Home

Conceptual Design and Setup of a Fully Digital Recording

1. 63 64 where fr 2I 1 2I 2I 1 2I 2 27 3 The c M have the properties m IM c amp l M F1 AM CUMF2 Substituting Eqs 62 and 64 into Eq 58 one obtains 24r hw 2 IM HelIM gt gt ff oP DVS bn 2 M a cP M Weis F Lc cP M v9 T cm v2 Om where the quadrupole coupling constant wg is defined as oT HQI DA 65 67 From Eq 57 one can see that in the principal axis system only the diagonal terms 2 Vi and the off diagonal terms V5 are not equal to zero The non vanishing matrix 24 2 PAC THEORY 2 5 Static Electric Quadrupole Interaction elements of the quadrupole Hamiltonian are therefore IM 2 HailIM Rwgzc m 68 and IM 2 Hor IM hwgc m 69 For the axially symmetric case 7 0 the quadrupole Hamiltonian is diagonal The eigenvalues Ea M IM Hg IM hwe 3M I I 1 70 are doubly degenerate in M As an example the spin state J 3 shall be considered as the two most commonly used PAC probes In Cd and Hf Ta have intermedi ate state with nuclear spin 5 The possible energy eigenvalues can easily be evaluated by Holts 1hwo Eol 2hug 71 1 wr E EE B Guo wn Eg 2 Ba 5 h 120g 72 ws Bo 2 Ba h 1800 For the axially symmetric case 7 0 all three PAC frequencies w1 w2 w3 for spin state I gt li
2. r lt n gt rest for lt n gt s if sequence file was not created yet Default is r0 no resting s lt n gt Evaluate only segments larger than lt n gt in SEQ file Default is s1 S lt n gt Evaluate only segments lower than lt n gt in SEQ file Default is S8000 v lt n gt 0 quiet mode 1 default mode with some messages 2 debug message mode Switches cij Reconstruction of dump files CIJ DAT into DIJ DAT files If this switch is set then operation of cocheck alters from its normal behaviour No sequence files are processed only dump files are being read in cfd Regard constant fraction discriminator result in timestamp Normal behaviour is not to add the CFD result to the timestamp h Print this help O force Open Window Mode on all channels rm remove SEQ files after coincidence search has been performed on them wb Save in binary mode 103 A 4 Configuration files and Command line options A APPENDIX A 4 6 Cocheck conf Path to data folder on Slaves Usually this path information is preceded by mnt pacX mount point of data folders projectPath data lt project name gt Energy windows energyWindow1 1000 2927 2928 4500 energyWindow2 1000 2731 2732 4000 energyWindow3 1000 2574 2575 4000 energyWindow4 500 2600 4000 6000 Time scaling factor in ps timeScale 10 t zero correction factor in ps Add to the timestamp of each Channel to account for signal propagat
3. 89 N The present 4 detector setup provides a set of 12 time spectra four 180 and eight 90 spectra Figure 13 depicts all the conventional detector pairings which can be constructed from the pool of the 12 time spectra The 12 spectra can systematically be combined so that 4 different conventional count ratios can be built These used to be referred to in literature as the two forward I II and two backward III IV count ratios In order to separate the perturbation function W 0 t from Eq 89 the following 60 5 EXPERIMENTS 5 2 Data Processing count rate ratio Z t is constructed from the individual count rates N 6 t Z t No2 180 t Ni3 180 t z Nos 90 t N12 90 t Ko kK 1EQ9E1E2E3 No W 180 t KoK1E0E1E2E3 a No W 90 t W2 180 t 90 W2 90 t Note that all in general unknown efficiencies cancel each other However if the 180 and 90 count rates are arbitrarily combined the terms and k might not necessarily vanish For instance is No2 180 t N13 180 t k W 180 t K 1208 NaO ANa T WO xp ON The remaining fraction of x and shall here be defined as the Start Stop efficiency between Detector i and j When the balanced ratio 90 is known one can gain all the other detector efficiencies by thoughtfully combining the count rates N t A matrix that describes the detector efficiencies can then be defined accordingly as For polycrystal
