If Xi um Scanrange_um gt 80 Yi um Scanrange_um gt 80 Scanrange um lt 0 DoAlert 0 The Scan Range is not valid Result 0 Endif If Pixels lt 0 Pixels gt 4096 DoAlert 0 The number of Pixels is not valid or makes no sense Result 0 Endif ControlInfo W ControlPanel Pixeltime Units If V_value 1 Pixeltime Pixeltime_display 1000 Else Pixeltime Pixeltime_display Endif If Pixeltime lt 300 DoAlert 0 The Time per Pixel is not valid Result 0 Endif Return result End A 2 Igor routines 173 Once controlled that the set of scanning parameters is within the possible limits CalculateParametersForAdw
109. crease in resolution see figure 2 2 Even though the improvement in resolution may at first sight seem to contravene the basic limits of optics it can be explained by a principle described by Lukosz 10 11 which states that resolution may be improved at the expense of field of view In the case of a confocal microscope the field of view is reduced by means of the back projected image of a point detector in conjunction with the focused point light source Nevertheless the field of view can be increased by scan ning There are basically two different ways of scanning which have been achieved by various methods in practical instruments The alternatives are either to scan a focused light beam across a stationary sample or to scan the sample mechanically across a stationary focused light spot In the first case scanning can be very fast and many images per second can be acquired In the second case scanning is much slower but as the optical path remains stationary undistorted images of very high quality are produced 8 The fluorescence and light scattering confocal microscope 2 2 Description of the home built confocal micro scope Figure 2 3 shows schematically the different components of the home built SCOM set up and helps as a guide for the detailed descriptions presented in the next sections of this chapter A number of light sources can be employed to obtain either monochromatic or white light illumination In every case the light
110. ctra of Zno 42Cdg 5g5e QDs excited at 514 5 nm 6 2 Experimental 113 400 300 Figure 6 2 Typical photolumines 200 cence image of the samples with indi vidual QDs excited at 514 5nm Blink ing is observed during the scanning for the image acquisition um 100 0 Counts per pixel Zno 42Cdo 585e QDs were diluted in toluene gt 99 7 Riedel de Ha n and spin casted onto glass N 1 Menzel Gl ser or Indium Tin Oxide ITO coated glass sub strates 50nm Fraunhofer Institut IST Braunschweig The concentration was ad justed in order to obtain a surface density of approximately one QD per um as shown in the micrograph of figure 6 2 6 2 2 Measurement The measurements were performed with the SCOM set up described in chapter 2 and consisted of imaging a region of the samples identifying a single QD and recording the photoluminescence emission of that QD as a function of time The avalanche photo diode APD was used as single photon detector and the Ar ion laser was used for circularly polarized excitation at 514 5nm The repetition rate of the Ar ion laser 60 MHz was too high to allow measurements of the relatively long photoluminescence excited state lifetime of the QDs Therefore continuous wave illumination was used Suitable dichroic and notch filters were used to separate the QDs emission from the excitation light The photoluminescence emission of around 350 individual Zno 42Cdo 535e QDs was rec
111. d because the actual value of QE is uncertain but expected to be close to unity Nevertheless it is possible as well to calculate the theoretical fluorescence signal of a fluorophore with non zero T QE lt 1 The effects on the detectable fraction of emitted flu orescence produced by quantum efficiencies smaller than one discussed in section 4 4 2 reflect directly in the theoretical fluorescence signals As the QE reduces the relative intensity of molecules with great out of plane components of their transition dipoles increases in comparison to the signals of in plane molecules 4 5 Conclusions The first conclusion to be drawn from the experimental observations shown in figures 4 3 and 4 6 is that single molecule fluorescence in the nanometric vicinity of a thin gold film can be excited and detected with a confocal epi illumination scheme through a thin gold film down to a separation distance of 15nm from the gold surface Second from the comparison of the images obtained with the different illumi nation modes and the theoretical calculations it can be concluded that the surface plasmon resonance in the gold thin film plays a fundamental role Surface plasmons provide the most effective propagation path for the excitation and fluorescence light In excitation if the surface plasmon resonance is not excited almost no excitation field is generated at the chromophores position In detection if the NA of the objec tive is not large eno
112. d in the case of single molecule fluorescence measurements Under these conditions every time a photon is detected the time elapsed from the last excitation pulse is computed mic t in figure 2 9 a Then if the laser pulses are narrow enough in comparison to the typical fluorescence decay time the fluorescence decay curve is directly obtained gt The photo objective focuses the monochromatic beams dispersed by the grating into diffraction limited spots with a size ranging from 1 9 to 3 5 um depending on the wavelength The pixel size of the CCD sensor is 6 75 um 20 The fluorescence and light scattering confocal microscope a b Detector signal Period 1 H _____ gt Period 2 2 S 2 Period 3 Period N l e 0 T mic t 2 c gt Q Oo 0 T mic t Figure 2 9 Time correlated single photon counting principle a The curve represents the exci tation light pulses and the short vertical lines represent detected fluorescence photons For each detected photon the time from the beginning of the experiment Mac t and the time from the last excitation pulse mic t are determined b The fluorescence decay curve is obtained by computing a histogram of the mic t times by making a histogram of the detected mic t times as shown schematically in figure 2 9 0 The microscope is equipped with an independent TCSPC module SPCM 630 Becker und Hickl GmbH The device can measure the mic t times with a resoluti
113. d so that the intensities can be compared The field distribution of all components present the same symmetry as in the case of the sample with the gold film and the differences between FB FL and TL illumination 78 Single molecule fluorescence through a thin gold film are basically the same described for the case with the gold film The only change observed is that the maximum intensities of the x component lie in this case in a direction perpendicular to the polarization direction With respect to the relative intensities it can be seen that the gold film reduces the intensity of the three components but to different extents The intensity of the x component is the most drastically affected the gold film reduces it approxi mately one order of magnitude The y and the z components are less affected the gold film reduces their intensities by a factor of approximately 4 and 2 5 respectively 4 4 4 Theoretical fluorescence signal The theoretical fluorescence signal corresponding to the three modes of illumina tion was calculated for an ideal fluorophore on both sides of the spacer air interface and with different orientations of its transition dipole by means of equation 3 40 The procedure consists of scaling the calculated fields shown in the previous section by the fraction of the de excitation corresponding to detected fluorescence T ger spacer FB Max 4 t 7 Signal p gt 20 T H
114. d the time after the excitation pulse at which those photons were detected mic t in section 2 2 4 Once an interesting region of the sample has been identified the TCSPC module should be started first in mode FIFO Then the confocal image scanning should begin so that in both the TCSPC and confocal data the complete information is stored In this manner every photon seen by the detector is doubly detected by the AD DA counter and the TCSPC and two sets of data are obtained The data from the confocal image recorded in Igor as described above contains for every pixel besides the position the time at which the counter was started and the time at which it was stopped with 25ns resolution The data obtained with the TCSPC consists of two values for every detected photon recorded in chronological order 15 All this information is saved together with the data in the Igor binary file as the note associated to the saved wave and one can get access to it with the Igor command note wavename 16The file is called ConfocalAnalyis pxp and is attached in a CD ROM to the copy of this dissertation at the MPI P library 2 4 Operation 35 Time ms 30 40 Confocal data TSCPS data c 100 4 3 2 1 5 1 10 53 1 ts 012345678 9 10 39 Time ns 38 E gt 37 gt 2 bor 38 35 0123 45 67 8 g 10 pi a ie ae L Time ns Figure 2 19 Fluorescence lifetime imaging a To produce a fl
115. d through the gold film with an epi illumination scanning confocal microscope The influence of the separation distance to the gold film is experimentally studied as well as different illumination modes to discriminate the contributions of transmitted and forbidden light for excitation The theoretical model presented in chapter 3 is applied for the calculation of theoretical single molecule fluorescence signals in the experimental conditions Theoretical and experimental signals are compared in order to interpret the observations 4 1 Introduction A number of processes of great technological and scientific importance involve a fluorophore working near a metallic layer Among the most important certainly are sensitized solar cells organic light emitting diodes and a variety of sensors The problem of spontaneous emission near a conducting surface has been al ready studied both experimentally and theoretically for several decades The first theoretical approach to the problem was reported by Sommerfeld in 1909 26 and after the pioneering experiments by Drexhage et al 41 43 a considerable number of experimental and theoretical investigations were dedicated to study the influence of a nearby metallic layer on the fluorescence emission 27 29 44 47 More recently room temperature optical spectroscopy of single fluorescent molecules By forbidden light it is meant light incident at angles higher than the angle of total internal r
116. d to the collection solid angle of the objective I it can be seen that in all cases practically the complete emission to the glass side is collected by the objective For the sample with gold film 99 of the radiation emitted by a perpendicular dipole to the glass side is collected For the parallel dipole 96 This collection efficiencies are slightly higher than the ones corresponding to the case of a sample without the gold film 93 and 91 for the perpendicular and parallel dipoles respectively The explanation for this stronger orientation of the emission is that analogously to the case of the excitation light fluorescence light emitted by the fluorophores propagates through the gold film via surface plasmon back coupling In order to compare the detection efficiency of the two cases it is necessary to compute the number of detectable photons emitted by a single molecule with and without the gold film Two factors need to be considered First the fraction of the de excitation rates corresponding to radiation collected by the objective equation 3 17 Second the number of excitation de excitation cycles that the molecule can 4 4 Modelling the experimental scheme 75 With Au Without Au L With Au Without Au Table 4 2 Influence of the gold film on the Kar 1 05 1 34 5 75 221 decay rates De excitation rates of a molecule e 0 29 1 18 272 2 00 on the air side of the spacer 24 nm air inter glass gt a face of
117. dence that the QD photoluminescence blinking does not occur between an off and an on state but between an off state and a distribution of on states They showed that the different emission intensities correspond to different excited state lifetimes To explain the observations they proposed the presence of fluctuating non radiative pathways Those results were confirmed by Fisher et al 114 Second Hohng and Ha reported an almost complete suppression of the photoluminescence blinking of QDs in the presence of 3 mercaptoethanol 115 However despite all the experimental and theoretical efforts dedicated to understand the QDs emission process the mechanism of the blinking remains uncertain and none of the proposed models provides a complete physical picture of the process In the experiments presented here the photoluminescence blinking of a new kind of QDs is studied monocrystalline alloyed Zng 1 Cdo 5sSe QDs 116 Naturally the first question addressed in this chapter is therefore whether the blinking of this new 112 Photoluminescence blinking of Zno 42Cdo53 Se nano crystals kind of QDs also present the universally observed power law Next if the ioniza tion model is correct a different blinking behavior is expected for QDs deposited on isolating and on semiconducting substrates In order to investigate this the lumi nescence blinking of Zng 42Cdp 535e QDs is studied on glass and Indium Tin Oxide ITO coated glass substrates as a
118. dinates and L parallel and perpendicular average Bibliography 1 10 11 12 13 14 M Fleischmann P J Hendra and A J McQuillan Raman spectra of pyri dine adsorbed at a silver electrode Chem Phys Lett vol 26 pp 163 166 1974 D L Jeanmaire and R van Duyne Surface raman spectroelectrochemistry Part I heterocyclic aromatic and aliphatic amines adsorbed on the anodized silver electrode J Electroanal Chem vol 84 pp 1 20 1977 M Albrecht and J Creighton Anomalously intense raman spectra of pyridine at a silver electrode J Am Chem Soc vol 99 pp 5215 5217 1977 M R Philpott Effect of surface plasmons on transitions in molecules J Chem Phys vol 62 pp 1812 1817 1975 S Nie and S R Emory Probing single molecules and single nanoparticles by surface enhanced raman scattering Science vol 275 pp 1102 1106 1997 K T Shimizu W K Woo B R Fisher H J Eisler and M G Bawendi Surface enhanced emission from single semiconductor nanocrystals Phys Rev Lett vol 89 p 117401 2002 M Minsky U s patent 3013467 microscopy apparatus vol 1961 pp Filed Nov 7 1957 Dec 19 M Minski Memoir on inventing the confocal scanning microscope Scan ning vol 10 pp 128 138 1988 C J R Sheppard and T Wilson Image formation in confocal scanning microscopes Optik vol 55 n
119. dspacer 0 nm b 02 0 0 0 10 20 30 40 50 60 er Transmitted light Forbidden light illumination mode illumination mode Full beam illumination mode Figure 4 2 Reflectivity of the sample system P polarization reflectivity coefficient of the samples with spacer thicknesses of 0 bare gold and 24nm as a function of the angle of incidence from the glass side The critical angle of total internal reflection 41 7 is marked as 6 The minima in reflectivity corresponds to the excitation of the surface plasmon resonance 58 angle of each illumination mode it is seen that the full beam FB and the forbidden light FL illumination modes can excite the SPR The TL illumination mode cannot 4 3 Single molecule fluorescence images through a thin gold film In this section it is demonstrated experimentally that it is indeed possible to excite and detect fluorescence of single molecules through the gold film and three aspects of the present configuration are investigated First the characteristics of the fluorescence images obtained Second images are acquired with different illu mination modes in order to find out what are the most effective pathways for the propagation of light imposed by the layered sample Third samples with different thicknesses of the polyelectrolyte spacer are studied in order to establish the min imum separation distance between the dye molecules and the gold film before the fluorescence becomes undetectable
120. e flag 1 END LP HP PointKinetic BAS Proze nummer 5 Delay 1 Eventsource 0 Number of Loops 0 Prioritat 1 Version 1 FastStop 0 AdbasicVersion 2000000 ATSRAM 0 OPT_LEVEL 1 SAVECOMPIL 0 DIM DATA_21 100000 AS LONG DIM DATA_22 100000 AS LONG DIM DATA_23 100000 AS LONG DIM j test AS INTEGER DIM time0 time3 time delay Kineticduration AS LONG DEFINE Xi PAR_1 DEFINE Yi PAR_2 DEFINE Pixeltime PAR_5 DEFINE j PAR_9 DEFINE Kineticduration PAR_15 DEFINE flag PAR_55 INCLUDE C ADwin ADbasic3 Inc adwgcent inc INIT globaldelay 1 flag 0 test 0 IF test 1 THEN PAR_1 32768 PAR_2 32768 PAR_5 100 PAR_15 2 ENDIF EVENT globaldelay 1 PAR_11 0 Xi PAR_1 Yi PAR_2 CNT_ENABLE 1 A 2 Igor routines 163 CNT_CLEAR 1 Pixeltime PAR_5 Pixeltime Pixeltime 40 usec PAR_5 Pixeltime timeO READ_TIMER Kineticduration Kineticduration 10000 DAC 2 Yi DAC 1 Xi timeO READ_TIMER j l DO START_PROCESS 6 DO flag PAR_55 UNTIL flag 1 time READ_TIMER delay time time0O UNTIL delay gt Kineticduration time3 READ_TIMER PAR_11 1 END FINISH A 2 Igor routines Igor routines are called Function Macro or Window and are grouped in Proce dures Igor can also call external routines programmed in other languages these external routines are called XOPs 133 and in this case were programmed in C see A 3 Any set of data i e an array of numbers is called Wave in Igor a Wave can be on
121. e two three or four dimensional In this section all the Procedures programmed to control the operation of the home build confocal microscope are presented Every Procedure and individual routine is explained and commented in Igor the symbol is used to add comments in the code Most of Igor commands have self explaining names and the code can be easily followed and in case of doubts there is a complete lexicon with all the programming commands in Igor s manual However it is worth to make a remark regarding the way Igor treats global variables If at the beginning of a Procedure reads pragma rtGlobals 0 then global variables and Waves do not need to be declared to be accessed in any function On the other hand if at the beginning of an Igor procedure states pragma rtGlobals 1 then all the global variables need to be declared in the corresponding functions 164 Set up control and data acquisition software with the commands Nvar and Svar Waves have to be declared as well with the command Wave Procedure ADWinlnitiation ADWinlnitiation Procedure has two functions The one via the ADWinBoot2 XOP boots and loads all the necessary routines to the local CPU of the AD DA PC card And the other via the ADWinInit2 XOP sends the stage to a initial position pragma rtGlobals 1 Function Boot Nvar Xi Yi Xi_prev Scanrange_prev Pixels_prev Xi 32768 0 Volt Yi 32768 0 Volt Xi_prev 0 Yi_prev 0 Scanra
122. e Wetzlar Germany GmbH was used Focusing the diode laser light The pulsed diode laser provides a multi mode non gaussian beam with very short collimation distance lt 1mm This divergent beam can be collimated with a standard 10x 0 2 NA microscope objective to an approximately 8 1 aspect ratio beam This beam is then focused by a 16x 0 3 NA objective on the tip of the single mode fiber The light coupling is very inefficient in this case Focusing the white light Light provided by the Xe arc lamp is highly divergent and cannot be efficiently coupled into the single mode fiber The best results were obtained by successively focusing the light with a 28mm 2 8 photo objective Nikon Japan and a 20x 0 4 NA microscope objective into the single mode fiber Although the photonic crystal fiber would be in principle ideal to obtain single mode white light spectral instabilities were observed The white light exiting the photonic crystal fiber showed spectral fluctuations in the sub ms range that prohib ited the use of this fiber to guide white light in any of the experiments 2 3 2 Collimation of the illumination beam The optical fiber acts as a point like light source providing a divergent beam which needs to be collimated in order to form the illumination beam The first step to collimate and align the illumination beam is to precisely define an optical axis by means of a reference laser beam and two aperture diaphragms 8On
123. e 7 3 is meant to show that the measured signals indeed correspond to scat tered light The sample area was scanned from bottom to top During the first lines the sample was not illuminated thus the image appears dark to a background inten sity level The next two dark areas were produced by placing a notch filter for the illumination wavelength A 514 5nm in the detection channel The notch filter renders the image dark to background level showing that the signals are composed of scattered light The effect of different sizes of the detection diaphragm aperture was also tested during the scanning for the image of figure 7 3 A large diaphragm aperture allows more reflected light to pass and be detected therefore the background intensity is higher Reducing the diaphragm aperture makes the overall signal less intense and the signal to background ratio higher There is an optimum size for the diaphragm 7 2 Light scattering of individual colloidal gold nanoparticles 139 20 700 600 500 400 300 200 100 0 Counts per pixel Larger pinhole 15 Notch re filter 5 3 m Closed er shutter um Figure 7 3 Light scattering image of individual gold particles The detail images on the left correspond to the dashed line squares in the main image aperture beyond which the signal reduces without any further increase of the signal to background 7 2 3 Spectra of colloidal gold nanoparticles Using white light illumi
124. e Tp it was possible to reproduce quan titatively all the photo induced characteristics of the QD blinking observed in the experiments decrease of the on time fraction increase of the cycles per second and decrease of the maximum on time So three characteristic parameters of the blinking can be accounted for by a single parameter in the simulations The photo induced lifetime of the on state Tp that best reproduced the experi mental data for each excitation power P is plotted in figure 6 18 as a function of P on linear log linear and log log scales The excitation intensity dependence of Tpz glass and Tpz rro are noticeable differ ent While the dependence of Tpz rro on P can be satisfactorily fitted by a single exponential the dependence of Tpr glass cannot Instead TPI glass aS a function of P can be represented by either a double exponential or a power law with exponent close to 1 The error bars in figure 6 18 were obtained by searching the extreme values of Tp that could still reproduce the experimental data within the statistical errors of the cycles per second and the on time fraction The photo induced lifetime of the on state Tp can be interpreted as the inverse of a photo induced rate from the on to the off state koff Pr 1 kopppr 6 9 TPI In the case of QDs on glass it is observed that the dependence of Tp with P roughly follows a power law with exponent 1 In this case Kaper X P 6 10 This indicates the
125. e both excitation of evanescent modes in the metal and direct energy transfer from the excited state of the molecule to the metal 63 64 These electromagnetic non radiative de excitation channels join the intrinsic ones to further reduce the fluorescence quantum yield of the molecule and at short distances lt A 30 are so dominant that they can quench the emission completely 41 The excitation and the emission of a single molecule in the experimental condi tions are modelled in the next two sections with the method presented in chapter 3 Table 4 1 lists the parameters used for the sample layers in the calculations The emission wavelength of the DilC1 5 and the dielectric constants of the layers were experimentally determined 46 The fluorescent dye molecules are considered to be in the interface between the polyelectrolyte spacer and air 70 Single molecule fluorescence through a thin gold film d nm A 633 nm A 670 nm glass 2 26 2 25 Table 4 1 Properties of the sample layers Thicknesses d and dielectric constants e for Au 44 12 65 i 1 01 14 94 i 1 1 the excitation and emission wavelengths A of the sample layers The glass substrates and the air are considered semi infinite spacer 0 1000 air Photo bleaching Fluorescent dye molecules undergo irreversible photo bleaching The reaction mechanism of photo bleaching is yet unclear but it is found experimentally that for the DiIC1 5 dye it o
126. e diaphragm is sufficient if one works with a rail that allows for translation along the optical axis 24 The fluorescence and light scattering confocal microscope a xX Diaphragm Diaphragm a y Main Reference CEE optical laser s Z axis beam b Achromat lens C Single mode Retarded fiber interference plate Figure 2 10 Collimation and alignment of the illumination beam a Definition of an optical axis b Introduction of the collimation lens centered and perpendicularly in the optical axis c Positioning of the optical fiber figure 2 10 a Then the collimating achromatic lens is introduced centered and perpendicular to the optical axis i e the reference laser beam should not deviate see figure 2 10 b From now on the collimating lens should not be moved in any direction but along the optical axis z Next the reference laser beam can be re moved and the optical fiber should be placed near the optical axis figure 2 10 c The distance d along the optical axis z between the fiber tip and the lens should be adjusted in order to collimate the light Collimation can be easily checked with the help of the retarded interference shear plate 09SPM001 Melles Griot GmbH which produces interference fringes parallel to the beam when it is collimated 23 Then the position in the plane perpendicular to the optical axis x y and the di rection 0 of the fiber should be adjusted The direction of the collimated
127. e direction of the molecule tran sition dipole In the experimental conditions this electric field is produced by the excitation light when it is focused by the microscope objective through the layered system of the samples In this section a theoretical method to calculate this field is presented 3 3 1 Electric field distribution near a geometric focus in a layered system The time independent part of the electric field E in a point r near the geomet ric focus of an aplanatic optical system such as the microscope objective can be calculated by adding the contributions of an angular distribution of electromagnetic 48 Single molecule fluorescence through a layered system plane waves traveling to the focus as 36 E r C N f Eoe dQ 3 19 Q where is a constant Eo is the complex amplitude of the plane waves at an origin normally the Gaussian focal point from which the position of the point of interest r is measured k is the wavevector of the radiation and Q is the solid angle defin ing the angular distribution of plane waves f is a scalar function called strength function it depends on the properties of the focusing system and accounts for the different intensities of rays focused in different directions Equation 3 19 is valid for an aberration free optical system and under geometric optics approximation i e the distance between the exit pupil plane and the image region dy in figure 3 3 and the linear dimensions of th
128. e ek 55 ii CONTENTS 4 Single molecule fluorescence through a thin gold film 57 AT Introd ti n y sa Arn eis eras a inh SL EARLS N ae ei 57 423 Experimental 2 2 ostein 3 eS doe eae es oP oe eae ae ay aa sede rl 59 4 2 1 Sample preparation 2 42 Area SO ee eg eo eo Se 59 4 2 2 Measurement et ee a a 60 4 3 Single molecule fluorescence images through a thin gold film 62 4 3 1 Full am images ecd acea 2 gatta mar sheet 62 4 3 2 Different illumination modes 64 4 3 3 Influence of the separation distance to the gold film 66 4 4 Modelling the experimental scheme 68 4 4 1 Fundamental concepts oaoa hee re in 68 4 4 2 Detectable fraction of the emitted fluorescence 71 4 4 3 Excitation field at the chromophores position 75 4 4 4 Theoretical fluorescence signal 78 A5 Conclusi Ns o s a Anp a re ae eM dt eg de R a 81 5 Single molecule fluorescence dynamics 85 5 1 Electronic transition rates ooo ee 85 5 2 Kinetic traces analysis methods gt ooa aa ee a au 4 88 5 2 1 Autocorrelation analysis wer Dede a de 88 5 2 2 Trace histogram analysis ooo a a de Be rn 90 52 3 COMPA iSl a n So sie Moen a ee en Ze re 95 5 3 Experimental au I a Be oe a Be dah oe aes eg 97 5 3 1 Sample preparation we pe ek ee Soe Sree Se es 99 5 3 2 Measurement ssa 3202 2 4 5 REO A Pe eee RA ged 100 5 4 Influence of a nearby thin Au film on
129. e exit pupil front lens of the microscope objective are assumed to be large in comparison to the wavelength This approach was first proposed by Richards and Wolf 36 37 and used by a number of authors to tailor the focus geometry 38 or to determine the three dimensional orientation of single molecules 39 40 Those works considered the focus of light to occur through an homogeneous medium or in the presence of a single interface In contrast in the model presented here the focusing of light occurs through a layered system The mentioned approach needs therefore to be generalized to consider the focusing of a plane wave front through a multilayer system The next paragraphs are dedicated to the calculation of the electric field at the Gaussian focus Eo the product k r and the introduction of the strength function f in order to write the integral 3 19 for the focusing of light through a layered system Figure 3 3 shows the geometry of the problem and introduces some of the variables and the coordinates used in the following calculations A monochromatic linearly polarized plane wave front is considered as the source field Esre which is focused through a layered system with a microscope objective In order to account for the amplitude of the plane waves at the focus Ep it is first necessary to find an expression for the electric field generated by a focused wave in any point of the layered system A Cartesian coordinate system is considered
130. e gold film present a much broader distri bution of Ta with an average of 0 61 x 10 1 s In this case the range of Ta values observed is narrower than the theoretically predicted In the presence of the gold film the ratio T gia T eta 18 approximately 5 5 and the experimentally observed ratio of maximum to minimum I is approximately 3 To illustrate this a simulated distribution is shown together with the experimental data The experiments do not probe the extremes of the P21 distribution because of two reasons The low I 2 end is not probed because it corresponds to molecules which transition dipole lies par allel to interface and as demonstrated in chapter 4 those molecules are practically undetectable through the gold film This is supported by the fact that even though the calculations predict for the parallel molecules near gold a minimum T around 25 smaller than in the case without gold the observed minimum values of I are practically equal in both cases The high I 2 end of the distribution is not probed because they cannot be measured due to the limiting time resolution of the APD 1ns 5 4 2 Influence on koff The distributions of koff for the molecules in the samples with and without gold present different average values and similar broadness However before drawing any conclusion from this results it is necessary to take into account that koff depends on the excitation rate Ij equation 5 5 and the exc
131. e kinetic trace less probable For this reason for all other parameters fixed a short kinetic trace has with higher probability a larger number of cycles per second In order to permit a proper comparison of the experimental data to the simula tions T was set to the length of the experimental traces for each excitation power shown in the graphs of figures 6 4 and 6 6 126 Photoluminescence blinking of Zno 42Cdo ss5e nano crystals m exponent of the power law probability density The histogram analysis method retrieves systematically values of m which are slightly smaller than the actual value All simulations were performed with m be tween 1 72 and 1 74 The values obtained from the simulations are shown in figure 6 8 the error bars indicate the range of values of m obtained from approximately 10 simulations with the other input parameters fixed Ion and Ioff background and net fluorescence intensity The trace histogram method retrieves systematically values slightly higher for Iopr and lower for Ion due to the mized bins explained in section 5 2 2 see also section 5 2 3 In the simulations the experimental values of orp and Jon net figure 6 7 were used as starting values Then the inputs were adjusted in order to obtain from the analysis values more similar to the experimental ones see figure 6 7 tmin Minimum allowed on or off time The influence of an arbitrary tmin was tested by varying it from 10 to 1 ms In ge
132. e recombination To explain the power law probabilities of the on and off times it is necessary to introduce tunnelling barriers that fluctuate in the time scale of the blinking 111 Both Shimizu and Kuno can explain the power law probability densities for the on and off periods but they differ in the following The model of Kuno since the tunnelling barriers fluctuate in the same time scale of the blinking predicts some residual memory effect i e short long periods should preferably be followed by short long periods The model of Shimitzu does not predict any residual memory effect because after each on off transition the random walk of the trap state starts from the same point This residual memory effect was not observed in the experiments 111 With respect to the physical processes involved in the blinking one of the first proposed models involves Auger ionization of the excited QD In this model a charged QD is a dark QD and therefore the average neutralization time corresponds to the average off time 102 Electrostatic force microscopy experiments showed that QDs can be charged both by thermal and photo induced processes 112 Still single rate charging and neutralization processes cannot explain the power law probability densities of the on and off periods 101 Two important contributions to the discussion of QD blinking have been reported more recently First Schlegel et al 113 introduced the concept and experimental evi
133. ectable fraction of the emitted fluorescence Following the approach presented in section 3 2 the electromagnetic decay rates of a fluorophore at the air spacer interface of the samples were calculated The pa rameters were set in order to represent the experimental conditions The wavelength was set to 670 nm the emission wavelength of the DilC1 5 The dielectric constants of the materials for this wavelength are listed in table 4 1 A gold film thickness of 44nm was considered and the spacer thickness was varied from 0 to 1000 nm in order to observe the influence on the rates of the separation distance between the molecules and the gold film Figure 4 9 shows the normalized decay rates of a parallel top and a perpen dicular bottom dipole on the air side of the spacer air interface as a function of the spacer thickness The total decay rate the decay rate to the air Tair and the glass semi space T glass and the electromagnetic non radiative decay rate Dir Per Fair Pgiass are plotted all normalized to the total emission of a free dipole in air For very thick spacers as the influence of the gold film vanishes all the rates tend to a constant value At very short distances from the metal film the excited molecule can very effectively transfer its excess energy to the metal Due to this dominant non radiative decay channels for separation distances in the order of A 30 the total decay rate diverges to extremely high
134. ectric constants e The uppermost and lowermost layers do and dy are considered to be semi infinite and all the interfaces are assumed to be plane and parallel The total thickness of the layered system is assumed to be small in comparison to the size of the incident plane wave front A plane electromagnetic wave defined by its wavevector k the parallel to the incidence plane component of its electric field EP and the perpendicular component E is considered to strike one of the outermost interfaces with an angle of incidence o The angle of incidence 69 determines the component of the wavevector parallel to the interfaces k which due to the electromagnetic boundary conditions remains constant through the whole multi layer system The electric field amplitude and phase in a given z position of the layered system which can be thought as the interference result of the waves successively reflected and refracted at the interfaces can be calculated via the TMA as a function of k or equivalently as a function of 6 arcsink k By considering the symmetry of the dipole field in front of a plane interface it is found that the emission rate of a molecule with an arbitrary out of plane orientation I see figure 3 1 can be calculated from the emission rates of the two limiting cases of dipoles parallel Tj and perpendicular I to the interface 34 I sin Ty cos Ty 3 5 Equation 3 5 is valid after 27 i
135. ecule due to excursions from the excited singlet state to the triplet state the autocorrelation decay is a single exponential see figure 5 3 Based on the 3 level model introduced in section 5 1 it is possible to obtain an analytical expression for this exponential decay 72 73 90 Single molecule fluorescence dynamics C r 1 Ae ae kon koff Fon a konlon kossloss Ion and Iofs are the intensities of the on and off states respectively kon and koff the rate constants of the onSSoff transitions as defined in section 5 1 equations 5 4 and 5 5 The equations 5 12 are valid under two conditions 68 First the times probed by C r are long in comparison to the total radiative decay time which means that photon anti bunching and Rabi oscillations are not visible Second the assumption is made that the inter system crossing rate los see figure 5 1 and the triplet relaxation rate 3 are much smaller than the total decay of the excited singlet 21 Ta nr P23 Single molecule triplet blinking normally fulfils the two conditions Values for A and x can be obtained from an exponential fit to the experimental autocorrelation see figure 5 3 Then by combining equations 5 8 and 5 12 it can be found that k A I APE d koff 5 13 All the parameters in equation 5 13 can be determined experimentally Iofs is the average experimental background intensity and is the overall average inten sity i e t
136. eflection 58 Single molecule fluorescence through a thin gold film became possible 48 49 and since then an increasing number of scientific questions are being addressed on a single molecule level Surprisingly only one experimental work was reported of single molecule fluorescence near a metal surface Yokota et al imaged single fluorescently labelled proteins with a surface plasmon fluorescence microscopy set up in 1998 50 In those experiments excitation was accomplished by the surface plasmon field at the metal solution interface via a Kretschmann prism coupling configuration and fluorescence collection was done with a high numerical aperture objective from the other side of the sample The same geometry was later modelled by Enderlein 51 k b c th Chromophore m I y E Mae a Y ff Spacer l o E Aufim 1 j y f Glass Dichroic N substrate 7 mirror Semple Se K Objective v v i Fluorescence i i light Figure 4 1 Experimental configuration a The Cartesian coordinate system adopted takes the polarization direction of the illumination beam as reference for the x axis and the z axis is perpendicular to the interfaces The orientation of the transition dipoles of the chromophores u is defined by amp and 8 The direction of emitted or incident radiation of wavevector k is defined by 8 and w b High angle of incidence illumination forbidden light FL c Low angle of incidence ill
137. ely charged dots increased upon illumination with light of energy above the absorption band edge of the QDs Following this line the different on time fraction behavior observed for QDs on glass and on ITO coated glass substrates could be explained by taking into account that glass is an insulator and ITO is a semiconductor of relatively low work function 4 4 4 7 eV 117 Then on ITO the electrons have a relatively high mobility and given the fact that the QDs are either positively charged or neutral they can rapidly build up a negative charge density around a QD As a consequence ejection of an electron from the QD to the substrate is less viable leading to longer on times for the QDs on ITO However it should be noted that the physical picture described above is contra dictory to the observation that even at very low excitation intensities the QDs on ITO seem to be half the time in the off state i e ionized Correlations between adjacent on and off times Additional information about the physical nature of the fluctuations involved in the blinking process may be obtained via correlations of adjacent on and off times Figure 6 11 shows as an example the scatter plots of the on times vs the successive on times the off times vs the successive off times and the on times vs the successive off times of the on and off periods detected from QDs on glass under an excitation power P 0 74kW cm A similar behavior is found for
138. em Phys vol 89 no 6 pp 3435 3441 1988 A L Efros and M Rosen The electronic structure of semiconductor nanocrystals Annu Rev Mater Sci vol 30 pp 475 521 2000 Observation of the Dark Exciton in CdSe quantum dots Phys Rev Lett vol 75 pp 3728 3731 1995 A L Efros and M Rosen Band edge exciton in quantum dots of semiconduc tors with a degenerate valence band Dark and bright exciton states Physical Review B vol 54 no 7 pp 4543 4856 1996 M Nirmal B O Dabbousi M G Bawendi J J Macklin J K Trautman T D Harris and L E Brus Fluorescence intermittency in single CdSe nanocrystals Nature vol 383 pp 802 804 1996 M Danek K F Jensen C B Murray and M G Bawendi Synthesis of luminescent thin film cdse znse quantum dot composites using cdse quantum dots passivated with an overlayer of znse Chem Mater vol 8 pp 173 180 1996 202 BIBLIOGRAPHY 104 105 106 107 108 109 110 111 112 113 114 115 B O Dabbousi J Rodriguez Viejo F V Mikulec J R Heine H Mattoussi R Ober K F Jensen and M G Bawendi cdse zns core shell quantum dots Synthesis and characterization of a size series of highly luminescent nanocrystallites J Phys Chem B vol 101 pp 9463 9475 1997 W Guo J J Li Y A Wang and X Peng Luminescent cdse cds core shell nanocrystals
139. ene PS colloids on glass substrates evaporation of a thin gold film ion milling and removal of the PS colloids PS colloidal nano spheres Polysciences Inc with a 400 nm were physisorbed on the glass coverslips by immersing the substrates for 30 minutes in the PS colloid Stock solutions of the PS colloid 2 1 solids in water were stored at 4 C Before deposition the PS colloid was allowed to warm to room temperature sonicated for 5 minutes and then diluted 10 times in ethanol The coverslips with adsorbed colloids were rinsed with ethanol and dried with a stream of nitrogen Next a process of subsequent deposition of a thin gold film was carried out twice followed by ion milling The samples were mounted in a home built holder that allows to control the angle of evaporation with respect to the metal source tilting angle as well as the rotation angle of the cover slips A 15nm gold film was thermally evaporated thermal evaporator Edwards A100 with a tilting angle of 30 After that the samples were rotated by 90 and a second gold film of 15nm was deposited with the same tilting angle The thicknesses of the evaporated films were controlled by monitoring the deposited mass with a quartz crystal microbalance Then Argon ion milling lon beam etching machine Microsys400 Roth und Rau Oberfl chentechnik GmbH was performed with a beam perpendicular to the substrate surface in order to remove all the gold not masked by the colloidal PS
140. ent with the ionization model that states that a positively charged QD is a dark QD and a neutral QD is a dark QD supported by the electrostatic measurements on single QDs performed by Krauss et al 112 They reported that at room temperature and in air half of the QDs were positively charged and half were neutral and that the fraction of charged dots increased upon illumination with light of energy above the band gap of the dots A careful observation of the correlation between adjacent on and off times shows a weak residual memory effect for consecutive on times and for the consecutive off times No dependence of this correlation on the excitation power was observed No memory effect was observed between an on time and the adjacent off time This supports the idea of the two different process one ruling the on times and another ruling the off times Although such a memory effect was predicted by the model of ionization through fluctuating tunnelling barriers proposed by Kuno 110 it was never observed in experiments From the analysis of the intensity distributions it is found that the QDs present a Poisson distributed off intensity with average equal to the background intensity showing that the QDs indeed stop emitting during the off periods In contrast the on intensity distribution is in general not Poissonian Some kinetic traces or parts of kinetic traces can be found in which the on intensity is practically Poisson distributed Even
141. ental scheme In order to properly interpret the single molecule fluorescence signals and to understand the advantages and limitations of the present scheme the fluorescence excitation and emission of a single molecule under the experimental conditions is modelled following the approach described in chapter 3 The first part of this chap ter introduces the concepts necessary to understand the theoretical results presented next Section 4 4 2 considers the de excitation rates and section 4 4 3 the excitation rate In section 4 4 4 the theoretical results obtained for the excitation and for the emission are brought together to calculate the theoretical fluorescence signals of a fluorophore under the experimental conditions 4 4 1 Fundamental concepts The excitation and emission properties of a fluorophore are greatly affected by the properties of the surrounding media In particular a nearby metallic surface can 6No single molecule fluorescence signal was detected on the samples with 1 bilayer spacer Therefore the signals must be smaller than the background intensity The value of Ij for the 1 bilayer spacer sample shown in figure 4 7 d was calculated assuming a signal equal to half the background intensity Reproduced with permission of the authors 4 4 Modelling the experimental scheme 69 a Ip Tr No Figure 4 8 Schematic representation metal of the excitation and de excitation rates of a fluorophore a In free space b
142. ential probability distribution of equation 5 16 is continuous in time and considers the 5 2 Kinetic traces analysis methods 93 detection time of one photon Then it has a pre factor equal to one jim Puxsat 1 Instead the histogram of figure 5 5 is discrete in time and considers the detection times of a number of photons Non No fr photons of the on and off state respectively are grouped in bins of bin width bw In this case the pre factors Aon and Aoryr for the discrete probability distributions are the number of photons to be found in a certain bin From equation 5 18 t bw N f Ij el t dt A i a N 1 e 5 20 where 7 stands for on or off depending on the population Then the total number of photons detected in each state can be directly calculated from the double exponential pre factors Aon and Aofp as Aj Finally considering that the average intensity of each of the states is simply the ratio between the total number of photons and the total time in that state I N T the total time that the molecule spends on the on or off state can be calculated as A T 2 2 I 1 el bw 5 22 Next with the information obtained from the inter photon time histogram an iterative process is carried out in order to find the optimum bin width and threshold that allow to distinguish between the on and off states with the highest accuracy and time resolution from a histogram of the kinetic trace
143. er behave optically as being on the air side of the spacer air interface Nevertheless it is worth to consider the two possibilities again and see whether that statement is confirmed or not The calculated total electromagnetic decay rates of molecules on the air side of the spacer air interface which transition dipole lie parallel and perpendicular to the polyelectrolyte surface are listed in table 5 3 Since these rates are normalized to the total emission of a dipole in air they are proportional to the actual rates The parallel and perpendicular dipoles are the limiting cases and molecules with orientations in between have intermediate decay rates In the absence of the gold film the ratio T Kia 1 T ior 15 1 65 and can quantitatively account for the observed range of Iz values The asymmetry of the distribution in favor of lower values of I arises from the facts that the molecules are randomly oriented in the sample and that they were randomly selected to be studied Then as there are more molecules oriented parallel to the surface the distribution is richer in lower values of Ts To illustrate this effect a Monte Carlo simulated distribution of T values is shown in figure 5 12 together with the experimental datat This result supports The simulations consider molecules randomly oriented in three dimensions and that are ran domly selected The simulated distributions as the experimental ones are made up of 33 and 42 values for
144. er light 2 3 Alignment 27 Collimated a Collimated beam beam beam beam Slightly Highly convergent divergent T a convergent divergent Figure 2 12 Huygenian ocular The system consists of two conjugated lenses one has a much shorter focal distance than the other f gt f2 a A collimated beam is focused by the first lens into the focal point of the second lens therefore producing a collimated beam of small size b If the incoming beam is slightly convergent divergent the first lens focuses it further from closer to the second lens which due to its short focal distance produces a highly convergent divergent beam and pinhole can be adjusted with micrometer precision in x y and z directions The adjustment should be done in sequential order Then by means of mirrors if necessary polarizing beam splitters or dichroic mirrors the reflected beam can be directed to the combination of detectors APD PMT spectrograph required by the experiment Alignment of the avalanche photo diode The reflected beam should be focused into the small active area of the APD with an 100 mm focal length achromatic lens For this purpose the APD is mounted on an xyz stage with micrometer precision The first alignment can be accomplished by eye but fine adjustments can only be done when a real fluorescent or scattering sample is measured by optimizing the APD position in order to maximize the signal Alignment of the p
145. es present all very similar intensities 1 0 a E 0 8 0 6 0 4 S A 0 2 0 0 500 600 700nm A a E 0 8 5 0 6 0 4 S A 0 2 0 0 500 600 700nm 35 40 45 50 A um Figure 7 5 Light scattering of C shaped gold nanoparticles On the right SEM top and scattered light bottom images of the particles On the left spectra of particles a and b for the two polarization states of the illumination beam Light scattering spectra were recorded from several particles with the illumi nation beam polarized along the symmetry axis of the C particles and along the perpendicular direction Some particles showed different behavior for the two polar izations and some particles showed almost indistinguishable spectra As example the spectra from the particles marked as a and b are shown on the left of figure 7 5 Particle a does not show a noticeable difference between the two polarizations The spectra differ slightly from each other in the short wavelength range but in both cases the maximum lies at around 596 nm In contrast particle b presents a clearly different behavior for the two illumination polarizations When illuminated with light polarized along the symmetry axis of the C the resonance is found at around 570 nm and for the perpendicular polarization at around 596 nm This effect can be explained by the fact that even though the fabrication process should lead to parti cles with uniform orientation the C shaped pa
146. es are constructed by taking into account the driving signal i e dividing the image in figure 2 6 in two through the center the images present two artifacts First due to the inertia of the piezoelectric drivers at the direction inversion points the images appear blurred at one edge the forward image on the left and the backward image on the rigt side Second due to the total delay the forward and backward images are shifted with respect to each other Correction of these artifacts is extremely important because they directly affect the accuracy in the determination of the position of fluorescent dyes or any other measurable feature on the sample For this reason in order to construct consistent forward and backward images that permit proper position determination the data acquisition software has to take into account the behavior of the piezoelectric stage The software code can be found in appendix A 3 In principle a complete correction could be accomplished by taking into account the capacitive monitor signal which is supposed to report the actual position of the stage at any time Such correction consists of three steps First a number of lines are scanned with the set parameters and the forward and backward linear regions of the monitor signal are determined as well as the number of invalid pixels at the inversion points Second a new scan is performed with an additional number of pixels equal to the invalid pixels in the x direct
147. es were programmed in a specific computer language ADBasic 20 provided by the AD DA manufac turer The second part is the PC user interface With this software the user can set the scanning parameters as well as observe on line the collected data and save it in an appropriate format The PC user interface was programmed in Igor 21 in order to take advantage of Igor s built in capabilities for data treatment Finally the third part consists of a set of routines required for the communication between the PC user interface and the local CPU of the AD DA converter These routines were programmed in C The code of all these programs is presented and described in the appendix A 3 CCD camera and spectra acquisition control The CCD camera can be controlled with specific software provided by the man ufacturer and details about the operation can be found in the operation manual of the camera 22 All the relevant parameters for the operation of the CCD camera can be controlled such as the active region of the CCD sensor and the collection time It is also possible to record up to 9999 frames as a function of time with a minimum collection time of 1ms maximum repetition rate of 1000 spectra per second 22 The fluorescence and light scattering confocal microscope TCSPC unit control The TCSPC unit has its control software control provided by the manufacturer details about the operation can be found in the user manual of the TCSPC mod