4. int status Read an integer from the file handle 44 4 SOFTWARE 4 2 Slave applications It holds the interrupt request flag in the 1st byte and the parallel port status word in the 2nd byte fread amp status sizeof status 1 f Load the command word int command __PACMOD_PMT Write the command to the device fwrite amp command sizeof command 1 f fclose f return 0 A file handle is created by the C function call fopen The fread function returns the parallel port status in the integer variable status The information is aligned as follows Bit 0 Interrupt flag Bit 1 7 reserved 0 Bit 8 15 Parallel port status See Table A 8 for details Bit 15 31 reserved 0 The fwrite function call sends a command word to the device driver The commands are predefined in the project header file pacmod h All implemented commands and their descriptions are shown in Table 3 4 2 Slave applications 4 2 1 Pacslave The slave process Pacslave performs operational tasks of the Slave computers It provides a TCP IP server socket for the Host application to connect to over the local area network The server socket is listening on port 15000 It was designated for the exchange of project settings between the Host and the Slaves Pacslave interfaces with the handshake bus and drives the Slave s logic signals However the main task 45 4 2 Slave applications 4 SOFTWARE Comma
5. Maxwell solutions the angular distribution of the y radiation can be calculated using the energy flux density the Poynting vector S E x B 4 E and B are perpendicular to each other and thus points into the direction of the y radiation In the far field region the norm of the Poynting vector follows from the Maxwell solutions 3 and 4 gt C gt gt m Bela er P 5 The magnitude of the energy flux density for both E l and M l radiation is pro portional to the expectation value of the spherical harmonic functions Thus the radiation characteristic of the electro magnetic wave is anisotropic with respect to 0 6 2 PAC THEORY 2 1 Theory of Angular Correlation and y Therefore exists an angular correlation among the transitions which gives in formation about the nucleus The spherical harmonic functions can be fully described in polar coordinates by the associated Legendre polynomials P cos 0 9 Yr Oo Pose 6 The application of the angular momentum operator L on the spherical harmonic functions Y addresses the various magnetic sublevels m for a specific spin state I Multipole radiation originates from transitions between magnetic sublevels causing changes in angular momentun of a spin state The expectation value is ILY The normalized angular distribution functions Ly 2 gt 2y give a set of functions which are related to the radiation characteristics 7 2 1 The
6. Wine Plan 5 arte heads SRG eae ek gr a 89 A 3 2 Analog Multiplexer Module 89 A 3 3 Synchronization and Channel Select Module 89 A 3 4 Reference clock module 04 89 A 4 Configuration files and Command line options 94 AA Pacslave s aree 84 82 84 3 a aaa 94 A 4 2 Pacslave conf Sana a ua ae ee 97 AA Packet area ee 99 AA Pachost Conf ni et ee ae Rn ee teten 101 AA y COCWEEK 2 2 320er Elle We 102 AAG Cocheck Conf a da 2 0 ae er eh 104 CONTENTS CONTENTS References 105 Lebenslauf 109 III LIST OF FIGURES List of Figures OS AN DoF UON Mm mM NY BR Be Be Be Be Be Be Be Be k DH OO WON DT FP WY KF CO Illustrations of angular anisotropy characteristics Energy scheme for J 5 2 intermediate state Functional block diagram 2 2 a ait Meran at pre Handshakes bus timing diagram 2 2222 Er ee Subsequent segments in memory 2 2 2 2 m e a a Block diagram of multiplexer circuit 422 eA woes Aa Single transient pulse 22 2 228 2 a8 ara Illustration of the CFD technique applied on transient waveform Time deviation results with CFD 222222 Flow chart for PAC processes Pachost and Pacslave Energy resolution of PMT detectors 2 2 22 22mm Time spectrum of a detector pair 2 nn mann Conventional Start Stop detector arrangements Effective anisotropy A22eff with respect to detector di
7. dark noise effects inside the detector such as thermal electrons may induce a 59 5 2 Data Processing 5 EXPERIMENTS Start gt Stop I 3 I 3 II 3 IV 3 Ann Pa a aan 1 1 1 1 Figure 13 Conventional Start Stop arrangements for the 4 channel acquisition sys tem The 12 time spectra can be systematically combined to achieve at least the 4 conventional PAC count ratios I IV trigger signal Therefore the probability detecting a Start event in detector 7 can be weighted by a value The probability of detecting a Stop event after successfully detecting a Start event can be measured in terms of the half life Ty of the intermedi ate state Values for Ty for the used isotopes in this work respectively are given in Tables 1 and 2 For each detector pair i j the raw data contents can be expressed as e7t N Dis 0 t siezi e NW 8 t Big 88 W 6 t is the angular correlation function 35 The average background count B is a constant fraction due to random coincidences collected by the spectrometer For times tthat are much shorter than the half life time of the nucleus the term e 7 1 For large t the count rate is D 0 t Bij The background corrected count rate is t TN N 0 t D 0 t Bij Se NoW 6 1