148. et al 110 111 but it was not observed experimentally 6 4 Modelling the QDs blinking The modelling of the QDs blinking is performed to address the following ques tions Are the on intensities deviated from a Poisson behavior really due to partly detected short on times produced by the power law probability density Can the photo induced shortening of the on time fraction be explained by the presence of an additional independent pathway which introduces a characteristic lifetime for the on state If that is the case how does this characteristic lifetime depend on the excitation intensity 124 Photoluminescence blinking of Zno 42Cdo ss5e nano crystals 6 4 1 Blinking model Since there is no analytical description of the QD blinking the simulations were performed via a Monte Carlo method that takes into account the experimentally observed characteristics of the blinking The emission of a QD is considered to switch between an on and an off state each of them with a characteristic intensity Jon and osp Two processes are responsible for the blinking one spontaneous onoff switching and one photo induced transition from the on state to the off state Under no illumination the probability of a certain length of an on or an off period P t and P t ff respectively follows a power law with an exponent be tween 1 and 2 figures 6 4 and 6 6 In order to make the power law probability density normalizable it is necessary to restrict
149. f this chapter are dedicated to detailed description of the micro scope The first section explains the confocal principle The second section describes the different components and their functions Finally the last two sections explain how to properly align and operate the instrument 2 1 The confocal principle An ideal optical microscope would examine each point of the specimen and mea sure the amount of light scattered or absorbed by that point However if many of such measurements were performed simultaneously every point in the image plane would be clouded by aberrant rays of scattered light coming from other points of the sample Marvin Minsky 7 8 found in 1955 a simple and elegant solution for this problem the confocal arrangement see figure 2 1 In the first place it is possible to illuminate only one point of the specimen at a time by using a microscope objective to focus the light spread by an aperture pinhole illumination pin hole in figures 2 1 a and 2 1 b As a consequence the amount of light in the specimen is reduced by orders of magnitude without reducing the focal brightness at all fundamentally important to prevent photo bleaching in single molecule fluorescence experiments Still due to multiple scattering some extra rays In this context point means a diffraction limited spot 6 The fluorescence and light scattering confocal microscope coming from different points of the sample dashed rays in
150. facturer nm Repetition Light source Coherent inc Argon ion Innova 90C laser Mode Locker PulseDrive APE GmbH Helium Neon Uniphase inc laser i Becker und Hickl GmbH Laser diode BHL 150 Xenon arc 300 800 OSRAM lamp with water 150000 XBO150 IR filter Table 2 1 Light sources CW stands for continuous wave and FWHM for full width at half maximum used to focus the illumination beam onto the samples The microscope objective is mounted on a xyz stage with micrometer precision Fluorescence measurements For fluorescence measurements laser light is used Line pass filters Omega Op tics Inc are used to refine the illumination wavelength and a dichroic mirror AHF AG is used to direct the illumination beam to the microscope objective and to separate the excitation from the fluorescence light Light scattering measurements For light scattering measurements annular illumination see next subsection is combined with areduced detection beam Instead of a dichroic mirror a 50 50 beam splitter OWIS GmbH is used to direct the illumination beam to the microscope objective and to separate the excitation from the fluorescence light Focusing angles and annular illumination The TEM00 light provided by the single mode fiber is collimated with an achro matic lens with minimized spherical aberration Therefore the illumination beam 2 2 Description of the home built confocal microscope 11 can be
151. fer is performed In contrast the TCSPC data is not inherently binned because the TCSPC unit records the detection time of each photon The data employed for further analysis is therefore the TCSPC one The 18The sample is moved to the position marked by the cursor A in the last active image also called top image because the operating system Windows places the last active window on top of the others 2 4 Operation 37 BISAC_v7 Igor Pro 4 02A ELE zZ ADWin Shutter Open Range um Measurement Surtace Scan Minetic S J a Ji v1 Jar lau Figure 2 20 Kinetic trace screen Screen shot of the user interface software during the recording of a fluorescence vs time trace The software records the data and updates the display periodically at a frequency defined by the Time parameter in the Kinetic division of the control panel data recorded for the PC user interface is discarded after each measurement but is of practical importance because it allows to visualize on line the kinetic trace 2 4 4 Spectra measurements The procedure to measure the spectrum of light from a specific point of the sample is basically the same as that to measure a kinetic traces Once a feature of interest is identified in a confocal image its position should be marked with the round A cursor and the Start button of the Kinetic division of the control panel should be pressed The sample is then moved to the desired position the shutter
152. ference Corr_Xmondisplay j fit Corr_XmondisplayF j j 1 While difference gt Pixelsize XiF j 1 If Corr_Xmondisplay XiF Corr_Xmondisplay 2 Pixels 1 XfB 2 Pixels 1 Elseif Corr_Xmondisplay XiF gt Corr_Xmondisplay 2 Pixels 1 j 2 Pixels 1 Do difference Corr_Xmondisplay XiF Corr_Xmondisplay j fel While difference gt Pixelsize XfB j 1 Else Set up control and data acquisition software A 2 Igor routines 171 XfB 2 Pixels 1 j XiF Do difference Corr_Xmondisplay 2 Pixels 1 Corr_Xmondisplay j j 1 While difference gt Pixelsize XiF j 1 Endif J 0 Do difference fit_Corr_XmondisplayB j Corr_Xmondisplay j Pixels jt l While difference gt Pixelsize XiB j 1 Pixels If Corr_Xmondisplay Pixels 1 Corr_Xmondisplay XiB XfF Pixels 1 Elseif Corr Xmondisplay Pixels 1 gt Corr Xmondisplay XiB j Pixels 1 Do difference Corr_Xmondisplay j Corr_Xmondisplay XiB fel While difference gt Pixelsize XfF j 1 Else should never happen aber j XiB Do difference Corr_Xmondisplay j Corr_Xmondisplay Pixels 1 j 1 While difference gt Pixelsize XiB j 1 XfF Pixels 1 Endif Lines for future if needed improvement to consider the case in which Calibration fails I XfF XiF lt GoodPixels XfB XiB lt GoodPixels return 0 endif If XfF XiF gt GoodPixels Do XfF 1 While XfF XiF gt GoodPixels Endif If XfB XiB gt GoodPixels Do XiB 1 While XfB XiB gt GoodPixels Endif 172 Set up
153. fied by dialysis with a membrane tube with a cut off molecular weight of 3500 Spectra por 6 Spectrum Laboratories Inc The dialysis was conducted for six days as the water was exchanged twice a day PSS was purified by a controlled precipitation method The polymer was dissolved in water and then the solution was transferred slowly drop by drop into ethanol gt 98 Riedel de Ha n close to its freezing temperature 50 to 60 C The precipitated polymer was then filtered and dried in vacuum After this treatment the fluorescence background was reduced considerably and reached a level low enough to allow for single molecule measure ments Namely a typical sample with four PAH PSS bilayers presented between none and two diffraction limited impurities with fluorescence signals comparable to the one of a single dye molecule within an area of 10 x 10 um Single fluorescent dye molecules 1 1 3 3 3 3 hexamethylindicarbocyanine iodide DiIC1 5 Molecular probes maximum excitation at Aere 638 nm maximum emis sion at Aem 670 nm were deposited electrostatically on a negatively charged sur face terminated with PSS by immersing the samples for one minute in a 107 M Milli Q water solution of the dye Then the samples were rinsed with Milli Q water and dried with a stream of nitrogen 4 2 2 Measurement The samples were studied in the SCOM set up described in chapter 2 Fluores cence of single molecules was excited and detected
154. figures 2 1 a and 2 1 b could reach the detectors However it is possible to reject those rays using a second microscope objective to image on the same confocal point of the specimen a second pinhole aperture detection pinhole in figures2 1 a and 2 1 b placed in front of the detector Then as shown in figure 2 1 a an elegant symmetric configuration is obtained consisting of a pinhole and an objective lens on each side of the specimen The term confocal should be clear now it is used to indicate that both illumination and collection pinholes lenses are focused on the same point of the object a c Specimen Dec 19 1961 Illumination pin hole Filed Mev 7 1997 Detector Pan hte fs pin hole Light source b Specimen l Illumination pin hole Beam splitter Light source llumination Light in focus Light positively defocused Detection Mia ha pin hole serres Light negatively defocused Detector Figure 2 1 Confocal principle c Original sketches of the confocal principle from the patent by M Minski 1957 7 a and b The original sketches are reproduced to better accompany the explanation in the text FIG 2 in c shows the original mechanical scanning stage by Minski The original drawing of the confocal microscope in the patent by M Minski from 1957 is shown in figure 2 1 c The sketch named FIG 1 in figure 2 1 c shows the schematic of the configuration described
155. function of the excitation intensity Finally the blinking of QDs is modelled via a Monte Carlo method In the modelling an single rate transition from the on to the ofl state is proposed to account for the photo induced effects observed in the blinking The dependence of this transition rate on the excitation intensity is analyzed 6 2 Experimental 6 2 1 Sample preparation Monocrystalline Zno 42Cdo 535e nano particles were provided by the Department of Materials Science of the National University of Singapore The preparation method and the ensemble photoluminescence characteristics of the Znog 42Cdo 5sSe QDs are described in 116 The nano crystals studied here have an average size of 6 2nm As shown in figure 6 1 the QDs present an ensemble band edge absorption peaking at 515nm and an ensemble emission maximum at 550nm with a full width at half maximum FWHM of 40nm Figure 6 1 b also shows the emission spectra of four individual Zno 42Cdo 555e QDs on a glass substrate the FWHM of single dot emission spectra ranges from 13 to 24nm a b eee it Glass ensemble ER ITO ensemble 3 Jo A Toluene 5 E cdl 2 __ Glass single dot 5 3 Ss 2 c lt N J im oo ER POEMES 480 520 560 600 520 540 560 580 600 nm A nm Figure 6 1 a Absorption of Zno 42Cdo 5s85e QDs in toluene on glass and on ITO coated glass substrates b Ensemble on glass and ITO and 4 individual on glass photoluminescence spe
156. g Pho toinduced conversion of silver nanospheres to nanoprisms Science vol 294 pp 1901 1903 2001 J P Kottmann and O J F Martin Spectral response of plasmon resonant nanoparticles with a non regular shape Optics Express vol 6 pp 213 219 2000 204 BIBLIOGRAPHY 129 L A Blanco and F J G de Abajo Spontaneous light emission in complex nanostructures Phys Rev B vol 69 p 205414 2004 130 T Turkevich P C Stevenson and J Hillier A study of the nucleation and growth processes in the synthesis of colloidal gold Discuss Faraday Soc vol 11 pp 55 75 1951 131 G Frens Controlled nucleation for the regulation of the particle size in monodisperse gold suspensions Nature Phys Sci vol 241 pp 20 22 1973 132 U Kreibig and M Vollmer Optical properties of metal clusters Springer Verlag Ist ed 1995 133 Wavemetrics inc Igor XOP Toolkit Reference Manual v 3 1 1998 Acknowledgements This lines are dedicated to express my gratitude the people who made my research possible and my stay at the MPI P pleasant In first place I want to thank the supervisor of my dissertation Prof Dr Wolf gang Knoll Thank you for giving me the opportunity of carrying out my Ph D programm at your group at the MPI P I am greatly thankful for the support and freedom you gave me during these years Dr Maximilian Kreiter thank you for your scientific support you
157. g confocal microscope ine Scan AvgCts 26 Cps 1467 a a 4 Scan Parameters a Irirh LAAL Measurement Line Scan Surface Scan ae wife BESTE AO amp Bvss Figure 2 17 Scanning screen Screen shot of the user interface software during the scanning to acquire an image Line scanning The computer control also allows to perform vertical and horizontal line scans The lines are scanned repeatedly until the user stops the process The procedure for the scanning of a vertical or horizontal line is initiated by pressing the buttons V Scan or H Scan of the division Measurement Line Scan of the control panel The range of the line scan is set by the position of the cursors in any of the images The round cursor A sets the starting point and the squared cursor B sets the final point for the line scan Figure 2 18 shows a screen shot while a horizontal line is being scanned The limits were set by the cursors in the upper image forward scanned Line scanning is very helpful to optimize the focus of an image of single molecules as well as to adjust the APD position One way to optimize the focus of a single molecule fluorescence image is to make a line scan through the center of a detected fluorescence spot and adjust the z position of the sample in order to make the profile narrowest For extra help it is possible to check the Gauss Fit check box and a Gaussian curve is fitted after every scanned line and the full width at
158. gle dye molecules can be placed from the gold film before their fluorescence becomes undetectable To address this question samples with polyelectrolyte spacers of four different thicknesses were studied 100 40 80 30 60 40 20 20 10 26 25 22 20 i 15 14 y 10 ne 10 9 5 kWicm 4 11 9 KW 20 22 24 26 28 40 42 4 46 48 um um Figure 4 6 Influence of the separation distance to the gold film Images 8x8 um 2ms per pixel of the samples with different spacer thicknesses a 4 bilayers 24nm b 3 bilayers 15 nm c 2 bilayers 10 nm d 1 bilayer 4 5nm The excitation power used is shown in each image Figure 4 6 shows FL images of the samples with 4 3 2 and 1 bilayer spac ers corresponding to separation distances of the chromophores from the gold film of 24 15 10 and 4 5nm respectively In all the samples the great majority of 4 3 Single molecule fluorescence images through a thin gold film 67 the detected signals present the same characteristic pattern discussed above These images correspond to samples from the same batch produced under the same con ditions Therefore the same surface density of dye molecules is expected for all of them The detection of single molecule fluorescence becomes more difficult as the spacer thickness decreases note on the bottom left of each image the increasing excitation intensity used as the spacer thickness decreases The average signal to background ratio
159. h graph of each individual trace was computed A power law y Ax was fitted to each off time length histogram and the obtained exponent is written in the legend of the corresponding graph The probability of a certain length of an off period shows a clear power law over more than 4 decades in time and more than 6 and in some cases even 8 decades in probability density with no clear dependence on the excitation intensity The 116 Photoluminescence blinking of Zno 42Cdo ss5e nano crystals 1000 P 2 04 kWiem lem e P ni ON 1000 P TARN cm 1000 Pe 0 95 KiWiem 4 OFF OFF 100 OFF 10 fit m 1 63 fit m 1 63 fit m 1 59 2 a 10 4 g 94 c c c 4 3 1 Q Q Q S Oo 14 5 0 1 01 4 01 4 0 01 0 01 T 70s T 150s 0 01 0 001 0 001 X oe ee En ee 001 01 1 10 001 01 1 10 0 01 01 1 10 Period length s Period length s Period length s a P 0 74 Wiem aa 1000 P 0 026 kWicm 100 OFF 109 7 OFF fit m 1 60 10 4 fit m 1 75 a 10 2 2 c ce c 3 14 gt gt 14 8 8 8 0 1 0 14 0014 T 0 01 4 0 001 0 001 4 0 001 cry 001 01 1 10 0 01 01 1 10 0 1 1 10 100 Period length s Period length s Period length s Figure 6 4 Histograms of the length of the on and off periods obtained from the common computation of several individual kinetic traces of different Zng 42Cdo 5ssSe QDs on glass under various excitation intensities P The same to
160. he gold particles deposited on silicon wafers as described in 57 see figure 7 1 Sample preparation Colloidal gold nanoparticles were physisorbed on thin 0 13 0 16 mm glass cov erslips N 1 Menzel Gl ser The glass substrates were cleaned successively with Hellmanex 2 Milli Q water and ethanol gt 98 Riedel de Ha n In addition to remove any rest of organic material the coverslips were heated for two hours at 500 C in air To deposit the particles the gold colloid was let in contact with the coverslips for a certain waiting time and then spin casted The waiting time was ad justed in order to obtain a surface density suitable for single particle measurements the best results were found between 10 and 60 seconds Measurement Light scattering images and spectra were acquired with the home built SCOM described in chapter 2 set up for light scattering measurements A blocking disc with a diameter of 6mm was used to produce annular illumination section 2 2 1 and the pinhole size was adjusted in order to optimize the signal to background as explained below Images were acquired under laser illumination Ar ion A 514 5nm and white light illumination Xe arc lamp was used to acquire the spectra 7 2 2 Images of colloidal gold nanoparticles Figure 7 2 shows a light scattering image of the 20nm size gold nanoparticles obtained with circularly polarized 514 5nm Ar ion laser illumination Light scattering images c
161. he total number of detected photons divided by the total collection time Finally kon can be obtained from equation 5 12 5 2 2 Trace histogram analysis If one makes a time histogram of the raw data shown in figure 5 2 one finds some bins with a high number of photons corresponding to the on state and some bins with a low number of photons corresponding to the off state of the molecule figure 5 4 a Then it possible to choose a threshold T a number of photons marked by the horizontal line in figure 5 4 a and classify all the bins with less than 7 photons as off bins and all the bins with T photons or more as on bins From this histogram analysis it is possible to calculate the average length of the on and off periods that lead to the values of kon and Ko However choosing arbitrary bin widths and thresholds can yield misleading in formation As can be seen in figure 5 4 b a broad bin width can help to distinguish better between the two states but the price to pay is that many short on and 5 2 Kinetic traces analysis methods 91 a s Bin width 0 75 ms 40 2 30 2 ad M T 10 l o aw Lat AF La hl aa LLE 3 0 3 2 3 4 3 6 3 8 4 0 b 250 Bin width 5 00 ms 200 8 150 100 50 0 T T T T 3 0 3 2 3 4 3 6 3 8 4 0 c 49 Bin width 0 05 ms 8 2 6 O 4 2 04 3 0 3 2 3 4 3 6 3 8 4 0 Time s Figure 5 4 Effect of the bin width on the kinetic trace histogram analysis Time histograms of the kinetic trace data sho
162. hotoluminescence blinking of Z ng 49C dp 585e nano crystals Colloidal semiconductor nano crystals also known as quantum dots QDs are attractive fluorophores for a variety of growing applications in spectroscopy of single biological molecules 74 75 and quantum information processing using single photon sources 76 However severe intermittence in emission also known as blinking has been universally observed and represents an intrinsic limitation for the practical use of QDs The fluorescence behavior of single QDs was reported to change dramatically when they are adsorbed on a rough metal film 6 The observed changes include a fivefold increase in the observed fluorescence intensity and a striking reduction of the blinking showing that it is indeed possible to enhance the performance of the QDs by electromagnetic interactions with metallic objects Nevertheless since processes behind the photoluminescence blinking of QDs are not yet understood the enhancing effect cannot be exploited in a controlled manner The most commonly mechanism suggested for the blinking of QDs relates the switching between dark and bright states to ionization and neutralization events in the QD If that is the case the electric transport properties of the surrounding media is expected to influence the blinking behavior To investigate this the emission of Zno 42Cdo 535e QDs deposited on glass and on ITO coated glass substrates is studied as a function of time
163. hotomultiplier tube Due to the big active area of the PMT the alignment is considerably simple The reflected beam should be focused onto the active area of the PMT with an achromatic 100mm focal length lens This can be accomplished by eye and no further adjustments are needed Illumination of the active area of the PMT with high intensities even when it is switched off can increase the dark count rate of the PMT Eventually it can take several days until the nominal dark count rate is recovered For this reason the active area of the PMT should be covered during the alignment procedure 28 The fluorescence and light scattering confocal microscope Alignment of the spectrograph The collimated reflected beam should be directed first to the VPH grating Light dispersed by the grating is collected by the photo objective and focused on the CCD sensor Both the photo objective and the CCD sensor should be centered in the spectral range of interest The latter can be accomplished with the help of two different laser light sources for example the 460 nm line of the Ar ion laser and the 633 nm of the HeNe laser as shown schematically in figure 2 13 CCD camera Photo objective Transmission Grating Figure 2 13 Alignment of the spectrograph Both the photo objective and the CCD camera can be moved in x y and z directions as well as tilted in amp and 0 Using a two color beam helps to center the photo objective in the spectr
164. i e axially stig matic and obeying the sine condition all around the z axis Hence the plane wave front of the source field in the object space is transformed into a spherical wave front in the image space without aberrations The strength function f is in this case 3 As shown in section 2 2 1 a point of the objective rear lens can be equivalently defined in polar coordinates R Y or by a set 0 7 because R sin 0 Rri nglass NA 3 3 The excitation 51 found to be 37 f Vcos 3 27 Introducing in integral 3 19 the electric field components 3 24 the path difference exponent 3 26 and the strength function 3 27 27 Omax E r cf f Vcos 9 Ep 8 y el sin M cosh 2P 0089 sing dO dy 3 28 0 Omin where j stands for x y or z and the solid angle Q is integrated in spherical coor dinates The azimuth angle w is integrated all around the z axis and the focusing angle range Omin and Omar depends on parameters such as the NA of the objective the diameter of the illumination beam and the blocking disc in annular illumination see section 2 2 1 The integral 3 28 permits the calculation up to a constant of the electric field in any point near the focus of the microscope objective when the latter is illuminated with a linearly polarized plane wave front The electric field distribution over a region near the focus can be calculated by performing the integration for each point r of the region Nevertheless
165. ic drivers and electronic delays The curves are the driving signal to scan the x direction the ca pacitive monitor signal of the piezo stage and the actual x position of the piezoelectric stage after correction see section 2 2 2 The thin lines are linear fits used to determine the linear range of the x position Further information can be found in the appendix A 3 While the test lines for calibration are being scanned the screen of the user interface looks like in figure 2 16 The driving and monitor signals of the piezo stage are displayed together with the corrected x position see section 2 2 2 Once the scanning to acquire an image starts the screen of the user interface looks like in figure 2 17 On the left the forward and backward scanned images are displayed and updated line by line as the acquired data is transferred to the PC In addition the profile of the last scanned line is also displayed on the upper right corner of the screen 12It is possible that the monitor signal suddenly appears extremely noisy This is due to in stabilities in the monitor output of the amplifier If this happens the BNC connector from the monitor output of the amplifier should be unplugged for some seconds and the scanning should be restarted 13The images displayed on the user interface software are rotated 90 counter clockwise with respect to the sample as seen from the front of the microscope 32 The fluorescence and light scatterin
166. ict very low signal to background ratios for the fluorescence signals of single molecules on the spacer side of the interface The calculated fluorescence signals provide the explanation to the fact that the TL signals do not coincide with the FB or FL As can be noted by comparing the theoretical fluorescence signals for FB or FL to the ones for TL the reason is that those signals correspond to different molecules with different orientations Then a FB FL signal would coincide with a TL signal in only two cases First when the orientation of a molecule has an optimum compromise between x and z components and even in this case the detected signal would be very weak in comparison to the dominant signals of molecules oriented closer to the interface normal Second when more than one molecule lie close on the same diffraction limited spot to each other Due to the changes in the total de excitation rate introduced by the gold presence it is found that the number of detectable photons emitted by a parallel molecule in the sample with the gold film is 26 of the photons the same molecule would emit in a sample without the gold film For the case of a perpendicular molecule due to the surface plasmon enhancement this percentage rises to 140 For the same sample geometry the detection through the gold film was compared to the detection from the air side with an 0 9 NA objective The surface plasmon mediated detection through the gold film is more
167. in translate the parameters into terms of AD digits for the AD DA card Function CalculateParametersForAdwin Xi Xi_um 32768 80 32768 Piezo table in closed loop moves 80um Yi Yi_um 32768 80 32768 for the 0 10V 32768 ADWin units range Scanrange Scanrange_um 32768 80 The 80 should be calibrated Xf Xi Scanrange Yf Yi Scanrange Pixelsize Scanrange Pixels Xf_um Xi_um Scanrange_um Yf_um Yi_um Scanrange_um End ExecuteSAC2ForScan executes the Sac2 XOP which then calls the routines to move the piezo stage Function ExecuteSAC2ForScan scantype 0 Execute Sac2 Xi Yi Scanrange Pixels Pixeltime ActualY scantype k transferdatawave End The collected data for each line is displayed by DataForDisplay in both the for ward and backward images as well as in the on line line display Function DataForDisplay Wave CtsDisplayF CtsDisplayB transferdatawave LineCtsDisplay CtsdisplayF k transferdatawave p XiF q 2 1 CtsdisplayB k transferdatawave XfB p q 2 1 LineCtsdisplay CtsdisplayB p k AvCounts sum LineCtsDisplay 0 GoodPixels 1 GoodPixels AvCPS AvCounts 1000000 Pixeltime DoUpdate End The function Scan is the main function for image scanning and it coordinates all the previously presented functions Function Scan Start ButtonControl String Start SetDataFolder root Variable paramctrl XiScale XfScale YiScale YfScale Wave fit_LineCtsDisplay 174 Set up control and data acq
168. ing global variables to the values calculated by FitLinearRange And the function RestoreDis play Values gives the user the possibility to recover the uncorrected parameters Function FitLinearRange Wave Xdisplay Xmondisplay t2display Corr_Xmondisplay W_coef Variable a b Yup Ydwn i j difference Pixels_aux RemoveFromGraph Z fit_Corr_XmondisplayF RemoveFromGraph Z fit_Corr_XmondisplayB XiF 0 4 Pixels XfF Pixels 1 Cursor P A Corr_Xmondisplay XiF Cursor P B Corr_Xmondisplay XfF CurveFit Q N line Xdisplay 0 Pixels 1 a bx CurveFit Q N H 01 line Corr_Xmondisplay XiF XfF a bx a W coef 0 b W_coef 1 Make O N Pixels 1 20 fit_Corr_XmondisplayF fit_Corr_XmondisplayF a b p AppendToGraph fit_Corr_XmondisplayF vs t2display XiB 1 4 Pixels XfB 2 Pixels 1 Cursor P A Corr_Xmondisplay XiB Cursor P B Corr_Xmondisplay XfB CurveFit Q N line Xdisplay Pixels 2 Pixels 1 a bx CurveFit Q N H 01 line Corr_Xmondisplay XiB XfB a W _coef 0 b W_coef 1 A 2 Igor routines 167 Make O N Pixels fit_Corr_XmondisplayB fit_Corr_XmondisplayB a b p Pixels Duplicate O R Pixels 2 Pixels 1 t2display t2displayB AppendToGraph fit_Corr_XmondisplayB vs t2displayB j 0 Do differencee Corr Xmondisplay j fit_Corr_XmondisplayF j j 1 While difference gt Pixelsize XiF j 1 j 0 Do difference fit_Corr_XmondisplayB j Corr_Xmondisplay j Pixels j 1 While difference gt Pixelsize XiB j 1 Pixels If Corr_Xmondisplay XiF
169. ion in order to be able to discard them and still keep the desired scan range Third the new linear regions of the monitor signal are found and the for ward and backward images are constructed only with those pixels Like this the forward images are constructed with the pixels between the points A and B in figure 2 6 and the backward images similarly with the pixels between the points C and D The so made images delimited in figure 2 6 by the vertical white lines would not be blurred anymore at one edge Yet they are still shifted with respect to each other The explanation for this extra shift is that the capacitive monitor signal is electronically delayed The only way to account for this is via an empirical calibra tion parameter It is observed that faster scanning leads to a larger shift between forward and backward images To quantify the shift several images of the same region of a sample produced at different scanning speeds are necessary Then the 2 2 Description of the home built confocal microscope 15 a 3004 2nd order polynomial fit Measured values 150 0 x lt j 150 300 5 10 15 20 Scanning speed m s b x pixel 0 50 100 150 200 250 300 Corrected signal 5 x Linear fit S 5 8 Q a x 0 100 200 300 400 500 600 700 Time ms 5 5 4 3 3 E gt 3 2 2 1 1 0 0 0 1 2 3 5 0 1 2 3 um um Figure 2 7 Scanning correction a Position shift of an arbitrar
170. isplayF 0 CtsdisplayB 0 Endif SetScale x XiScale XfScale um CtsdisplayF SetScale y YiScale YfScale um CtsdisplayF SetScale x XiScale XfScale um CtsdisplayB SetScale y YiScale YfScale um CtsdisplayB Make O I U N 2 Pixels GoodPixels 6 transferdatawave transferdatawave 0 Make O I U N GoodPixels LineCtsDisplay LineCtsDisplay 0 fit_LineCtsDisplay 0 SetScale x XiScale XfScale um LineCtsDisplay SetScale x XiScale XfScale um fit_LineCtsDisplay ControlUpdate A W LineScan k 0 ActualY Yi Pixelsize 2 OpenShutter Do ActualY ActualY Pixelsize ExecuteSAC2ForScan DataForDisplay k 1 DoUpdate While k lt GoodPixels CloseShutter Beep End 176 Set up control and data acquisition software Confocal Line The Confocal Line Procedure groups the routines for line scanning The Igor function ExecuteSAC2ForLine calls the XOP SAC2 with the required parameters pragma rtGlobals 0 Function ExecuteSAC2ForLine k 0 Execute Sac2 Xi Yi Scanrange Pixels Pixeltime ActualY scantype k transferdatawaveL End The collected data for each line is displayed on the on line line display by DataForDisplay Function DataForLineDisplay graphname String graphname Wave LineCtsDisplay transferdatawave transferdatawaveL If empstr graphname SurfaceScanF 0 LineCtsDisplay transferdatawaveL p XiF g 2 1 Endif If cmpstr graphname SurfaceScanB 0
171. itation rate depends on many parameters such as the illumination intensity the sample geometry the molecules absorption cross section and absorption dipole orientation From equations 5 5 and 5 2 it can be seen that Kopp _ Ta 5 25 Ton logy Tor ae and since T can be calculated up to a constant value via the method presented in 3 2 it is possible to obtain information about 23 from the distributions of the ratio koff Ion off The simulation was performed with the theoretical values of T total and P tar multiplied by the proportionality factor obtained for the data without gold 0 26 see previous footnote 104 Single molecule fluorescence dynamics Figure 5 13 shows the experimental distributions of kop Lon Ioff for the cases with and without gold Table 5 4 lists the calculated detectable decay rate DI 3 of parallel and perpendicular dipoles on the air side of the air spacer interface as well as the average for an ensemble of arbitrarily oriented molecules The distribution of ko lon Ioff is broader for the molecules near gold The reason for this is that the molecules near gold with arbitrary out of plane orientations span a larger range of rg than the molecules in the sample without gold As a consequence the range of Ion equation 5 2 is also larger for the molecules near gold By means of equation 5 25 it is possible to calculate a quantity proportional to To3 as Po oc e
172. ive adjustments of the fiber tip position in x y and z directions should be done as the light intensity coming out of the fiber is monitored Once the x y and z positions are optimized further fine adjustments of the tilting 0 and amp of the fiber can be done When using laser light in order to control the polarization of the illumination beam a A 2 and a 4 plates are placed between the laser and the fiber The angular position of the plates have to be adjusted in order to produce a polarization state in the laser beam such that after the polarization effects of the fiber the light exiting the fiber has the desired polarization Removing of the polymer cladding is facilitated by immersing the fiber tips in acetone for 2 5 minutes 2 3 Alignment 23 The microscope is equipped with polarization filter films OWIS GmbH to con trol the linear or circular polarization state of the illumination beam The polariza tion of the Xe arc lamp white light cannot be controlled before the fiber However a certain polarization state can be filtered from the illumination beam at the exit of the fiber by means of a polarizing film Focusing the gas phase laser light The gas phase lasers He Ne and Ar ion provide gaussian beams with long co herence distances In this case a standard microscope objective is sufficient to focus the beam and to obtain highly efficient coupling into the fiber In particular a 16x 0 3NA microscope objectiv
173. ixels each Extra pixels were scanned in the x direction in order to discard the pixels corresponding to the delay near the direction inversion points 4The scanning speed here mentioned is an actual speed in m s then it relates the scanning range the pixel time and the number of pixels A reasonable value for the scanning rate is for example 40 m s corresponding to an image of 10 x 10 um with 250 x 250 pixels and a collection time of 1 ms 14 The fluorescence and light scattering confocal microscope Figure 2 6 shows a complete 320 x 128 pixels fluorescence image obtained with the confocal microscope The image has a mirror symmetry because it is formed by scanning forward and backward the same region of the sample The left and top axis show the pixels of the image The left axis shows the x position of the stage in um and the bottom axis the time necessary to scan forward and backward an x line The dashed line curve is the driving signal i e the desired x position sent to the piezoelectric stage to perform a forward and backward scan along the x direction The solid line curve is the corresponding monitor signal provided by the capacitive closed loop which is meant to report the actual position of the piezoelectric stage The delay between the driving signal and the position of the stage reported by the capacitive loop can be clearly seen by comparing the driving to the monitor signals If the individual forward and backward imag
174. l fluorescence signals Comparison of the experimen tally observed fluorescence signals right note the dynamic blinking during the scanning and the theoretical fluorescence signal of an ideal fluorophore which transition dipole is parallel to the z direction left with the experiments The grey filled curves in figure 4 18 shows the profiles of three experimental fluorescence signals of single molecules obtained with the three illumination modes The images of those signals are shown in figure 4 5 and the points from which the profiles were taken are marked with white lines The FB and FL profiles correspond to the same molecule which is to be judged from spatial distribution of the fluorescence signal and the fact that presents one of the most intense signals oriented almost perfectly along the z direction The TL profile corresponds to a molecule that gave one of the strongest TL signals therefore it should be oriented parallel to the interface and close to the x direction Plotted in black together with the FB and FL experimental profiles are the profiles of the modelled signals corresponding to a fluorophore oriented along the z axis calculated for FB and FL illumination conditions respectively Together with the experimental TL profile the profile of the TL modelled signal of a fluorophore oriented along the x axis is plotted also in black The theoretical profiles were not fitted to the experiment They were all scaled by the same fac
175. l kinetic traces of different Zno 42Cdo 535e QDs on ITO under various excitation intensities P The same total time T from each kinetic trace was computed The off time length histograms are fitted with a power law which exponent m is shown in each graph for QDs on glass and on ITO coated glass substrates calculated as Ioff 6 1 where Ty rf is the total off time Two different on intensities shown in figure 6 7 b as a function of the excitation intensity were calculated as N Non Ion net a I Ton avg 7m m t T ff auvg Ton Torf on Loss 6 2 where Ton is the total on time Ion net represents the intrinsic emission capacity of the QDs and it is observed to increase with the excitation intensity Ton avg is the time averaged on intensity It represents the time average photoluminescence emission of one QD or the average emission of an ensemble of QDs and remains almost constant over the whole studied excitation intensity range This shows how the blinking process limits the ultimate 1Note that Ion net differs from the Ion defined in chapter 5 in that Ion ne does not include the background intensity Ioff 6 3 QD kinetic traces 119 a b 14x10 Glass 120x10 di 2 eg ITO ENO 12 4 Simulated 100 4 ji pense Pi 10 e 804 ITO awg e i 84 604 ga 7 al En 40 Pe i 20 4 we a T T T T T z T T T T 0 0 0 5 1 0 1 5 2 0 0 0 0 5 1 0 1 5 2 0 P kWicm
176. l pe 5 26 isa Ioff r avg Then taking the experimental average value of kopp Lon Ioff and the calcu lated average value of ee it can be calculated that the chromophores have in the sample without gold an average T a3 0 043 x 1 34 0 058 and in the sample with gold 93 x 0 052 x 1 09 0 057 Here it should be noted that the average value of 1 used in this calculation is an isotropic average that considers all orientations equally probable As explained above this is not the case in the experiments with gold because the molecules parallel to the interface are hardly detectable Nev ertheless under this conditions no effect of the gold film on the ISC rate 93 is observed No Au 7 Au lt X gt 0 0043 c 0 0028 na 6 7 E 8 4 8 2 0 4 8 12x10 4 8 12x10 koff lon loff koff lon loff Figure 5 13 Distribution of kop Lon Ioff obtained via the autocorrelation analysis of the kinetic traces from molecules in the samples with and without the gold film The average and standard deviation of each distribution are shown in the graphs 3This rate was called in chapters 3 and 4 5 4 Influence of a nearby thin Au film on the electronic transition ratk 5 Table 5 4 Calculated normalized radiative de excitation rate of a ee Ree L avg parallel and perpendicular dipole without Au 1 08 1 87 1 34 emitted into the collection solid with Au 0 29 2 72 1 09 angle of the microscope o
177. lass 1 503 which leads to a maximum focusing angle of 68 6 The rear lens of the objective has a radius of 4 45 mm therefore the radius of the effective focal sphere is 4 78 mm 12 The fluorescence and light scattering confocal microscope 2 2 2 Scanning To form an image with the home built confocal system the sample is stepwise mechanically scanned over the focus of the microscope objective as the detected photons are counted at each step pixel of the image In comparison to optical scanning mechanical scanning offers quality at the expense of speed 15 Further more having no moving optical elements makes it easier to adapt new elements in the optical path and also makes the alignment more stable Initial position Scanning control parameters D 4 8333383323 7 a a 3323333337 Scan range Initial position um ail ae Re he a ad Scan range um x a Pixel time us or us 2 Number of pixels Figure 2 5 Scanning process The parameters used to control the scanning are listed on the right Four parameters are used to control the scanning the initial position the scan ning range the number of pixels the same for both dimensions and the pixel time counting time at each pixel The scanning process depicted in figure 2 5 is as follows The sample is first moved to the initial position Then it is moved step wise forward and backward along the arbitrarily called x direction arrows 1 and
178. layer polyelectrolyte spacer following the method described in 4 2 1 The samples without gold were prepared with a 2 5 bilayer polyelectrolyte film deposited onto 3 aminopropyltriethoxy silane 3 APTES Aldrich functionalized glass substrates The functionalization of the glass substrates was accomplished by the following steps Cleaning with surfactant Hellmanex II Hellma GmbH and water Treatment with a 1 1 5 mixture of HzO2 NH3 H2O at 80 C for 30 minutes Silanes self assembly by immersing the substrates for one hour in a 0 1 M Milli Q water solution of 3 APTES Rinsing with Milli Q water and drying with nitrogen Annealing for one hour at 120 C The so prepared substrates contain free amino groups which in water render 100 Single molecule fluorescence dynamics H 6 Chromophores a yee Air eio z l Sample A i B e a p Microscope p objective Dichroic mirror Au film Sample B Fluorescence light Figure 5 11 Schematic of the experimental configuration Fluorescent dye molecules with ar bitrary orientations defined by 8 and of their transition dipoles u are placed at the interface between a polyelectrolyte film and air In sample A the polyelectrolyte film is deposited directly on a glass substrate In sample B the polyelectrolyte film acts as a dielectric spacer between the chromophores and a 44nm gold film the surface positively charged Then a 2 5 bilayer polyelectrolyte film was de p
179. ld c Due to the inversion symmetry of the z component of the source field with respect to the z axis in any incidence plane the field generated along the z axis have zero z component To write the field along the arbitrary in plane direction y a on the original xy coordinates system the contributions of Be and Ee on the x and y axis of the original coordinates system have to be added E p p a zp EZ cosa Ee sina E cos a EY sin a 3 35 E p zp EZ sina a cosa E7 EY sin a cos a l Next it necessary to show that E is zero both along the y and z axis First following the same reasoning that lead to the nullity of Ep and EP it can be seen that EY has to be zero as well Second in a similar way as depicted in figure 3 5 c for any plane of incidence the p component of the source field focused through the one half of the objective lens generates a field along the z axis opposite to the one generated by the p component of the source field focused through the other half of the objective Then P is zero and as a consequence it is possible to obtain the 4This is basically the same projection that was done to obtain Eg from Ey in equation 3 23 54 Single molecule fluorescence through a layered system projection on any other radial direction by E p p a zp E7 cosa 3 36 Summarizing the field in any point p p of a plane parallel to the focal plane z zp can be calculated by means of
180. ld be used The data is saved in Igor binary format 25 together with the adjacent information entered in the other fields of the Save Data division of the control panel Comments Wavelength Power etc In addition the scanning parameters Pizeltime Xi Yi Scan Range Pixels are stored as well as the necessary information to reconstruct the forward and backward images from the data Software to extract the information from this type of data was programmed as well but it is not discussed in this dissertation 2 4 3 Time correlated measurements Two kinds of time correlated measurements are described in this section Life time imaging and fluorescence vs time traces Both are performed with the help of the TCSPC module see 2 2 3 Details about the operation of the TCSPC module can be found in its user manual 19 In these kinds of measurement data needs to be collected simultaneously by the AD DA converter and the TCSPC unit There fore it is necessary to divide the detector signal simply by a BNC T connector and send it to both the AD DA converter and TCSPC unit see sections 2 2 3 and 2 2 4 for the input signal requirements of the AD DA and TCSPC Lifetime imaging Lifetime imaging consists of recording a fluorescence image in which also the fluorescence excited state lifetime is stored That means that each pixel of the image should contain information about the position the number of photons counted at that position an
181. le small graph and table in figure 4 10 From the boundary conditions of the electric field equations 4 3 the electromag netic decay rates for the parallel dipole are the same on both sides of the interface Instead for a perpendicular dipole all the electromagnetic decay rates on the spacer 72 Single molecule fluorescence through a thin gold film Parallel dipole 90 rem w 3 Ti a Lew 9 5 3 air L w i F um E see g aA ER 5 12 f f obj r g A lt gm F of i o 2 f Q ne N 7 0 8 Ea T 3 E 0 4 So 0 0 P 3 eee Herrer prre rrn 0 200 400 600 800 1000 0 20 40 60 80 100 Spacer thickness nm Spacer thickness nm 6 Perpendicular dipole py G 0 8 o TeRi f a 6 S 4 Tair s sia Se z v 2 Ihr y S 4 D 3 dahhh oaoot tooo o 3 Q E 32 2 6 1 zZ 0 T r 0 i r 0 200 400 600 800 1000 0 20 40 60 80 100 Spacer thickness nm Spacer thickness nm Figure 4 9 Electromagnetic decay rates Electromagnetic decay rates normalized to the total emission of a free dipole in air for a parallel top and for a perpendicular bottom dipole on the air side of the air spacer interface of the samples emitting at A 670nm On the right detail of the region of spacer thickness between 0 and 100 nm the detectable fraction of the radiative decay Tob is plotted as well side are decreased by a factor Espacer Eair The value of Paet for a molecule with arbitrary
182. lled and a complete mathematical modelling is possible The first studies were performed with fluorophores placed at a controlled separation distance from a gold film The in fluence of the locally enhanced surface plasmon electromagnetic field on molecular fluorescence was investigated on a single molecule level First a theoretical model was set up to calculate the fluorescence signal of a single molecule in a plane layered system including the electric field distribution in the focus of the SCOM and the emission rates of a chromophore chapter 3 Second the excitation and emission of single molecule fluorescence through a thin gold film was investigated experimentally and modelled chapter 4 Third the influence of the nearby gold film on the electronic transition rates responsible of the fluorescence process was studied chapter 5 If fluorescent markers are being considered nanometer size colloidal semicon ducting crystallites also known as quantum dots QD cannot be ignored Since the middle 70s the QDs have provoked a tremendous fundamental and technical inter est Owing to their size dependent photoluminescence which is tunable across the complete visible spectrum the QDs find application as light emitting devices lasers and biological labels However the emission process in semiconducting QDs involves very complicated processes and the emitting state remains controversial Recently the advent of QD studies on a single dot level
183. ls_display If Check Values 0 Abort Scan aborted Parameters out of range Endif CalculateParametersForAdwin DoWindow B PointKinetic DoWindow F SurfaceScanF DoWindow F SurfaceScanB DoWindow F LineScan Yi Yi PixelsToCorrectInX Pixelsize Correction using previous calibration parameters Scanrange Scanrange PixelsToDiscard Pixelsize Correction using prev calibration parameters Pixels round Scanrange Pixelsize Make O I U N 2 Pixels 2 Pixels 5 transferdatawaveL transferdatawaveL 0 Make O I U N Pixels LineCtsDisplay LineCtsDisplay 0 fit_LineCtsDisplay 0 SetScale x Yi_um Yf_um um LineCtsDisplay SetScale x Yi_um Yf_um um fit_LineCtsDisplay DoUpdate ActualY Xi Scantype 2 Init A 2 Igor routines 179 OpenShutter Do ExecuteSAC2ForLine DataForLineDisplay graphname While 1 End Confocal Kinetic This procedure groups the necessary routines to accomplish fixed position mea surements In particular fluorescence vs time traces The main function is MainK inetic which via the ExecuteKinetic and the Kinetic XOP drives the piezo stage to the measurement position and collects the data The data is displayed in intervals fixed by the global variable Kinetic Duration pragma rtGlobals 0 Function ExecuteKinetic Execute Kinetic Xi Yi PixeltimeforKinetic Kineticduration Points transferkineticwave End Function MainKinetic StartKinetic ButtonControl String StartKinetic
184. m shown in figure 5 6 The values of kon and kops are obtained from the exponential fits 5 2 3 Comparison Table 5 1 presents the results obtained from the analysis of the data of shown in figure 5 2 via the autocorrelation and trace histogram methods Both methods provide similar but not equal results The trace histogram method retrieves a value of Ion slightly lower than the real one and a value of Topp slightly higher than the real one This is a consequence of the mixed bins explained above some of the on photons are computed in mixed bins that do not reach the threshold to be counted as on bins Monte Carlo simulated data In order to know which analysis is more reliable both methods were tested with Monte Carlo simulations A three level system like the one described in section 5 1 was programmed in Igor 21 The input parameters are the transition rates Ti2 Poi P23 131 the background count rate Ioff and the total time The Monte Carlo simulation generates a set of TCSPC mac t photon detection times 96 Single molecule fluorescence dynamics ihe Method Parameter lon 1 s loff 1 s ton s toff s Interphoton times Histogram correlation Table 5 1 Results of the analysis of the TC SPC data shown in figure 5 2 The autocor relation and trace histogram methods provide similar but not exactly the same results kon 1 s koff 1 s Toff ms Ton ms that can be then analyzed wi
185. m indicating a size of approximately 20nm 132 a b 0 25 0 8 t A 533 nm 0 20 A 520 nm 5 06 T 0 154 ot a J 0 4 a 0 05 4 0 24 0 00 4 500 600 700nm 500 600 700nm A A Figure 7 4 Light scattering spectra of individual gold particles 7 3 Light scattering of individual C shaped gold nanoparticles Metallic nano structures of more complex shape are more interesting than spher ical nano particles because they can present several surface plasmon resonances and stronger field localization effects C shaped nano structures are supposed to present at least two resonances and a strongly enhanced field localization within the open gap Far field investigations should allow to characterize the resonances of these particles and get some understanding of the local field enhancement 7 3 1 Experimental Substrate preparation Thin glass cover slips N 1 Menzel Gl ser were cleaned by immersing them in freshly prepared piranha solution 7 3 concentrated sulfuric acid Aldrich GmbH 30 hydrogen peroxide Aldrich GmbH for 1 hour and then sonicated for 15 min utes The slides were rinsed copiously with Milli Q water and dried under a stream of nitrogen The cover slips were used immediately after this process 7 3 Light scattering of individual C shaped gold nanoparticles 141 Fabrication of the C shaped nano structures The fabrication of the C shaped nanoparticles consists of four steps deposition of polystyr
186. maller and five dash line circles were placed on top of each image on the same position corresponding to five detected signals with TL illumination By comparing the FB to the TL images it can be seen that very weak signals in the FB image 4 3 Single molecule fluorescence images through a thin gold film 65 30 200 140 20 80 10 20 48 50 52 54 56 um 44 43 120 30 42 100 100 E 80 2b 80 41 ee mam 60 20 u 60 d 40 15 40 20 nut 20 46 47 48 49 50 46 47 48 49 50 46 47 48 49 50 um um um Figure 4 5 Different modes of illumination Typical images upper row 8 x 8um and lower row 5 x 5um of the samples with a spacer of 4 bilayers 24nm obtained with full beam FB left transmitted light TL center and forbidden light FL right illumination Profiles along the white vertical lines are compared to the theoretical signals in figure 4 18 correspond to the round TL signals Those FB signals are so weak that can be easily covered by the much stronger characteristic patterns and very careful observation is necessary not to confuse them with the background Even though the background is reduced they become weaker under FL illumination The strongest signal in both the FB and FL images has a very clear pattern and shows almost no signal in the TL image The facts that the images obtained with FB and FL illumination are so simi lar and the images obtained with TL illumination present so weak signals indica
187. mber indices 2 1 value 0 dataYbji MDSetNumericWavePoint Value transferdatawaveB indices value i i 0 while i lt Pixels indices 0 i row number indices 1 k column number indices 2 2 value 0 dataCtsb i MDSet Numeric WavePoint Value transferdatawaveB indices value i i 0 while i lt Pixels A 3 C routines 185 indices 0 i row number indices 1 k column number indices 2 3 value 0 datat1b i MDSetNumericWavePoint Value transferdatawaveB indices value i i 0 while i lt Pixels indices 0 i row number indices 1 k column number indices 2 4 value 0 datat2b i MDSetNumericWavePoint Value transferdatawaveB indices value itt DataTransferLine waveHndl transferdatawaveL long Pixels int i long DataCts 10000 long indices 3 dimensions of the wave double value 1 dimensions of the value Get1Data 3 DataCts 1 Pixels i 0 while i lt Pixels indices 0 i row number indices 1 0 column number indices 2 2 value 0 DataCts i MDSetNumericWavePoint Value transferdatawaveL indices value itt static void Scan long Xi long Yi long Scanrange long Pixels long Pixeltime long Y long scantype long k waveHndl transferdatawaveF waveHndl transferdatawaveB int check SetPar 1 Xi Xi gt PAR_1 SetPar 2 Yi Yi gt PAR_2 SetPar 3 Scanrange Scanrange gt PAR_3 SetPar 4
188. mes 130 6 18 Photo induced lifetime of the on state vs excitation intensity 131 7 1 Size distribution and SEM of the gold particles 2 137 7 2 Light scattering of individual Au nanoparticles 138 7 3 Light scattering imaging 04948 ae an enden 139 7 4 Light scattering spectra of individual gold particles 140 7 5 Light scattering of C shaped gold nanoparticles 142 Abbreviations AD DA APD a B p 0 Y CCD E FB FL FWHM NA analog digital digital analog avalanche photo diode angles charge coupled device electric field dielectric constant full beam forbidden light full width at half maximum transition rate intensity transition rate wavevector wavelength transition dipole numerical apperture refractive index 194 Abbreviations PMT photomultiplier tube QD quantum dot QE fluorescence quantum efficiency r position R radius linear correlation coefficient Fiat radius of the objective rear lens p radius in polar coordinates SBR signal to background ratio SCOM scanning confocal optical microscope SERS surface enhanced Raman scattering SM single molecule SPR surface plasmon resonance TCSPC time correlated single photon counting TEM00 transverse electromagnetic mode 00 t time TMA transfer matrix algorithm TL transmitted light TTL transistor transistor logic VPH volume phase holographic T characteristic time x y Zz Cartesian coor
189. mited to 3 level single molecule fluorescence blinking In contrast the trace histogram method is more versatile and can be applied to a other kinds of blinking process 5 3 Experimental The experimental scheme is shown in figure 5 11 Single fluorescent dye molecules DiIC1 5 Molecular probes are deposited at the interface between a layer by layer polyelectrolyte film and air Two kinds of samples were prepared In the ones the polyelectrolyte film was deposited directly on a glass substrate sample A in figure 5 11 In the others the polyelectrolyte film was used to place the chromophores at a controlled distance from a thin gold film sample B in figure 5 11 exactly as described in chapter 4 Like this the chromophores environment is the same in both samples and the 98 Single molecule fluorescence dynamics a b 10 2 1 o Simulated data Exponential fit 29 2 10 a Simulated data ro OD 16 Exponential fit 10 m 10 1 10 2 we Be oe 4o6 2 ar T La p LA 10 10 10 10 0 0 05 1 0 15 2 0 T ms Inter photon time ms c 2 O 40 4 a 7 LATTA T cn man PR ij o bey IMUM Yi Way I Iina Y Ha H mtoa a i lglg Pi a 3 4 0 4 2 44 46 48 5 0 0 5000 Time s Cts d 100 4 100 Simulated data m Simulated data Experimental data 10 Exponential fit 10 2 2 oO 14 1 0 1 4 0 1 T T T T T _ xb tF tT f 6S OD 5 0 5 amp 6 2 23 3 3 20 40
190. molecule in front of a mirror Ber Bunsenges Phys Chem vol 72 no 2 p 329 1968 K Drexhage Monomolecular layers and light Sci Am vol 222 no 3 p 108 1970 H Kuhn Classical aspects of energy transfer in molecular systmes J Chem Phys vol 53 no 1 pp 101 108 1970 W L Barnes Fluorescence near interfaces the role of photonic mode den sity Journal of Modern Optics vol 45 no 4 pp 661 699 1998 K Vasilev W Knoll and M Kreiter Fluorescence intensities of chro mophores in front of a thin metal film J Chem Phys vol 120 no 7 pp 3439 3445 2004 K Vasilev F D Stefani V Jacobsen W Knoll and M Kreiter Reduced photobleaching of chromophores close to a metal surface J Chem Phys 2004 X S Xie and R C Dunn Probing single molecule dynamics vol 265 no 5170 pp 361 364 1994 Science 198 BIBLIOGRAPHY 49 50 51 52 53 54 59 56 57 58 59 62 63 64 X S Xie and J K Trautman Optical studies of single molecules at room temperature Annu Rev Phys Chem vol 49 pp 441 480 1998 H Yokota K Saito and T Yanagida Single molecule imaging of fluores cently labeled proteins on metal by surface plasmons in aqueous solution Phys Rev Lett vol 80 no 20 pp 4606 4609 1998 J Enderlein Single molecule fluorescence near a metal layer
191. mples as a function of the spacer 5 Average taken over 20 randomly chosen signals for the 4 and 3 bilayer spacer samples and over 5 signals for the 2 bilayer spacer samples 68 Single molecule fluorescence through a thin gold film thickness The excitation was in that case performed by a surface plasmon field at the gold spacer interface generated by a laser beam of constant intensity incident at the angle of surface plasmon resonance Then in order to compare the single molecule quenching behavior observed in the present experiments to the ensemble one measured by Vasilev et al the maximum single molecule signals detected on each type of sample Im were identified and normalized by In I arg 4 1 ep where Jag is the average background intensity tp is the counting time per pixel and P is the excitation power Figures 4 7 a b and c show the profiles of representative maximum signals detected on the samples with 4 3 and 2 bilayer spacers marked with a white line on the corresponding images of figure 4 6 These maximum values were normalized as explained above equation 4 1 and plotted as a function of the spacer thickness together with ensemble results of Vasilev et al in figure 4 7 d Both quenching trends are in excellent agreement providing further evidence that the excitation of the single dye molecules is accomplished via surface plasmons at the metal polyelectrolyte interface 4 4 Modelling the experim
192. n 3 7 does not depend on u and the normalized total electromagnetic decay rates for the parallel and perpendicular dipoles can be calculated as ER 3 T kp krz TP Kp k kh Ts kp Tiotal 2 k3 Im if 9 j 9 Kr dk 0 3 11 Sell Tiotal L or 2K Im if Kn rp k dk 0 gt Finally from equations 3 7 and 3 9 the total electromagnetic decay rate of a molecule with and arbitrary out of plane orientation amp is calculated as Pita sin Tital l T cos Tiotal 1 3 12 The case of a dipole on the inner side of the first interface So far the total electromagnetic decay rate was calculated for a dipole in front of the layered system It is simple to extend this result to the case of a dipole infinitely close to the interface from the inner side of the first layer Due to the fact that the materials are non magnetic the only change to the Poynting computation of the power flux arises from the discontinuity of the perpendicular component of the electric field Then Itall takes the same value on both sides of the interface and Deru needs to be scaled by a factor er e where e is the dielectric constant of the first layer 46 Single molecule fluorescence through a layered system 3 2 3 Non radiative electromagnetic de excitation rate The non radiative electromagnetic decay rate T can be calculated as the dif ference of the total electromagnetic de excitation rate I and the total radi
193. nation the scattering spectra of individual nanoparticles were recorded In order to account for the spectral behavior of the system Xe arc lamp lenses mirrors and CCD camera the spectrum of the light reflected at the glass air interface was recorded from an empty area of the samples and used as reference The scattering spectral response of a single particle S A was obtained by Spart A Sref A Srep A S A 7 1 where Syar A is the intensity of scattered light by a given particle and S r A is the reflected intensity both for the wavelength A Spectra from a number of particles were acquired All the spectra present a single peak corresponding to the excitation of the surface plasmon resonance SPR in the quasi spherical particles 132 Different SPR wavelengths were found with the trend that longer SPR wavelengths correspond to stronger signals i e larger particles No further studies were carried out because the aim of the present ex periments is to provide a proof of principle for the light scattering measurements and to characterize the resolution of the home built microscope Figure 7 4 a shows a typical scattering spectra of the colloidal gold particles showing a resonance at 140 Light scattering from single metallic nano structures 533nm Figure 7 4 b shows a spectra of one of the weakest detected signals This spectra is expected to correspond to the smallest detectable particle and presents a SPR at 520n
194. nd indicates that the QDs posses several emitting states supporting the results of Schlegel 113 and Fisher 114 Except for the on intensity and the weak memory effect the Monte Carlo model can reproduce all the experimental characteristics of the blinking The only in put parameter that was varied in order to reproduced the experimental data is the characteristic lifetime of the on times rTpr Three experimental photoinduced char acteristics of the blinking decrease of the on time fraction increase of the cycles per second and decrease of the maximum on time could be reproduced quantitatively by Tpr This lifetime of the on state shows a completely different behavior as func tion of the excitation intensity for the QDs on glass and on ITO coated glass In the case of QDs on glass Tp can be satisfactorily fitted by a power law with exponent 1 providing evidence of a one photon photo induced shortening of the on times In the case of ITO Tp can be fitted satisfactorily with a single exponential dependence on the excitation intensity but no theoretical support was found for this dependence Nevertheless given the so different dependencies of Tp on the excitation intensity for the QDs on glass and on ITO it is most likely that two completely different photo induced mechanisms are involved Chapter 7 Confocal microscopy measurements of light scattering from single metallic nano structures This chapter is meant to demonstrate the pe
195. neral a smaller tmin produces shorter on and off times with higher probability and a consequent higher number of cycles per second Several 7 kinetic traces were simulated with different values of tmin and all the other input parameters fixed and then analyzed with the trace histogram method No changes in the detected number of cycles per second were noticed for tmin below 10 ms The reason is that as tmin becomes smaller the generated shorter on and off times become undetectable and their influence vanishes Figure 6 13 illustrates this effect The simulated on and off periods are shown together with the on and off periods observed in the kinetic trace histogram The tmin input was set to 1075 ms in all simulations This value is small enough and no noticeable influence on the number of cycles per second is expected Lower values were not used because they increased greatly the computational time Tpr photo induced on state lifetime Tp is the only parameter that can introduce a difference between the on and off times distributions A shorter Tp makes long on times less probable with a consequent higher number of cycles per second and shorter total on time fraction 6 4 Modelling the QDs blinking 127 Counts 0 2 4 SI g 10 Period length s ee Counts 6 90 6 95 7 00 7 05 7 10 _ 7 Period length s Counts 6 960 6 962 6 964 6 966 6 968 6 970 Period length s Figure 6 13 Simulated
196. ng spectrograph was adapted to the set up The spectrograph consists of a volume phase holographic VPH transmission grating HVG 590 and HVG 690 Kaiser Optical Systems Inc a photo objective 85 mm f1 4 Nikon Inc and a high quantum efficiency charge coupled device CCD camera SensiCam QE PCO Imaging GmbH Table 2 3 presents the relevant technical information for these components The collimated detected beam is directed to the VPH grating either by a silvered mirror or by a beam splitter Light is diffracted by the VPH grating at angles according to classical diffraction and the energy distribution is governed by the Bragg condition The photo objective focuses the spectrally dispersed light into a line of the CCD sensor of the camera The spectral ranges covered by the HVG 590 and HVG 690 gratings are from 390 to 790 nm and from 500 to 800nm respectively The size of the detectable spectral range depends on the angular dispersion of the grating the focal length 2 2 Description of the home built confocal microscope 19 VPH grating Photo CCD camera Objective Kaiser Optical Nikon inc PCO Imaging GmbH Systems inc ii SensiCam QE Sensor 1376x1040 pixels pixel size 6 45x6 45 um dynamic range 12 bit non linearity lt 1 Angular dispersion Focal distance 35 mm nm Maximum HVG 590 0 0657 diaphragm Ainm QE apperture 400 43 500 60 600 55 700 40 800 21 HVG 650
197. nge_prev 0 Pixeltime_prev 0 Pixels_prev 0 Pixeltime_prev 0 Init Execute ADWinBoot2 ADWinBoot2 is an XOP written in C End Function Init Variable result t1 t2 Execute ADWinlnit2 Xi Yi ADWinlnit2 is an XOP written in C t1 DateTime Do t2 DateTime While t2 t1 lt 14 Return result End Procedure Calibration This procedure realizes all the necessary corrections for the delay of the piezo electric drivers ScanCalib scans via the Function ExecuteScanCalib and the Scan Calibration XOP three lines with the set parameters and records the capacitive monitor signal from the amplifier Then it calls CalibrationDisplay which calculates the corrected position and displays it together with the driving signal and the ca pacitive monitor signal The corrected position is calculated from the monitor signal A 2 Igor routines 165 with the calibration parameters The function CheckValues belongs to Procedure Confocal Imaging A 2 pragma rtGlobals 0 Function ScanCalib Variable paramctrl j Offset Xi_um Xi_um_display Yi_um Yi_um_display Pixels Pixels_display Scanrange_um Scanrange_um_display If Check Values 0 Abort Scan aborted Parameters out of range Endif DoUpdate CalculateParametersForADWin Init DoWindow B PointKinetic DoWindow F Calibration Make O I U N 2 Pixels Xdisplay Make O I U N 2 Pixels Xmondisplay Make O I U N 2 Pixels t2display Make O I U N 2 Pixels Corr Xmondi
198. ngle molecule fluorescence through a layered system a b Ere c ty y ty Bon 7 Pi N j Y E y ia So I E 1 sl y ie r x li E 5 J E r x a a S E b E Figure 3 4 Symmetry considerations in the focal plane a Due to the mirror symmetry of the problem with respect to the x axis no y components of the fields are generated along the x axis b Due to the inversion symmetry with respect to the y axis imposed by the source field direction no y components are generated along the y axis c Then the fields generated along the x or y axis have only x in plane component given point ofthe y axis Then as shown in figure 3 4 b half a period later both the source field and the generated y component have the same amplitude and point in the opposite direction On the other hand it is possible to choose an x axis pointing in the opposite direction This would invert the direction of the source field but not the generated y component Then there are two valid solutions for the generated y component produced by the same source field but pointing in opposite directions Therefore the y component of the electric field along the y axis needs to be null As a consequence a source field polarized along the x axis generates in plane z zp fields along the x and y axis that only have x components as depicted in figure 3 4 c To describe the in plane field distribution i
199. ntegration over the azimuth angle in figure 3 1 3 2 The emission 43 Figure 3 2 General multi layer system showing the relevant param eters for the transfer matrix algo rithm TMA and the coordinate system It is therefore sufficient to calculate the two particular cases After the introduction of the normalization factors equation 3 4 and 2r integration around the azimuthal angle the emission rates to a region of space in medium m for a parallel and perpendicular dipole can be calculated by the following integrals kmax 3k k ms ee Bell Eyl 2 dk 0 Ang Km Kp 3 6 3kp Fk ng ee E k dk 1er tae IB dk 0 where ko 27 A and km Nm ko are the wave vectors of the emitted radiation in vacuum and in medium m respectively nz is the refractive index of the reference medium i e the rates are normalized to the total emission of a dipole in a medium of refractive index nr E k are expressions for the Cartesian components of the electric field at the dipole position as a function of k They are obtained via the TMA for plane waves incident from medium m and depend on the thicknesses and dielectric constants of the layers and the wavelength of the emitted radiation The integration limits define the region of space of interest To compute the far field emission to a complete semi space the integration should be accomplished between zero and the wavevector of the emitted radiation
200. ntrary on the spacer side of the interface the molecules with a greater in plane component parallel to the polarization direction present the strongest signals with a spatial distribution that is the one characteristic of the intensity of the x component of the electric field spacer BPI 90 67 57 PL i 45 22 5 0 gt 4 8 0 Max a Signal ai 7 6 3 0 er 10 7 72 it r A D 0 gt 0 225 45 675 90 0 225 45 675 9 f Figure 4 15 Modelled FL illumination fluorescence signals Calculated fluorescence signal of an ideal QE 1 fluorophore under FL illumination on the air left and on the spacer right sides of the air spacer interface The orientation of the transition dipole of the fluorophore is defined by od and 8 according to figure 4 1 At the bottom the maxima of the signals for 3 0 as a function of od 8Up to the small polarization dependence of the transmission coefficient of the dichroic mirror and the filters 80 Single molecule fluorescence through a thin gold film The modelled fluorescence signals corresponding to forbidden light illumination are presented in figure 4 15 The differences with the results for FB illumination are due to the fact that under FL illumination the x component of the electric field is reduced markedly more than the other two see section 4 4 3 On the air side the molecules with greater out of plane component of their transition dipole present
201. ntrol panel This action in addition to resetting the local processor of the AD DA converter clears up the local memory of the AD DA converter and loads all necessary routines II The PC user interface file is called SAC_v7 pxp 30 The fluorescence and light scattering confocal microscope 2 4 1 Sample requirements and mounting Due to the short working distance of the high NA microscope objective the sam ples should be prepared on thin microscope coverslips 0 13 0 16mm The sample should be placed on the sample holder see figure 2 15 fixed with the magnetic film and then moved with care to the piezoelectric stage The region of interest of the sample should be positioned on top of the front lens of the microscope objective with sufficient immersion oil between the objective and the sample A A Magnetic film Sample Sample holder Immersion Oil Piezoelectric stage Figure 2 15 Sample holder The sample holder consists of a thin square plate 75x75 mm of steel with a perforation 20mm in the center The sample should be placed on top of the sample holder and fixed with a magnetic film Then the complete set should be positioned on the piezoelectric stage After placing a new sample on the microscope the focus should be adjusted It is advisable to start by placing the focus of the objective on the glass air or sample surrounding medium interface as described in section 2 3 4 2 4 2 Imaging The fir
202. nty in the detection efficiency 7 in equation 3 1 and that the electric field in Peze is calculated up to a constant this theoretical signal is proportional to the experimental one For a given orientation of the transition 3 5 Conclusions 55 dipole of the fluorophore Peze defines the spatial distribution of the fluorescence sig nal intensity and T ge scales the signals according to the detectable radiation emitted from a molecule with the given out of plane orientation 3 5 Conclusions A method to model the fluorescence signal of single molecules was set up by considering the chromophores as oscillating dipoles interacting classically with the electromagnetic field placed at the outer interface of a layered system and with arbitrary orientations The method is of general applicability to any non magnetic plane layered system and considers the excitation and the emission separately Calculations performed with this model can be useful to design and compare experimental configurations i e different illumination modes materials and thicknesses of the layered system and detection schemes and to determine the three dimensional orientation of single chromophores under different conditions Chapter 4 Excitation and detection of single molecule fluorescence through a thin gold film In this chapter it is demonstrated that fluorescence of single molecules in the nanometric vicinity of a thin gold film can be detected and studie
203. o 4 pp 331 342 1980 W Lukosz Optical systems with resolving powers exceeding classical limit J Opt Soc Am vol 56 no 11 p 1463 1966 W Lukosz Optical systems with resolving powers exceeding classical limit 2 J Opt Soc Am vol 57 no 7 p 932 1967 J C Knight and P S J Russell Photonic crystal fibers New ways to guide light Science vol 282 pp 1476 1478 1998 P Russell Photonic crystal fibers Science vol 199 pp 358 392 2003 M Born and E Wolf Principles of optics Oxford England Pergamon Press 6th ed 1980 196 BIBLIOGRAPHY 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 T Wilson Techniques of optical scanning microscopy J Phys E Sci In strum vol 22 pp 532 517 1989 T Wilson Theory and practice of scanning optical microscopy London Aca demic Press 1st ed 1984 D V O Connor and D Phillips Time Correlated Single Photon Counting London Academic Press 1984 L M Bollinger and G E Thomas Measurement of the time dependence of scintilation intensity by delayed coincidence method The Review of Scientific Instruments vol 32 no 9 pp 1044 1150 1961 Becker und Hickl GmbH SPCM 630 User manual http www becker hickl de Last accessed in Jan 2004 J ger GmbH ADwin AD DA converters http www adwin de Last ac cessed in Jan
204. o compare different detection schemes and experimental geometries in order to optimize the collection of fluores cence photons Both methods were combined to calculate the SCOM fluorescence signal of a chromophore in a general layered system The fluorescence excitation and emission of single molecules through a thin gold film was investigated experimentally and modelled Chromophores were placed at a controlled separation distance from the gold film by means of a polyelectrolyte spacer layer It was demonstrated that due to the mediation of surface plasmons single molecule fluorescence near a thin gold film can be excited and detected with an epi illumination scheme through the film Single molecule fluorescence as close as 15 nm to the gold film was studied in this manner In comparison to the detection from the air side the surface plasmon medi ated detection through the gold film resulted more efficient In comparison to the 146 Summary case without gold the number of detectable photons emitted by perpendicular flu orophores was found to be enhanced by a factor of 1 4 The latter is a consequence of the increased photo stability produced by the surface plasmon enhanced total de excitation rate An excellent quantitative agreement between the experimental and modelled single molecule fluorescence signals was found if the molecules were considered to behave optically as on the air side of the interface The fluorescence dynamics
205. objective NA 1 4 the refrac tive index of the focusing medium Ngiass 1 503 and on the beam and blocking disc diameter as explained in section 2 2 1 For FB and FL illumination modes the beam diameter was adjusted to completely illuminate the rear lens of the objective which has a diameter of 8 9mm In the case of TL illumination the beam diameter was reduced to 5mm The blocking disc used in the FL illumination mode had a diameter of 5mm Then the ranges of angles of incidence 0 corresponding to each illumination mode are Full beam FB illumination 0 lt 0 lt 68 6 Forbidden light FL illumination 31 6 lt 0 lt 68 6 Transmitted light TL illumination 0 lt 6 lt 31 6 As their name indicate the aim of the different modes of illumination is to discrim inate the contributions of transmitted and forbidden light for the excitation of the dye molecules Figure 4 2 aids to explain this point The curves are the theoretical reflectivity coefficient of the sample system with a spacer thickness of Onm bare gold and 24nm 4 PAH PSS bilayers for a A 633nm p polarized plane wave as a function of the angle of incidence from the glass side Below the 6 axis the angular ranges corresponding to the different modes of illumination are indicated For low incidence angles the reflectivity of the samples is relatively high because the gold film acts as a good mirror Above the critical angle of total internal re
206. of the samples were recorded 5 4 Influence of a nearby thin Au film on the electronic transition raii and analyzed with the trace histogram and autocorrelation methods Since the au tocorrelation method is more accurate only the results obtained with this method were taken into account Nevertheless only the kinetic traces that showed a three level behavior with both methods were considered i e the autocorrelation and the histograms of the length of the on and of periods were satisfactorily fitted by a single exponential 5 4 Influence of a nearby thin Au film on the elec tronic transition rates Figure 5 12 shows the distributions of the singlet decay rate P inverse of the excited state lifetime kon and koff obtained from 33 molecules in the sample without the gold film and 42 molecules in the sample with the gold film The influence of the gold presence on each rate is analyzed next X gt 92 wo 4 o 1 6273 n j 3 44 NO 3 2 g 3 Au O 4 2 oO 0 iceman ae TS r 0 3 06 09 1 2x10 1 2 3 4 5x10 0 1 2 3 4x10 T gt 1s koff 1 s kon 1 s 8 4 2 z3 s3 2 Au 5 5 5 8 8 4 8 2 4 I 03 06 09 1 2x10 1 2 3 4 5x10 0 1 2 3 4x10 I 54 1 8 koff 1 8 kon 1 s Figure 5 12 Distributions of values of I z kon and korj obtained via the autocorrelation analysis of the kinetic traces from molecules in the samples with bottom and without top the gold film The averages and standard deviations
207. on de excitation cycles before irreversible photo bleaching occurs 136 Light scattering from single metallic nano structures In order to successfully implement the plasmon resonances in practical applica tions it is necessary to tailor the resonance frequencies and the field localization in a controlled manner To achieve this the investigation of nanoparticles of different shapes and materials is of fundamental importance In recent years a great progress has been made in the fabrication of metallic nano structures of different shapes 123 127 Theoretical investigations have been also developed 128 129 but accurate solutions for the local fields of plasmon reso nant particles of arbitrary shape remains however a theoretical challenge For this reason experimental investigations are very important The aim of this chapter is to demonstrate the capabilities of the home built SCOM to perform light scattering measurements with particular application to the study of surface plasmon resonances of metallic nanoparticles For this purpose two different gold nanoparticles were studied spherical colloidal nanoparticles and C shaped nano structures 7 2 Light scattering of individual colloidal gold nanoparticles Colloidal gold nano particles have been deeply studied and can be readily avail able thus they provide an ideal test sample to characterize the performance of the home build SCOM for light scattering measurements For this rea
208. on of 12ps up to a rate of 8 MHz Furthermore the TCSPC unit also records for every detected photon the detection time from the beginning of the experiment Mac t in figure 2 9 with a resolution of 50 ns The TCSPC module needs only two input signals The pulse train signal from the laser which is typically a sinusoidal signal with a frequency equal to the repetition rate of the laser pulses and the single photon detector signal In both cases the signals need to have an amplitude in between 50mV and 1V Therefore both the APD and the PMT signals need to be treated with specific electronics before feeding the TCSPC The APD signal is processed by a router HRT 82 Becker und Hickl GmbH that transforms the positive TTL pulses provided by the APD into the required negative pulses The PMT signal is already negative but too weak for the TCSPC therefore it needs to be amplified with a 26dB broadband 5kHz 1 6GHz amplifier HFAC 26 Becker und Hickl GmbH The ultimate time resolution for the determination of fluorescence lifetimes is usually limited by one of the following two factors First the width of the laser 6In fact the TCSPC unit takes advantage of the low intensity conditions It determines the time from the photon detection until the next excitation pulse and calculates the mic t time later by difference In this manner the TCSPC unit performs the time calculation only when a photon is detected and not for all the pulses
209. on experimental parameters such as temperature and excitation intensity The on and off periods present non trivial power law statistics and can render the QD dot dark for periods of hundreds of seconds Surface passivation by a suitable inorganic or organic layer was reported to improve the quantum efficiency and photo bleaching stability of QDs 103 106 but the same blinking behavior was universally observed in capped and uncapped QDs of different kinds 107 110 This blinking represents an intrinsic limitation for the QDs applications because it limits their quantum efficiency and brightness and 6 1 Brief Introduction and current status 111 makes their use in molecular tracking very complicated Several models have been suggested to account for the dynamics of the blinking but most of them fail to explain the power law statistics such as the quantum jump model the activated kinetics model the static barrier tunnelling model all reviewed in 110 Two models can account for the on and off times statistics First Shimizu et al suggested a trap state that randomly wanders in the energy space and even tually shifts into resonance with the excited state 109 At each crossing of the trap and excited states the QD can switch from on to off or viceversa Second Kuno et al suggested a model in which an excited electron or hole can tunnel to a trap state rendering the QD dark until tunnelling back to the excited state allows the radia tiv
210. or different values of the threshold sweeping from zero to the average of the on population Jo bw The optimum threshold for the given bin width is found if less than 1 bin is wrongly classified In case no threshold fulfils that condition the bin width is increased and the procedure is repeated successively In this man ner the minimum bin width that allows to place a threshold between the Poisson distributed on and off populations that fulfils the condition Wrong bins lt 1 is found Then a time histogram of the data is constructed with the optimum bin width and the on and off periods are identified via the optimum threshold with the highest reliability and time resolution possible Figure 5 6 shows the optimum bin width histogram of the data shown in figure 5 2 and the optimum threshold horizontal line On the bottom of figure 5 6 a smaller region of the histogram is shown corresponding to the photons detected between the fourth and fifth second The onSoff fluorescence fluctuations can be clearly seen On the bottom right of figure 5 6 the histogram of the photons per bin intensity obtained in the kinetic trace histogram solid grey is shown The on intensity and the off intensity can be clearly distinguished The curves are Poisson distributions defined by equation 5 23 scaled to the corresponding maximum of the on or off intensity In the small graph only the region from 0 to 1000 counts in the horizontal scale is plotted in
211. orded as a function of time with the TCSPC module of the set up The pho ton detection mac t times contain the information about photoluminescence fluctu ations see section 2 2 4 Kinetic traces were recorded under different excitation powers which were measured before the recording of each kinetic trace with a photo power meter S120 Thor Labs Inc 114 Photoluminescence blinking of Zno 42Cdo ss5e nano crystals 6 3 QD kinetic traces Photoluminescence vs time traces kinetic traces of single Zng 42Cdo 535e QDs were analyzed with the trace histogram method described in section 5 2 2 6 3 1 General characteristics Although at first sight they might seem similar to the kinetic traces of single dye molecules see chapter 5 the kinetic traces obtained from QDs on both glass and ITO coated glass substrates hide a much more complex process Five repre sentative kinetic traces from QDs are plotted in figure 6 3 left together with the histograms of the photons per bin i e intensity center and the distributions of the length of the on and off periods right N On 2 2 10 aN OF 5 E Q Q 01 4 0 001 4 T T T T L 0 001 0 1 J On 100 on 2 2 c c 1 3 2 amp 8 0 01 o 20 40 60 80 100 120 140 0 10x10 00101 1 m On 2 g 1 m Off c C 4 gt 14 f oO O 001 0 0001 001 01 10 On 2 2 10 4 on 5 5 4 Q Q 014 O 5 i 0 001 001 01 1 2 2 100 of Cc c 4 5 31 N O oO 0
212. order to observe better the on intensity distribution and the region between the on and off intensities It can be seen that in the region between the on and ofl intensities the experimental intensity deviates from the Poisson distributions to higher values This corresponds to mized bins in which a 5 2 Kinetic traces analysis methods 95 part of an on time and a part of an off time occurred This can be due to on times shorter than the bin width or to bins computed at the beginning or the end of an on time The latter being the most probable because the average on time Ton is much larger than the bin width see below a b 100 Experimental data F Experimetal data 4 Exponential fit 24 Exponential fit 2 10 10 d 2 4 2 24 6 k O 1 a 4 24 2 IT S S S S tie T T T 1 5 10 15 20 25 30 20 40 60 80 100 On period length ms Off period length ms Figure 5 7 Histograms of the length of the on periods a and of the off periods b obtained from the analysis of the kinetic trace histogram of figure 5 6 The optimized kinetic trace histogram can be analyzed to find the on and off bins and determine the length of the on and off periods Histograms of the length of the on and off periods show exponential decays with exponents equal to kon and koff respectively equations 5 6 Figure 5 7 shows the histograms of the length of the on a and off periods b obtained from the analysis of the kinetic trace histogra
213. orientation of its transition dipole is given by equation 3 14 For parallel and perpendicular dipoles the fraction takes the same value on any side of the interface but for intermediate orientations it does not As an example figure 4 11 a shows I y as a function of the out of plane orien tation of the transition dipole for molecules on either side of the spacer 24nm air interface It can be seen that in general the fluorescence emission collected by the objective is higher for molecules on the air side of the interface In the experimental scheme for a 4 bilayer 24nm spacer almost 50 of the de excitation rate of a perpendicular dipole on the air side of the interface corresponds to detectable fluo rescence For the case of a parallel dipole this fraction is slightly higher than 25 And for molecules which transition dipoles are oriented with an angle above 30 with respect to the interface plane Taget gt 40 4 4 Modelling the experimental scheme 73 1 0 aaa Parallel Perpendicular Aussee af nm 0 8 0 6 Idet 0 4 s s Bonn ynnentten ns u pi f 2 ee i i 0 2 N e 205 X AAs Fs M 0 200 400 600 800 1000 Spacer thickness nm Figure 4 10 Detectable fraction of the fluorescence emission as function of the spacer thickness Taer for a parallel and a perpendicular dipole on the air side of the spacer air interface as a function of the spacer thickne
214. orrelation of single molecule fluo rescence Phys Rev Lett vol 70 no 23 pp 3584 3587 1993 R M Dickson A B Cubitt R Y Tsien and W E Moerner On off blink ing and switching behavior of single molecules fo green fluorescent protein Nature vol 388 pp 355 358 1997 K D Weston P J Carson H Metiu and S K Buratto Room temperature characteristics of single dye molecules adsorbed on a glass surface J Chem Phys vol 109 no 17 pp 7474 7485 1998 M Orrit J Bernard R Brown and B Lounis Progress in Optics vol XXV ch Optical spectrocopy of single molecules in solids pp 61 144 Amsterdam The Netherlands Elsevier 1996 S Y Kilin T M Maevskaya A P Nizovtsev V N Shatokhin P R Berman C von Borczyskowski J Wrachtrup and L Fleury Stochastic dynamics of a single impurity molecule from the viewpoint of continuous measurement theory Phys Rev A vol 57 pp 1400 1411 1998 A N Kapanidisa and S Weiss Fluorescent probes and bioconjugation chemistries for single molecule fluorescence analysis of biomolecules J Chem Phys vol 117 pp 10953 10964 2002 S Weiss Measuring conformational dynamics of biomolecules by single molecule fluorescence spectroscopy Nat Struct Biol vol 7 pp 724 729 2000 B Lounis and W E Moerner Single photons on demand from a single molecule at room temperature Nature vol 407 pp 491 493 2
215. oscope 13 equipped with a local processor and local memory that enable to perform real time measurements with a resolution of 25 ns regardless of the speed of the controlling computer The data acquisition also is accomplished by the same AD DA converter For each pixel of an x scanned line forward and backward the position the counts and the starting and final counting times are stored in the local memory of the AD DA converter Then the sample is moved in the y direction and before starting the next x line all the information is transferred to the controlling computer Like this the counting times are measured with a resolution of 25 ns The so collected data has the information corresponding to two images of the same region one forward scanned and the other backward scanned The individual images need to be extracted from the complete data Forward and backward images from the collected data While scanning at a reasonable speed 0 5 2lines s the piezoelectric stage response in the x direction is delayed with respect to the driving signal In the y direction due to the time necessary for the data transfer the piezoelectric stage has sufficient time to reach the set position x pixel 0 50 100 150 200 250 300 y pixel x position um T T 0 100 200 300 400 500 600 700 800 Time ms Figure 2 6 Scanning delay Complete data collected to construct backward and forward images of 5 x 5 um 128 x 128 p
216. osed of rates which are normalized to the total emission of a free dipole in a reference medium T Therefore in order to account for a QE lt 1 r an intrinsic non radiative decay rate normalized as well to the total radiative de excitation rate needs to be added to r in the denominator of Paet as shown in 3 3 The excitation 47 equation 3 2 It is easy to express this intrinsic normalized non radiative decay rate I as a function of the QE gt Hr Oe i 3 16 Ti is a function exclusively of QE and is therefore not expected to depend on external parameters such as the properties of the layered system or the position and orientation of the fluorophore with respect to the interfaces Then the detectable fraction of the emitted radiation for an arbitrarily oriented fluorophore with T 4 0 writes Tolo sin Q Tov cos Papi ul T sin A Pry Pia Bee Tae 3 17 3 3 The excitation Far from saturation the excitation rate of a fluorophore positioned at r and with a transition dipole moment u is proportional to the square modulus of the projection of the electric field at that position E r along the dipole direction Dene t amp u E r 3 18 As a consequence in scanning confocal optical microscopy SCOM the spa tial distribution of the fluorescence signal of a single molecule reflects the spatial distribution of the electric field intensity along th
217. osited by the layer by layer method starting with PSS as described in the pre vious chapter section 4 2 1 Finally single fluorescent dye molecules 1 1 3 3 3 3 hexamethylindicarbocyanine iodide DiIC1 5 Molecular probes maximum excita tion at Acre 638 nm maximum emission at Aem 670 nm were deposited electro statically on a negatively charged surface terminated with PSS by immersing the samples for one minute in a 10 1 M Milli Q water solution of the dyes Then the samples were rinsed with Milli Q water and dried with a stream of nitrogen 5 3 2 Measurement The measurements were performed with the SCOM set up described in chapter 2 and consisted of imaging a region of the samples identifying a single molecule and recording the fluorescence emission of that molecule as a function of time The avalanche photo diode APD was used as single photon detector and the red pulsed diode laser was used for excitation at 633nm with a repetition rate of 50 MHz The total instrumental response was limited by the APD response to Ins The fluores cence emission was recorded as a function of time with the TCSPC module see section 2 2 4 The mac t times contain information about fluorescence fluctuations in a time scale longer than several hundreds of us such as the triplet blinking The mic t times allow to obtain the excited state lifetime with a resolution limited by the APD response Kinetic traces from around 35 molecules in each
218. otoluminescence blinking of Zno 42Cdo ss5e nano crystals 6 5 Conclusions Photoluminescence blinking of Zno 42Cdo 535e quantum dots was measured for the first time The length of the on and off periods present the power law probability density that was universally observed in other QDs of different compositions and structures This further supports the idea of a common blinking mechanism for all semiconducting QDs The blinking behavior of QDs on glass insulator and on ITO coated glass semi conductor substrates was studied under different excitation intensities The probability density of the off times shows a power law that spans up to 8 decades in probability density and 4 decades independently of the excitation intensity and the nature of the substrate In contrast the probability of the on times shows the same power law for short times but long on times become less probable as the excitation intensity increases The decrease in the number of long on times and as a consequence of the on time fraction is much more pronounced for the QDs on glass Then it can be concluded that there are two different processes governing the lifetime of the on and the off times at least for the long times At very low excitation intensities the QDs spend approximately half the time in the on state and the other half in the off state As the excitation intensity increases the fraction of time that the QDs spend in the on state reduces This is in agreem
219. out how to place the microscope coverslip laser illumination with sufficiently high intensity should be used and all filters should be removed Like this the light reflected at the glass air interface can be easily seen By moving the objective up and down the focus of the objective should be placed at the glass air interface so that the reflected beam consists of light from the focal spot At this point as the reflected light comes from the focus it is collected and collimated by the microscope objective Therefore it is possible to control that the focus of the objective is at the glass air interface by controlling the collimation of the reflected beam There are two ways of controlling the collimation of the reflected beam One is by using the collimation tester shear plate 23 as described in section 2 3 2 The other more practical method is to direct the reflected beam to the microscope Huygenian ocular there is a flippable mirror installed for this purpose and then project it onto a screen The ocular amplifies any deviation from the collimation and the adjustment can be accomplished by eye see figure 2 12 The collimated reflected beam should be directed to the detection path with a sil vered mirror There it is focused with an achromatic lens into the confocal pin hole and collimated again with a second achromatic lens The positions of both lenses lOTMPORTANT do not look into the eyepiece if no filter is blocking the las
220. pole at 24nm distance from the gold film is close to unity unaffected by the gold film see figure 4 9 the QE reduces their detectable emission fraction almost linearly In contrast the reduction for a perpendicular dipole is damped 74 Single molecule fluorescence through a thin gold film a b 0 454 040 a N x o 3 R 0 35 gt onthe air side x U 0 30 gt on the spacer side B hen ND EEE EEE QE 0 1 a nimo 0 04 Ze SEE PT BERET I gt CEB 0 01 0 20 40 60 80 0 20 40 60 80 L 6 I oli p I Figure 4 11 Detectable fraction of the fluorescence emission as function of the orientation a Taet for an ideal fluorophore QE 1 on either side of the spacer 24nm air interface as a function of out of plane orientation of the transition dipole b Effect of a smaller QE for the fluorophore on the air side due to the strongest enhancement of their total electromagnetic decay rate As an example it can be seen from figure 4 11 b that a QE 0 5 reduces the detectable fraction of a parallel dipole by 48 and the one of a perpendicular dipole by less than 15 Comparison to the detection in a sample without the gold film The de excitation rates were calculated as well for molecules on the air side of the spacer air interface of samples without the gold film Table 4 2 presents the results for the samples with and without the gold film By comparing the rates to the complete glass semi space Ugiass an
221. provided by the source is coupled into a single mode optical fiber to obtain a point like light source equivalent to the illumination pin hole described in the previous section Light coming out of the fiber is collimated to form the illumination beam With the help of a dichroic mirror or a beam splitter the illumination beam is directed to a high numerical aperture microscope objective that focuses it onto the sample Light coming from the sample is collected by the same objective and directed by a mirror to the single photon detectors and or the spectrograph In order to acquire an image or to study different regions the sample can be mechanically scanned over the focused beam by means of a piezoelectric xyz stage Furthermore although not included in figure 2 3 the microscope is equipped with a second single mode fiber aligned in the confocal system that can be coupled to any of the light sources This allows to perform simultaneous illumination with two wavelengths In the following for simplicity and because both fibers are equivalent only one fiber is considered The set up is very flexible Different modes of illumination and detection can be implemented to perform a variety of studies such as fluorescence Raman and light scattering measurements Following all the components of the set up and their functions are described in detail Section 2 2 1 describes the illumination section 2 2 2 the scanning and section 2 2 3 the detection The
222. provides TTL pulses that can be directly counted by the AD DA converter The PMT instead provides very weak fast and negative pulses The easiest way to transform this signal into a TTL pulse is by means of an Oscilloscope 18 The fluorescence and light scattering confocal microscope Sinal Dark Active Quantum Linear hot z counts aren officie acy Output response Manufacturer hra CPS diameter pulses up to model mm Ainm x10 cps Positive TTL Perkin Elmer Optoelectronics inc 2 5 V 50 9 SPCM AQR13 FWHM 30 ns 3 mV 50 Q Hamamatsu Photonics KK FWHM 2 ns H7422 40 Table 2 2 Single photon counting detectors APD avalanche photo diode PMT photo multiplier tube CPS counts per second A wavelength FWHM full width at half maximum All the quantities were corroborated in the laboratory except the ones marked with which are taken from the manufacturer specifications The 50Q input channel of a 300 MHz oscilloscope Tektronix 2465 is supplied with the PMT signal Once the triggering is adjusted the monitor output of the os cilloscope provides one TTL for each PMT pulse The power supply the Peltier cooling and the gain of the PMT are computer controlled with a controller and a software specifically designed for this detector DCC 100 Becker und Hickl GmbH The transmission grating spectrograph In order to measure single molecule fluorescence and light scattering spectra a transmission grati
223. ptured the attention of chemists physi cists and engineers from around the world It is not hard to see the motivation for such interest The effect was large completely unexpected difficult to understand and of enormous practical utility if it could be understood and exploited The investigation of the SERS still continues and the understanding of the phe nomenon has increased considerably Nowadays it is accepted that the SERS effect is caused by greatly enhanced electromagnetic fields generated by surface plasmon resonances SPR in certain hot spots of the rough substrate At this point it is important to note that such strong and localized electromagnetic fields do not only find applications in the SERS For example they can also be employed in optical tweezers and to modify radiative rates in a variety of processes such as molecular fluorescence Recent advances in microscopy have made it possible to use single metallic particles as SERS substrates and to obtain the Raman spectra of single molecules adsorbed on them 5 Almost simultaneously the field that is today called nanotechnology developed and thanks to that it is possible to produce an enormous variety of structures in the sub micrometer scale In particular metallic structures with nanometer size and different shapes can be manufactured and their surface plasmon resonances can 2 Introduction be tailored These days it is possible to think of a structure composed of me
224. rary Chapter 3 Modelling the fluorescence signal of a single molecule through a layered system A theoretical method is set up to model the scanning confocal microscopy fluo rescence signal of single molecules through a general non magnetic layered system The method considers the fluorescent dye molecule as a point oscillating dipole of fixed orientation interacting classically with the electromagnetic field First in this chapter the problem is described and the relevant parameters are introduced Then expressions for the de excitation rates and for the spatial distribution of the exci tation field and are derived for single molecules of arbitrary orientation Finally the results obtained for the excitation and the emission are combined to give an expression for the single molecule fluorescence signal 3 1 Description of the problem The fluorescence signal of a single molecule positioned at the first interface of a layered system is to be modelled Excitation and detection of fluorescence is consid ered to be accomplished with the high numerical aperture microscope objective of an episcopic illumination scanning confocal optical microscope as the one described in chapter 2 The detectable fluorescence intensity emitted by a single molecule in the absence of triplet blinking is given by L Ism leze N Pen 3 1 where Peze is the fluorescence excitation rate I and Inar are respectively the total 1The concept
225. reduces together with the number of detected molecules For the 4 bilayer spacer samples the average signal to background ratio is found to be 6 for the 3 bilayer spacer samples it is 3 5 and for the 2 bilayer spacer samples it is 1 3 Fluorescence signals were much rarer in the samples with 2 bilayer spacers and no fluorescence signals were detected in the samples with 1 bilayer spacer a 200 b 70 4 bilayer 3 bilayer n 160 60 120 2 ey 3 Er 8 80 oO 30 40 20 10 0 0 05 1 0 15 2 0 0 0 05 1 0 15 2 0 um um d c 35 E E Single molecules Ka au 2 bilayer 2 Ensemble 30 ETS 2 N ER g 25 20 S 5 no i 3 20 zE 6 Eg 15 Se 3 i 0 0 05 1 0 15 2 0 5 10 15 20 um Spacer thickness nm Figure 4 7 Quenching behavior a b and c Profiles corresponding to the vertical white lines on the images of the samples with a 4 3 and 2 bilayer spacer of figure 4 6 These are representative maximum signals for each kind of sample d Normalized maximum single molecule fluorescence signals I equation 4 1 from a b and c and the ensemble fluorescence intensity 46 as a function of the spacer thickness The present results on a single molecule level present the same quenching behav ior observed on an ensemble of fluorophores on the same sample system carried out by Vasilev et al 46 Vasilev et al measured the fluorescence intensity emitted by an ensemble of molecules to the glass side of the sa
226. rformance of the home built confo cal microscope for light scattering measurements First colloidal gold nanoparticles are used as a test sample to proof the light scattering measurement principle and to characterize the size of the smallest detectable scatterer Then light scattering images and spectra are acquired from two C shaped gold nanoparticles Excitation with two polarization states is employed to try to identify different resonances 7 1 Introduction The interest in metallic nanoparticles is nowadays driven by the phenomenon of surface enhancement of optical spectroscopies where the signals of molecules ad sorbed on rough metal surfaces or metallic nanoparticles can be strongly enhanced In the case surface enhanced Raman scattering SERS the enhancement factor can reach values above 10 allowing for single molecules detection 5 120 121 Even though the adsorption of the molecule on the metal can participate chemically in the Raman enhancement 122 it is believed that the greatly enhanced local elec tromagnetic fields generated upon excitation of surface plasmons in the metallic nanoparticles are responsible for the major contribution to the SERS effect Fluo rescence can be also enhanced by the strong surface plasmon fields Alterations of the local photonic mode density can lead to an higher quantum efficiency and shorter lifetime 45 62 of a chromophore and in molecular fluorescence can lead to a higher number of excitati
227. robing the evanescent field of propagating plasmon surface polaritons by fluorescence and raman spectroscopies J Chem Phys vol 98 no 12 pp 10093 10095 1993 J R Lakowicz Radiative decay engineering Biophysical and biomedical applications Analytical Biochemistry vol 298 pp 1 24 2001 B N J Persson and N D Lang Electron hole pair of excited states near a metal Phys Rev B vol 26 no 10 pp 5409 5415 1982 B N J Persson and E Zaremba Electron hole pair production at metal surfaces Phys Rev B vol 31 no 4 pp 1863 1872 1985 BIBLIOGRAPHY 199 65 66 67 68 71 72 73 76 77 P W Atkins Physical Chemistry W H Freeman amp Company 6th ed 1997 K D Weston P J Carson J A DeAro and S K Buratto Single molecule detection fluorescence of surface bound species in vacuum Chemical Physics Letters vol 308 pp 58 64 1999 J Widengren U Mets and R Rigler Fluorescence correlation spectroscopy of triplet states in solution A theoretical and experimental study J Phys Chem vol 99 pp 13368 13379 1995 J Bernard H Talon and M Orrit Photon bunching in the fluorescence from single molecules A probe for intersystem crossing J Chem Phys vol 98 no 2 pp 850 859 1993 A Zumbusch L Fleury R Brown J Bernard and M Orrit Probing indi vidual two level systems in a polymer by c
228. rough The size of the diaphragm can be at first adjusted by eye Once a scattering sample is placed on the microscope the size of the diaphragm can be optimized in order to get the best signal to background 2 4 Operation In this section instructions about how to perform the basic operations with the microscope are given First it is explained what the sample requirements are and how the sample should be mounted Second all the operations related to imaging are explained Finally the instructions to perform time correlated measurements and to record spectra are given J 3 EX Sean Parameters paramers Calbration ADWin gem Pixel time PB i Fresse ee i Shutter Open Xi um ain Xi um Pixels 2 D Pixels Measurement Kinetic Line Scan Surface Scan 1 1 Pixel time ms Gauss Ft M 2 2 Nitrogen EZ Figure 2 14 Control panel of the PC user interface used to operate the home built confocal microscope Many of the computer controlled functions of the microscope explained below can be managed from the PC user interface The control panel for this interface is shown in figure 2 14 For further details about the microscope control and data acquisition software please refer to appendix A 3 Before performing any operation from the PC user interface software it is nec essary to boot the local processor of the AD DA converter by pressing the button called Boot in the ADWin division of the co
229. rticles do not all have the exact same 7 4 Conclusions 143 orientation As it can be observed in the SEM image of figure 7 5 for the particle on the left the two polarization states represent two different situations but for the particle on the right both polarizations constitute almost identical situations The previous experimental results represent a nice example to show how im portant is to investigate these particles individually In an ensemble measurement the polarization effects would be hidden by averaging the response of particles with slightly different orientations 7 4 Conclusions The constructed home built SCOM can perform light scattering measurements with high sensitivity Individual scatterers with a sub wavelength size can be easily detected and studied For example colloidal gold particles can be detected down to a size of approximately 20 nm The light scattering behavior of C shaped gold nano structures was studied for two polarization states the electric field parallel and perpendicular to the sym metry axis of the C The two polarization states correspond in some particles to two different resonances but some other particles showed the same spectra for both polarizations This controversy which shows the importance of performing studies on a single particle level might be explained by taking into account that the parti cles have slightly different orientations Further measurements and complementary experimen
230. s Scan Acquisition Control ADBasic ScanInit T94 ADBPrLoad D Doktorarbeit Control CompleteRoutines Scan Acquisition Control ADBasic LP HP PointKinetic T95 ADBPrLoad D Doktorarbeit Control CompleteRoutines Scan Acquisition Control ADBasic HP KineticPixel T96 ADBPrLoad D Doktorarbeit Control CompleteRoutines Scan Acquisition Control ADBasic LP HP LineScan_Calibration T97 break main ioRecHandle This is the initial entry point at which the host application calls XOP The message sent by the host must be INIT main does any necessary initialization and then sets the XOPEntry field of the ioRecHandle to the address to be called for future messages HOST_IMPORT void main IORecHandle ioRecHandle ifdef XOP_GLOBALS_ARE_A4_BASED ifdef _MWERKS__ For CodeWarrior 68K XOPs SetCurrentA4 Set up correct A4 This allows globals to work SendXOPA4Tolgor ioRecHandle GetA4 And communicate it to Igor endif endif 182 Set up control and data acquisition software XOPlnit ioRecHandle do standard XOP initialization SetXOPEntry XOPEntry set entry point for future calls SetXOPResult OL All structures are 2 byte aligned if GENERATINGPOWERPC pragma options align reset endif ifdef WINDOWS_ pragma pack endif ADWinlnit sends the piezoelectric stage to a given position via the ScanInit BAS A 1 ADWinlnit
231. s become detectable To account for this and try to detect a real effect the number of cycles per second were divided by the SBR 6 9 b The number of cycles per second shows a practically linear increase with the SBR except for last or the last two points for the QDs on glass which deviates to higher values On time fraction The fraction of time that the QDs spend in the on state was computed as Ton Ton Toff and is shown in figure 6 10 as a function of the excitation intensity P At very low intensities the QDs spend approximately half of the time in the on state and as the excitation intensity increases the on time fraction reduces The reduction is much stronger for the QDs on glass substrates 6 3 QD kinetic traces 121 0 6 4 O Glass O ITO 0 5 i Simulated 0 4 Figure 6 10 Experimental and simulated 03 1 on time fraction for the QDs on glass and ITO coated glass substrates as a function of 0 24 Sal I the excitation intensity P ai T T T 0 1 a ON time fraction 0 0 0 5 1 0 1 5 2 0 P kW em This results are in qualitative agreement with the ionization model that states that a positively charged QD is a dark QD and a neutral QD is a bright QD Krauss et al 112 carried out electrostatic measurements on single QDs and showed that at room temperature and in dry air approximately 50 of the QDs were positively charged and 50 were neutral and that the percentage of positiv
232. s near gold was not completely probed The latter is due to two reasons the fact that parallel molecules are practically undetectable through the gold film and the limiting time resolution of the APD Furthermore the number of studied molecules might have been insufficient to probe such a broad distribution No influence of the gold presence was observed on the ISC rate I 93 This is 5 5 Conclusions 107 expected because the effect of the gold film is to introduce additional available elec tromagnetic modes that should not influence the spin unpairing process responsible of the non radiative transition 2 3 A noticeable 2 fold in average increase of kon T31 due the gold film was observed The gold film is able to accelerate the transition 3 1 A positive corre lation is found between Ig and kon indicating that the gold influences the radiative part of the 3 1 transition and that the transition dipole associated must have an orientation similar to the singlet transition dipole Further investigations are necessary to complete the results obtained in this chapter It would be beneficial to study a larger number molecules especially in the case with gold A faster photo detector needs to be employed in order to probe the complete Pz distribution of the molecules near gold Additional information could be obtained by correlating the transition rates to the three dimensional orientation of the individual molecules Chapter 6 P
233. s of photon detection 6 3 QD kinetic traces 117 3 100 fit m 1 68 3 1 5 Figure 6 5 Histogram of the on and off 8 times of a single kinetic trace Both the on 0 01 and off periods follow the power law but with a different maximum times 0 0001 0 01 0 1 1 10 Peiod length s times it is in practice impossible to compare them directly Therefore it is neces sary to find a few characteristic parameters that can describe the blinking From the analysis of the kinetic traces several parameters can be obtained to characterize the blinking behavior In this case on and off intensities the exponent of the off time length histogram power law the number of detected on off cycles per second and the on time fraction were chosen and are analyzed below Later in this chapter a Monte carlo procedure is employed to simulate the blink ing of QDs and to reproduce the experimental data Some of the characteristic quantities obtained from the trace histogram analysis will be used as input for the Monte Carlo simulations and some others as measure for comparison between the simulated and experimental traces Since there is no established mechanism for the photoluminescence blinking of QDs it is meaningful to try to frame the experimental results within the most com monly suggested ionization mechanism because that could lead to the design of new experiments of theoretical calculations to confirm or refute the hypothesis
234. se of a plane interface due to symmetry reasons the parallel to the interface component of the back reacted field generated by a perpendicular dipole vanishes and viceversa Therefore the total back reacted field generated by a dipole of arbitrary orientation can be written as a composition of the back reacted fields generated by the parallel and perpendicular components of the dipole Esr and Eyr respectively Ep Sin d cos 3 E Ep singsin 8 3 8 For cos where and 8 are the azimuth and polar angles that define an arbitrary direction for the transition dipole as shown in figure 3 1 Then the product Ep u for a dipole with arbitrary orientation writes Ey sind cos 8 sind cos 3 Ey Ey sindsind u singsin B Err cos cos d u sin Esri cos Eprt 3 9 The back reacted fields generated by a parallel and a perpendicular dipole can 3 2 The emission 45 be calculated by the following integrals 34 Eile k ia ete Bi up ji _ Sp ie rp k L pr rs k dk 2 2 kn 3 10 ip ff ko Bers k dk 74 Aeger Kr rp k dk 0 where kpg is the wave vector of the emitted radiation in the reference medium and kr its projection on the z direction rp and rg are the reflectivity coefficients of the layered system from the reference medium for a p and s polarized plane wave respectively They are functions of k and can be calculated with the TMA Then the term in brackets in equatio
235. son the first ex periments were performed with gold nanoparticles with an average size of 20nm which due to background limitations is near the minimum detectable size with this kind of far field instruments 7 2 1 Experimental Colloidal gold synthesis Colloidal gold particles nanoparticles were prepared by citrate reduction of gold chloride The method is well documented 130 131 and produces gold particles with a narrow size distribution The synthetic procedure is as follows in a 500ml 2 neck round bottom flask 300 ml of a 0 01 Milli Q water solution of HAuCl Aldrich GmbH are brought to boiling temperature under refluxing and stirring then 10 5ml of 1 trisodium citrate Aldrich GmbH in Milli Q water is injected rapidly After the change of color accomplished within 3 to 5 minutes the mixture is kept boiling for another 20 minutes After that the heating source is removed but the stirring is continued until the solution reaches room temperature The so prepared colloidal suspension has a conductivity of 380 20 mS cm and an initial pH of 6 0 0 3 The particle size 7 2 Light scattering of individual colloidal gold nanoparticles 137 12 10 oa 8 5 6 Figure 7 1 Size distribution 8 right of the colloidal gold par ticles obtained by analysis of the 2 SEM image left 0 T T T T T T 18 19 20 21 22 23 24 Diameter nm distribution was obtained by analyzing scanning electron microscopy SEM images of t
236. splay CompleteImage 0 Xdisplay 0 Xmondisplay 0 t2display 0 Make O I U N 2 Pixels 4 calibrationwave ActualY Yi Pixelsize 2 XminusXmon 0 RemoveFromGraph Z fit_Xmondisplay fit_Xdisplay fit_XmondisplayB RemoveFromGraph Z fit XmondisplayF PrevXdisplay j 0 Do ExecuteScanCalibration CalibrationDisplay j 1 While j lt 3 Return 1 End Function ExecuteScanCalibration Execute ScanCalibration Xi Yi Scanrange Pixels Pixeltime ActualY calibrationwave End 166 Set up control and data acquisition software Function CalibrationDisplay Variable TimeZero Offset Wave Xdisplay Xmondisplay t2display Corr_Xmondisplay Calibrationwave Xdisplay Calibrationwave p 0 Xmondisplay Calibrationwave p 1 t2display Calibrationwave p 3 TimeZero t2display O t2display t2display TimeZero Offset Xdisplay 0 Xmondisplay 0 Xmondisplay Xmondisplay Offset Corr Xmondisplay 0 Pixels 1 Xmondisplay p 10852 Pixelsize Pixeltime 299665 Pixelsize Pixeltime 2 Corr_Xmondisplay Pixels 2 Pixels 1 Xmondisplay p 10852 Pixelsize Pixeltime 299665 Pixelsize Pixeltime 2 End FitLinearRange fits the linear range of the corrected monitor signal and calcu lates how many pixels have to be added and the initial position for the measurement scan in order to have the desired number of pixels and scanning range in the linear range of the stage position The CorrectParameters function sets the scann
237. ss The small graph shows a close up of the region of spacer thicknesses used in the experiments The list displays the values of T a c for the parallel and for the perpendicular dipole for spacers from 4 to 24nm and their ratio black solid curve right axis in the small graph Comparison to the detection from the air side It is interesting to compare the present detection scheme to the detection from the air side with a high NA microscope objective For this reason the ratio T r 0 lt 0 lt 64 2 Gia Was also calculated This ratio represents the fraction of the de excitation rate of a fluorophore on the air side of the spacer air interface that corresponds to fluorescence emitted to the air side into the collection solid angle of an 0 9 NA ob jective Tair o lt 9 lt 64 2 T Gra is 0 25 for a parallel dipole and 0 07 for a perpendicular one Then in comparison to the detection from the air side the present detection scheme through the gold film is equivalent for a parallel dipole but is almost 7 times more effective for a perpendicular dipole The case of a fluorophore with T 4 0 Figure 4 11 b shows the detectable fluorescence fraction of the emission Tat of a fluorophore on the air side of the spacer 24nm air interface of the samples as a function of its out of plane orientation for different intrinsic quantum efficiencies equation 3 17 It can be seen that since the total electromagnetic decay rate of a parallel di
238. st step to make an image is to mount a sample and to focus the microscope Then from the control panel figure 2 14 the parameters for scanning an image can be set the initial position Xi and Yi in um the scanning range Scan Range in um the number of pixels per line Pirels and the counting time per pixel Pizeltime in ms or us The process of acquiring an image is initiated by pressing the button Start in the division called Surface Scan of the control panel figure 2 14 If the Calibration check box is checked a calibration procedure is performed be fore the image scanning in order to account for the delayed response of the piezoelec tric stage i e a number of lines is scanned with the set parameters and the correc tions parameters necessary to construct consistent forward and backward scanned images from the collected data are determined An overview of this calibration procedure is given in section 2 2 2 The code of the algorithm used is presented and commented in the appendix A 3 2 4 Operation 31 X Xmon F 320 732 Pixels to cor in X 16 X Xmon B 288 675 XiF 36 EXIF 193 Pxis to discard 36 XiB 242 BXfB 399 A bi AG G Bvss Figure 2 16 Calibration screen Screen shot of the user interface software during the calibra tion procedure before the start of the scanning to acquire an image The software displays the parameters of the calibration procedure that corrects the effects of the inertia of the piezoelectr
239. sted in table 4 1 In order to calculate the fields produced by the different illumination modes the integration limits for 0 in equation 3 28 were set in accordance to the range of angles of incidence of each illumination mode see section 4 2 2 Figure 4 12 shows the intensity distribution of the x y and z components of the electric field generated on the air side of the spacer air interface of the samples In order to compare the relative intensities of the components all the fields were normalized with the same factor in order to produce a FB z component with a maximum intensity of 100 The maximum intensity of each component is shown on the upper left of the each image and a line profile is shown to the right of the corresponding image The symmetry characteristics of the patterns is the same for the three modes 76 Single molecule fluorescence through a thin gold film 60 FB 20 0 00 05 10 15 20 um O an oa o8B8888 00 05 10 0 02 10 08 001 06 04 02 0 00 00 00 05 10 15 20 00 05 10 15 20 um um 00 05 10 15 20 um FL 0 o8 888 00 05 10 15 20 00 05 10 15 20 um 8 8 TL amp on ba ow Sh in 00 05 10 15 20 um Figure 4 12 Electric field distribution in the samples Theoretical calculations of the square modulus of the x y and z components of the electric field on the air side of the spacer air interface of the samples From top to bottom the results corresponding to full beam FB
240. stituto de Tecnologia Prof J Sabato Buenos Aires Scholarship granted by the Comisi n Na cional de Energia At mica Average grade 9 0 scale from 0 to 10 Awards Iron and Steel Society annual scholarship 1999 Thesis Surface plasmon resonance spectroscopy applied to the de tection and study of DNA hybridization reactions Experimental part carried out at the Max Planck Institut f r Polymerforschung under the direction of Prof Dr W Knoll Title Materials Engineer Ph D program at Johannes Gutenberg Universitat Mainz and Max Planck Intitut f r Polymerforschung
241. t In light scattering measurements instead of a dichroic mirror a 50 50 beam splitter is used The alignment procedure is the same in both cases so the following explanation for the dichroic mirror is also valid for the 50 50 beam splitter The aim of this procedure is to direct the collimated illumination beam centered and parallel to the high NA microscope objective that focuses the excitation light onto the sample To achieve this the positions of the dichroic mirror and the microscope objective should be adjusted First arrangement of the positions of both the objective and mirror can be done by eye For further fine adjustment of the positions the objective can be moved in the plane parallel to the sample and the dichroic mirror can be moved in all three coordinates and can also be tilted All movement are accomplished with micrometer precision The fine correction of the relative position of the illumination beam and mi croscope objective can be done with the help of the reflections of the illumination beam in the internal lenses of the microscope objective see figure 2 11 Relatively high laser intensity is required to easily see the reflections and the collimation lens should be displaced further from the fiber tip in order to make the beam slightly convergent A screen placed just before the objective and with a small aperture 6mm to let the convergent beam pass through allows the visualization of the beam reflections in the in
242. t t 1 Version 1 FastStop 0 AdbasicVersion 2000000 ATSRAM 0 OPT_LEVEL 1 SAVECOMPIL 0 DIM test Tpixels j k AS INTEGER DIM DATA_1 10000 AS LONG X DIM DATA_2 10000 JAS LONG Y DIM DATA_11 10000 AS LONG X DIM DATA_12 10000 AS LONG Y DIM DATA_31 10000 AS LONG Xmon DIM DATA_32 10000 AS LONG Xmon DEFINE Xi PAR_1 DEFINE Yi PAR_2 DEFINE Scanrange PAR_3 DEFINE Pixels PAR_4 DEFINE Pixeltime PAR_5 DEFINE Pixelsize PAR_6 DEFINE X PAR_7 DEFINE Y PAR_8 DEFINE j PAR_9 DEFINE terminated PAR_11 DEFINE flag PAR_55 INCLUDE C ADwin ADbasic3 Inc adwgcent inc INIT terminated 0 A 1 AD Basic routines 157 GLOBALDELAY 1 SET_MUX 0 test 0 IF test 1 THEN Xi 32768 Yi 32768 Scanrange 3277 Pixels 128 Pixeltime 5000 Y 45000 ENDIF EVENT PAR_11 0 Pixelsize Scanrange Pixels Pixeltime Pixeltime 40 usec X Xi Pixelsize 2 Y PAR_8 flag 1 j 0 CNT_ENABLE 1 CNT_CLEAR 1 DAC 2 Y DO j j 1 X X Pixelsize DATA_1 j X DATA_2 jJ Y START_PROCESS 2 DO X PAR_7 UNTIL flag 1 DATA _31 jJ ADC 1 UNTIL j Pixels DO j j START_PROCESS 2 DO X PAR_7 UNTIL flag 1 DATA_31 jJ ADC 1 DATA_1 j X DATA_2 jJ Y X X Pixelsize UNTIL j 2 Pixels X X Pixelsize END 158 Set up control and data acquisition software FINISH PAR_11 1 HP Pixel Vert BAS Proze nummer 8 Delay 1 Eventsource 0 Number of Loops 0 Prioritat 0 Version 1 FastStop 0
243. t is then sufficient to know EY and EY E r x 0 zp E7 0 EF E 0 3 32 E r 0 4 zp EY 0 BY Ey 0 To obtain the field along an arbitrary in plane direction a it is convenient to consider a primed coordinate system rotated an angle a around z in order to place the direction of interest along the x axis The source field Esre 1 0 0 can be decomposed in two components along the rotated axis as shown in figure 3 5 a Biss Bunt Esrcy Baraat cos q 3 33 Pay sing Taking into account that no field components perpendicular to the source field are 3 3 The excitation 53 generated equations 3 32 Esre and Esrcy generate the fields depicted in figure 3 5 b which can be calculated by scaling the components E and EY equations 3 29 and 3 30 with the corresponding components of the source field Then the fields generated along the x axis are ge m x gt x EF Erca EZ cosa e 3 34 Ey E Erca EY sina a b c E ie 4 x y Ere Ere y y P I f Ne gem Eo 4 E p NE rey 4 Y Z Xa _ EIN Aa EN AN x Pg x Z a PA g N fr N Borex j j x Ey N f Figure 3 5 Electric fields along an arbitrary direction a In plane fields generated by the source field in a coordinate system x y rotated a around z axis with respect to the original coordinates system xy b Electric fields generated by the x and y components of the source fie
244. tal lic nanoparticles and a chromophore or Raman scatterer in a defined geometry in order to produce an ultra effective marker or to imagine a metallic nano structure en gineered to function as a nano optical tweezers Even though such a nano structure cannot be fabricated in a controlled manner yet several research groups around the world are working on it and it should not be long until this is achieved The aim of this Ph D thesis is to settle the basis for the quantitative assessment of effects in individual such functional nano structures The first step taken was the design and construction of a scanning confocal op tical microscope SCOM that allows to measure from the same diffraction limited spot time resolved fluorescence and SERS with single molecule sensitivity and light scattering with highest resolution achievable with a far field method chapter 2 This instrument allows to investigate the surface plasmon resonances of individ ual metallic nanoparticles chapter 7 and their influence on the Raman scattering and or fluorescence processes Then a model system was sought to realize the first systematic study Surface plasmon resonances can be excited not only in metallic nanoparticles but also in planar surfaces Such a simple geometry although it provides a relatively small field enhancement represents a very convenient platform for systematic studies be cause it is easy to fabricate their geometric parameters can be contro
245. tal time T from each kinetic trace was computed The off time length histograms are fitted with a power law which exponent m is shown in each graph probability of a certain length of an on period shows the same power law behavior for short times but as the excitation intensity increases the probability of a long on period takes smaller values than the power law This gradual deviation from the power law arises from the common computation of the on times of different kinetic traces and is not observed in the on times of individual traces The individual traces are simply truncated at a maximum on time shorter than the maximum off time 109 Figure 6 5 shows an example The same experiment was conducted with the Zng 42Cdp 535e QDs deposited on ITO coated glass substrates The on and off time length histograms are shown in figure 6 6 in the same fashion as the results for glass substrates In this case the behavior is similar but the deviations of the on time length histogram from the power law are less pronounced indicating that the dependence on excitation intensity is weaker The probability of the length of an off period seems to be independent of the excitation intensity and the nature of the substrate This suggests at least for the long periods the presence of two distinct processes One responsible for the length of an on period and the other ruling the length of an off period Since the kinetic traces are composed of tens of million
246. te that light propagates most effectively through the layered system via surface plas mons the TL illumination mode cannot excite surface plasmon see figure 4 2 Excitation of the dye molecules is therefore accomplished principally by the evanes cent field of the surface plasmon which decays exponentially from the gold surface with a 1 e decay length of approximately 240nm This axial field localization is approximately two times higher than the one obtained in a confocal system when A 633 nm light is focused in a dielectric medium by a 1 4 NA microscope objective The electric field in the layered system from which the decay length was calculated was obtained via the TMA 66 Single molecule fluorescence through a thin gold film 4 3 3 Influence of the separation distance to the gold film Molecular fluorescence can be strongly suppressed by a close nearby metallic surface This effect frequently called quenching was observed in a number of en semble investigations of fluorescence near metals 42 45 46 60 The quenching is due the fact that energy transfer to the metal provides additional electromagnetic non radiative decay channels for an excited molecule that predominate at very short distances the emission rates of a fluorophore in a layered system is further treated in 3 2 Therefore for the practical use of the present detection scheme it is of great importance to know the minimum separation distance at which the sin
247. ternal lenses of the microscope objective If the beam and objective are to be concentric and parallel all the reflections should be concentric as well and their images should coincide on the screen aperture which should be concentric with the convergent illumination beam To attain this the position of the objective should be adjusted iteratively with tilting of the dichroic mirror In the experiments presented in this dissertation human hairs were used 26 The fluorescence and light scattering confocal microscope Microscope objective Inner lenses reflection 7 eflections Pr Screen H f Convergent O v beam Dichroic A a gt mirror p x y Figure 2 11 Alignment of the dichroic mirror and microscope objective The objective can be moved in the xy plane The mirror can be moved and tilted in all directions x y z 0 and When the illumination beam and the microscope objective are centered and parallel all the reflections of the beam on the inner lenses of the objective coincide in the center 2 3 4 Alignments in the detection The aim of these alignment procedures is to assure the selective detection of light coming only from the focal spot with the different detectors To proceed with the alignment it is necessary to have a visible beam of light collected by the objective from its focus To achieve this a microscope coverslip should be placed in the sample holder see section 2 4 1 for details ab
248. tes should share a common molecular geometry 5 5 Conclusions The autocorrelation and the trace histogram methods for the analysis of single molecule fluorescence blinking were introduced and their performance was compared via the analysis of Monte Carlo simulated data The trace histogram method was improved with respect to reported versions in order to systematically find the best compromise between time resolution and accuracy in distinguishing between on and off states The autocorrelation method is more demanding in terms of computing effort and reliable specially for the analysis of short traces However can be applied only to a three level blinking system The trace histogram method on the other hand is more versatile and can be employed to investigate other kinds of blinking The influence of a nearby metallic surface on the electronic transition rates was investigated via studies of single molecule fluorescence blinking and excited state lifetime The T21 koff and kon of individual dye molecules in the presence and in the absence of a nearby thin gold film were compared The gold film provides additional de excitation channels for the excited molecules which strongly depend on the chromophores orientation This is clearly observed as a broadening of the distribution of I s Comparison to theoretical calculations can explain the observed effect quantitatively in the case without gold and indicates that the distribution of chromophore
249. th the autocorrelation and trace histogram methods and the reliability with which both methods retrieve the original rates can be tested Simulation parameters T a 40000 1 s Ta 5x10 1 5 F3 3000 1 s 2 31 75 Ns 2 I pr 3500 1 8 F time 18 s koff 238 1 s Koa 75 Als 0 50 100 150 200 photons 231623 Photon number Figure 5 8 Monte Carlo simulated detection times of photons emitted by a single molecule based on a 3 level system section 5 1 The transition rates used for the simulation are listed on the right together with values of kon and koff A simulated TCSPC mac t times trace with a length of 18 seconds is shown in figure 5 8 The input parameters are listed on the right of figure 5 8 and were chosen to generate simulated data similar to the experimental one figure 5 2 that was analyzed in the previous two sections Figure 5 9 a shows the autocorrelation of the photon detection times and figure 5 9 b the inter photon times histogram Figure 5 9 c presents a histogram of the simulated kinetic trace with the optimum bin width The optimum bin width for the simulated data is very similar to the one determined for the experimental data and the threshold is the same The mized bins are observed in the simulated data as well they represent an intrinsic limitation of the trace histogram method Figure 5 3 Experimental 97 Method Simulation Interphoton Parameter Histogram correlation lon
250. the detection schemes is that light collected from the sample is directed to the detection channel via a silvered mirror OWIS GmbH figure 2 3 Then the light is focused with a 100 mm focal length achromatic lens OWIS GmbH into a 150 um confocal to the microscope objective pinhole New Focus Inc and collimated again with a second 100 mm focal length achromatic lens OWIS GmbH For imaging this collimated beam is directed and focused with another 100mm focal length achromatic lens OWIS GmbH to one of the single photon counting detectors For spectrally resolved measurements the detected collimated beam is directed to the transmission grating spectrograph Further information about the single photon detectors and the transmission grating spectrograph is given below In addition the microscope is equipped with an ocular Aziomat Plan W 10 x 25 Carl Zeiss Germany that can be used for focusing see section 2 3 4 and visual inspection of the samples Other optical elements such as polarizing beam splitters or dichroic mirrors can be installed in the detection channel according to the requirements of a particular experiment The photon counting for imaging is accomplished by a digital counter of the same AD DA converter that controls the scanning This counter only requires TTL Transistor Transistor Logic input pulses Therefore any detector can be adapted to the set up as long as its output signal is or is transformed to a TTL pulse
251. through the gold film with an epi illumination scheme as shown in figure 4 1 a He Ne laser light with a wave length A 633nm was used for excitation The illumination beam was linearly polarized and the intensities used ranged from 3 to 12kW cm Suitable dichroic mirror Notch and long pass filters were used to separate fluorescence from reflected excitation light Fluorescence micrographs were recorded by scanning the samples with three different modes of illumination as depicted in figure 4 1 First images were acquired under full beam FB illumination Second by using a blocking disc images were acquired under annular illumination corresponding to large angles of incidence in the following this scheme is called forbidden light FL illumination These intensity values are not the actual intensities at the focus They are calculated simply by dividing the full beam intensity by the area of the focal spot A given by the theoretical diffraction limit A tR R 0 61 A NA Therefore they would only represent the actual intensity at the focus of a transparent non absorbing sample 4 2 Experimental 61 Finally images were also recorded by illuminating with a reduced beam diameter corresponding to small angles of incidence in the following this scheme is called transmitted light TL illumination The range of angles of incidence corresponding to each illumination mode de pends on the numerical aperture of the microscope
252. tic traces The simulations were performed to reproduce the experimental data of the QDs on glass for the different excitation intensities P The results are presented in the same fashion as figure 6 4 6 4 Modelling the QDs blinking 129 Like this the only parameter left to introduce a difference between the on and off times and reproduce the experimental photo induced effects is Tp The value of Tpr was varied in order to find an optimum agreement between simulation and experimental data on time fraction and cycles per second was reached Figure 6 15 shows the histograms of the on and off period lengths simulated for the QDs on glass and figure 6 16 for the QDs on ITO coated glass To compare better to the experimental data a power law was fitted to the off time histograms with a fixed exponent equal to the experimental one 1000 1000 P 2 04 kWicm P 1 61 kWicm P 0 95 kWicm 100 100 100 4 m 1 63 m 1 60 m 1 60 10 4 10 4 g 104 2 2 amp 14 2 1 4 a 14 8 ard 0 1 4 01 4 0 01 4 0014 T 70s 0 014 T 108s T 124s 0 001 4 0 001 re ger 0 001 Imre re ee rer 0 01 0 1 1 10 0 1 1 10 100 0 1 1 10 100 Period length s Period length s Period length s 1000 P 0 74 kWicm 1000 P 0 18 kWicm P 0 026 kWicm 100 100 4 m 1 64 100 4 m 1 65 m 1 78 10 g 10 4 2 10 a 5 E 3 1 2 1 2 14 a 8 0 1 01 4 01 0 01 4 0 01 4 a T 170s T 250s oo1 T 500s 0 001
253. time correlated single photon counting unit is de scribed in section 2 2 4 and the computer control in section 2 2 5 2 2 1 Light sources and illumination A variety of light sources can be adapted to the microscope Table 2 1 lists the principal characteristics of the light sources used in the experiments presented in this dissertation Light provided by any of these sources is focused with a suitable microscope objective and coupled into a single mode optical fiber For this the fiber is mounted on a positioning xyz0 stage New Focus Inc with sub micrometer precision When using laser light a A 2 and a A 4 plates OWIS GmbH are placed before the light is coupled into the fiber and adjusted to compensate its polarization effects Two Although no Raman scattering measurements were performed in this work the changes in set up configuration required to allow such measurements are straight forward 2 2 Description of the home built confocal microscope 9 PC TCSPC Lens rol Xe arc gr N Lamp Ar Laser 515 488 nm Shutter Microscope objective Mode Locker 120 ps 633 nm CW 200ps Single mode optical fiber r PC Amplifier AD DA Piezoelectric XYZ stage Microscope objective Oscilloscope Dichroic mirror or Ocular 7 beam splitter Diaphragm Flipable mirror Pinhole Filters Beam Polarizing APD litter Beam Splitter A
254. to the ones obtained FL illumination The intensity of all peaks is decreased but this time the y and z components are the most affected The size of all patterns is increased due to the smaller effective NA corresponding to TL illumination As inferred from the experimental results the calculated electric fields show that lisht propagates most effectively through the layered samples via surface plasmon excitation The FB and FL illumination modes as they can excite the surface plasmon resonance SPR generate strong fields of similar intensities In contrast the TL illumination mode cannot excite the SPR and generates therefore a very weak field at the spacer air interface The electric field distribution on the polymer side of the spacer air interface was calculated too but they are not shown The only difference to the results on the air side is the reduction of the z component due to the boundary conditions for the electric field equations 4 3 Thus the results for the x and y components are the same as those shown figure 4 12 and the results for the z components show the same spatial distribution but their intensities are a factor sacer 5 67 weaker In order to examine the influence of the gold film the electric field distribution was also calculated on the air side of the spacer air interface of a sample without the gold film The results are presented in figure 4 13 The same normalization factor as in the case with gold was use
255. tor given by FB BGreg SI FB theo 4 5 Max where FB P is the maximum of the experimental FB signal BG pp is the average max theo max experimental FB background intensity and FB is the maximum of the modelled FB signal After normalization the experimental average background was added to the corresponding profile By applying the same scaling factor to the theoretical fluorescence signals shown in figures 4 14 4 15 and 4 16 and comparing them to the experimental background it is possible to identify the experimentally detectable signals The signals sur rounded by the dashed lines are the ones that present a peak value equal or higher 4 5 Conclusions 83 120 FB 1004 FL TL 100 aa 25 60 S 60 1 20 Z 40 ail 20 154 20 T T T T 7 T T T l T T 7 00 05 10 15 20 00 05 10 15 20 00 05 10 15 20 um um um Figure 4 18 Experimental and theoretical fluorescence signals for the different illumination modes Comparison of the profiles of the experimentally observed grey filled and the theo retical black fluorescence signals for the different illumination modes The experimental profiles correspond to the white lines in the images of figure 4 5 than 50 of the corresponding experimental background this represents a signal to background ratio equal or higher than the TL signals shown in figure 4 18 Those signals are in principle detectable With the same procedure it is possible to pred
256. ts will be carried out to test this hypothesis In combination with the results presented in the previous chapters the results presented here envisage combined measurements of light scattering and fluorescence or Raman scattering on a single particle level Such measurements can be useful to perform a quantitative study of the SERS or enhanced fluorescence in well defined geometries Chapter 8 Summary A sample scanning confocal optical microscope SCOM was designed and con structed in order to perform local measurements of fluorescence light scattering and Raman scattering The components of the SCOM and their functions were described and instructions for the alignment and operation of the microscope were given This instrument allows to measure time resolved fluorescence Raman scattering and light scattering from the same diffraction limited spot Fluorescence from single molecules and light scattering from metallic nanoparticles with a minimum size around 20nm can be studied Two theoretical methods were presented First a theoretical method for the description of the electric field distribution in the focus of the SCOM This enables the design of illumination modes for different purposes such as the determination of the three dimensional orientation of single chromophores or the excitation of partic ular resonances in metallic structures Second a method for the calculation of the de excitation rates of a chromophore This permits t
257. ugh to collect the highly oriented surface plasmon back coupled fluorescence light almost no fluorescence emission is detected The excitation field generated by the surface plasmons decays exponentially from the gold surface with a typical length shorter than the axial field localization obtained in a 1 4 NA confocal system For the geometry of the samples the electric field decays in air with a 1 e distance of 240 nm which is approximately 2 times smaller than the confocal z resolution for 633 nm Another evident observation is that the great majority of the experimentally detected fluorescence signals present the spatial distribution corresponding to the intensity of the z component of the electric field generated at the spacer air interface Figure 4 17 shows an example In fact in all the images recorded containing around 500 fluorescence signals of single molecules no fluorescence signal was detected with a clear x pattern Thus a direct comparison of the calculated fluorescence signals for FB and FL illumination to the experimental images allows to conclude that the single fluorophores behave optically as being placed on the air side of the spacer air interface Furthermore the theoretically calculated fluorescence signals for an ideal QE 1 fluorophore on the air side of the interface are in quantitative agreement 82 Single molecule fluorescence through a thin gold film 0 4 2 um Figure 4 17 Modelled and experimenta
258. uisition software If saved 1 DoAlert 2 The displayed data is not saved Do you want to save it Switch V_flag Case 1 dataname SaveCompleteData dataname break Case 3 Abort Endswitch Endif saved 0 XiScale Xi_um_display XfScale Xi_um_display Scanrange_um_display YiScale Yi_um_display YfScale Yi_um_display Scanrange_um_display XiF 0 XiB Pixels ControlInfo W ControlPanel Calibrate If V_value 1 Else CheckBox KeepPrevFit value 0 win ControlPanel ScanCalib FitLinearRange GoodPixels Pixels If Correct Parameters 0 Abort Scan aborted Parameters out of range after calibration Endif DoUpdate ScanFinalCalib FitFinalLinearRange DoUpdate PixelsToDiscard 0 Xi_um Xi_um_display Yi_um Yi_um_display Scanrange_um Scanrange_um_display Pixels Pixels_display GoodPixels Pixels If Check Values 0 Abort Scan aborted Parameters out of range Endif DoUpdate CalculateParametersForAdWin Init Endif Xi_prev Xi_um_display Yi_prev Yi_um_display A 2 Igor routines 175 Scanrange_prev Scanrange_um _display Pixels_prev Pixels_display Pixeltime_prev Pixeltime DoWindow B PointKinetic DoWindow B Calibration DoWindow F SurfaceScanF DoWindow F SurfaceScanB DoWindow F LineScan ValDisplay Width disable 1 win LineScan Make O I U N GoodPixels GoodPixels CtsdisplayF Make O I U N GoodPixels GoodPixels CtsdisplayB ControlInfo W ControlPanel Cleandata If V_value 1 Ctsd
259. ules is defined by and The di rection of the incident or emitted radiation of r wavevector k is defined by w and 0 yi dielectric contrast the results on either side of the interface can be very different due to the electromagnetic boundary conditions for the electric field For a plane interface between media A and B those boundary conditions write Eja Eis 3 3 eaELA eBELB where and eg are the dielectric constants of the respective mediums The symbols and L stand for the components of the electric field parallel and perpendicular to the interface plane 3 2 The emission The influence of the surroundings particularly a nearby conducting surface on the spontaneous emission of an oscillating point dipole was first discussed by Som merfeld 26 and later extended by Chance et al 27 Analytical solutions to this problem have been obtained by introducing the Hertz vector potential or dyadic Green s functions 27 In this chapter an alternative approach originally due to Weyl 28 and used by a number of authors 29 30 is generalized to be applied to any non magnetic layered system The dipolar field is considered as a superposi tion of s and p polarized plane waves and the influence of the layered system is described by its plane wave reflection coefficients It is convenient to normalize all the rates to the total emission of a free dipole in a given homogenous reference medium 2Wk fea 3 4 12
260. umination transmitted light TL The experimental scheme presented here differs from the one of Yokota et al A sample scanning confocal microscope was used in an episcopic illumination epi illumination arrangement from the gold side see figure 4 1 a The fluorescence of single molecules is excited and detected from the same side of the sample through the gold film In comparison to the experimental scheme of Yokota et al this scheme presents two important advantages First the sample is not constantly illuminated because the confocal scanning technique allows to excite and detect fluorescence on one diffraction limited spot of the sample at a time This considerably reduces irreversible photo bleaching of the samples which is of major importance in single molecule experiments Second the chromophores side of the sample remains free therefore allowing the application of a complementary technique At this point it is worth remarking that due to its metallic properties in the visible range its chemical stability and the possibility of obtaining atomically flat terraces gold has 4 2 Experimental 59 became one of the most widely used substrates for a number of techniques such as scanning tunnelling microscopy STM 52 atomic force microscopy AFM 53 and electrochemistry EC 54 In this chapter experimental and theoretical investigations are combined to ad dress the following questions Is it at all possible to detect
261. under different excitation intensities The blinking of QDs is simulated via a general model in order to identify the influence of the most important experimental parameters The photo induced ef fects of the blinking are taken into account in the simulations via an independent single rate transition from the on to the off state 110 Photoluminescence blinking of Zno 42Cdo ss5e nano crystals 6 1 Brief Introduction and current status Colloidal semiconducting nano crystals also known as quantum dots QDs are of tremendous interest due to their applications as light emitting devices 77 78 lasers 79 80 and biological labels 81 83 In comparison to organic fluorescent dye molecules QDs have a number of comparative advantages they are much brighter they are more photostable and they have a broad band edge absorption and a narrow emission band 81 82 84 The first studied II VI semiconducting crystallites were CdS QDs in the context of photo electrochemical reactions at interfaces 85 86 CdS QDs with a size in the order of 50 show quantum confinement effects that lead to a size dependent lowest excited electronic state The earliest studies on colloidal semiconducting particles were focused on the size dependence of the photoluminescence and the redox potential 87 91 Later on owing to their size dependent tunable emission across the complete visible spectrum CdSe nano crystals have become the most extensively investigated Q
262. uorescence lifetime image the confocal data and the TCSPC data need to be linearized in time and synchronized b Fluorescence lifetime image constructed from the data shown in a c Fluorescence decay curves of a single fluorescent dye and of the background are shown for comparison the detection time after the excitation pulse mic t with 12 ps resolution and the detection time from the beginning of the measurement mac t with 50 ns resolution see section 2 2 4 Then to obtain a fluorescence lifetime image it is necessary to combine these two sets of data in order to assign a position a pixel to each detected photon in the TCSPC Another piece of software was developed to synchronize the two sets of data and construct the lifetime image To perform the synchronization both data sets are linearized in time with an arbitrary origin A histogram of the TCSPC mac t data is done with a bin width equal to the pixel time of the confocal data and then compared to the latter figure 2 19 a If both data sets do not match in time the origin of the TCSPC data is shifted to later times and a small part of the TCSPC discarded a new histogram is computed and compared to the confocal data The process is repeated until an optimum match is achieved Figure 2 19 a Once the TCSPC and confocal data are synchronized the detection time TCSPC mac t time of every TCSPC detected photon is compared to the initial and final collection times of each
263. values For intermediate spacer thickness lt A all the rates present damped oscillations In addition the emission rate into the collection solid angle of the objective Poj was calculated and it is shown in the small graphs on the left of figure 4 9 Fluorescence detection is accomplished with a 1 4 NA microscope objective from the glass substrates which have a refractive index Ngiass 1 503 The collected radiation is therefore the fraction emitted to the glass side up to an angle 0 68 6 from the normal to the interfaces Then To is obtained by integrating equations 3 6 between zero and kmar Nglass Ko sin 68 6 where ko is the wavevector of the emitted radiation A 670 nm in vacuum From these calculated rates it is possible to obtain the fraction of the electro magnetic de excitation rate corresponding to radiation collected in a time unit by the microscope objective Pae Ton T E2 equation 3 14 This ratio represents for an ideal molecule with a null intrinsic non radiative decay rate T 0 the detectable fluorescence emission In figure 4 10 T g r is plotted as a function of the spacer thickness for dipoles parallel and perpendicular to the sample plane In both cases laet oscillates as a function of the separation distance to the gold film However in the range of spacer thicknesses used in the experiments Paet increases monotonically with the spacer thickness and is always higher for the perpendicular dipo
264. wn in figure 5 2 computed with different bin widths a 0 75 ms b 5ms c 0 05ms To distinguish the on from the off bins it is necessary to set a threshold for example the horizontal line shown in a off periods remain hidden inside the big bins The opposite occurs if one takes very narrow bins figure 5 4 c the increase in time resolution is paid by higher uncer tainty in distinguishing which bin is on and which is off Then there is a bin width range that makes the optimum compromise between time resolution and accuracy in distinguishing the on from the off state The procedure presented here is dedi cated to find this bin width place a suitable threshold and extract the information from the data histogram This method is based on the fact the on state and the off state have a character istic intensity J 7 standing for on or off Then the number of detected photons N in a unit time in any of the two on or off states is a Poisson random variable with an average rate J and the probability of detecting k photons in a time unit is k Lj I e m 5 14 P n k The number of photons N detected not in a time unit but in a time interval 92 Single molecule fluorescence dynamics t gt 0 is then LHE elit Pum k u 5 15 Calling X the waiting time until the next photon the probability that X is larger than a time t can be calculated as N e 5 16 gt Pxy 1 e 5 17 r me 5 18 This has the
265. y R 0 indicates no evidence of a linear relationship between X and Y 6 4 Modelling the QDs blinking 123 where X and Y stand for the logarithm of the corresponding time period plotted along the x or y axis in the scatter plots of figure 6 11 X and Y are the averages of X and Y respectively and the sums are performed for all the n points of each scatter plot Glass ITO 06 0 6 x ON vs next ON o a Sr bon 02 OFF vs next OFF 0 2 r e ON next OFF ee ON vs next OFF 0 0 gt girma 0 0 E 05 10 15 20 05 10 15 20 P kWicm P kW cm Figure 6 12 Correlation coefficient of adjacent on and off times as a function of the excitation intensity P Figure 6 12 shows the correlation coefficients obtained from the on and off periods of QDs on glass and on ITO for the different excitation intensities No evident dependence of the correlation behavior on the excitation power was found The correlation coefficient R between successive on times and between successive off times oscillates around 0 5 and the correlation coefficient between adjacent on and off times oscillates around zero This results support the idea that the processes responsible of turning a QD bright is distinct from the process that turns a QD dark In addition it indicates that both processes have certain residual memory This residual memory effect was predicted by the model of ionization through fluctuating tunnelling barriers pro posed by Kuno
266. y feature as a function of the scanning speed b the corrected monitor signal was added to figure 2 6 c forward F and backward B images after the corrections 16 The fluorescence and light scattering confocal microscope position of one feature can be followed on both forward and backward images as a function of the scanning speed In this manner as shown in figure 2 7 a the position shift vs scanning speed can be fitted with a polynomial function and the empirical correction parameters are obtained Using the correction parameters the monitor signal can be corrected to give the actual position of the sample In figure 2 7 b the same data of figure 2 6 is shown but this time the empirically corrected monitor signal is shown Then the pixels corresponding to the forward and backward linear regions of the corrected monitor signal are found following the algorithm detailed in the appendix A 3 and the forward and backward images are constructed These images delimited in figure 2 7 b by the vertical white lines and shown in more de tail in figure 2 7 c are consistent the average shift between forward and backward images is smaller than one pixel size 2 2 3 Detection The detection channel can be adapted to different experimental designs In the experiments presented in this dissertation fluorescence and light scattering mea surements were performed and the corresponding detection schemes are described below Common to all

CONFOCAL MICROSCOPY APPLIED TO THE STUDY OF SINGLE

Contents

Download Pdf Manuals

Related Search

Related Contents