8. Hyperfine Interactions 159 1 4 2004 379 383 J A Gardner C H Herden G Herden C Dreyer G Bulatova J Sullivan and P Sullivan Simulaneous Braille Tactile Graphics and Ink with Tiger Ink Attachment and Duxbury Proceedings of the 2004 CSUN International Conference on Technol ogy and Persons with Disabilities Los Angeles CA March 2004 J A Gardner C H Herden G Herden A Neeley C Dreyer H Stowell V Bu latov G Bulatova L Ungier and Y Sun Viewplus EmPrint Ink Pro Formatter and Designer Seeing and Feeling is Believing Proceedings of the 2005 CSUN Inter national Conference on Technology and Persons with Disabilities Los Angeles CA March 2005 J R der C H Herden J A Gardner and K D Becker Fully Digital Time Differ ential Perturbed Angular Correlation TDPAC Spectrometer submitted to Nuclear Instruments and Methods Section A NIMA S 07 00089 J R der C H Herden J A Gardner K D Becker M Uhrmacher and H Hofs ss Actual Concepts of Digital PAC Spectroscopy submitted to XIV International Conference on Hyperfine Interactions Iguazu Brazil August 2007 110 Konferenzbeitr ge C H Herden M Alves K D Becker and J A Gardner A new generation TD PAC spectrometer XIII International Conference on Hyperfine Interactions Bonn Germany 23 27 August 2004 Poster J R der C H Herden J A Gardner K D Becker M
9. October 1996 N J Stone Table of Nuclear Magnetic Dipole and Electric Quadrupole Moments Atomic Data and Nuclear Data Tables 90 1 2005 75 176 A Messiah Quantum Mechanics Volume 1 North Holland Publishing Company Amsterdam 1961 H Fischer and H Kaul Mathematik ftir Physiker Band 1 Teubner Studienbiicher Mathematik McGraw Hill Dictionary of Scientific and Technical Terms McGraw Hill Companies Inc 1974 2003 answers com 2003 Wikipedia The Free Encyclopedia www wikipedia org 2003 E Matthias W Schneider and R M Steffen Nuclear Level Splitting Caused by a Combined Electric Quadrupole and Magnetic Dipole Interaction Phys Rev 125 1962 261 268 A R Edmonds Angular Momentum in Quantum Mechanics Chapter 3 and 6 and tables Princeton University Press 1957 125 132 J Lu Stochastic Models of Perturbed Angular Correlation due to Diffusion of Defects in Materials PhD thesis Brigham Young University June 1995 H J ger J A Gardner J C Haygarth and R L Rasera Structural Characteriza tion of High Temperature Zirconia Ceramics by Perturbed Angular Correlation Spec troscopy J Am Ceram Soc 69 1986 458 463 105 REFERENCES REFERENCES 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 H J ger J A Gardner H T Su and R L Rasera Microcomputer controlled perturbed angular correlation
10. The unit cell of a chalcopyrite consists basically of two sphalerites that were doubled in c direction and two different cations In sphalerite exists a face centered cubic arrangement of A sites and B sites in the center of each tetrahedron For CdSiP2 the cells with silicon atoms alternate with cells containing cadmium The phospor is positioned on the chalcopyrite C sites The double cells mean that chal copyrite is tetragonal but the fact that the unit cell consists of two cubes means chalcopyrite mimics cubic symmetry very often TDPAC investigations gave very 78 6 RESULTS 6 2 PAC measurements 111 Cd in CdSiP ii a al Nu Ay h phi I h Ax et FD Fourier intensity 0 50 100 150 200 250 300 350 0 0 0 2 04 0 6 t ns o 10 rad s Figure 20 Time dependent anisotropy A22 ef fG22 t and corresponding Fourier trans form for In Cd in polycrystalline CdSiP3 at room temperature The sample shows quadrupole interaction with an axially symmetric EFG 7 0 The fundamental PAC frequency was determined as w 107 2 Mrad s leading to a quadrupole cou pling constant of wo 17 87 MHz clear spectra and in previous works it was shown that the In Cd cation is really positioned on the chalcopyrite A site 35 The nearest neighbour to the Cd atoms are the P atoms The time dependent anisotropy in Figure 20 is a perfect example for an axially symmetric EFG 7 0 in a polycrystalline sample The quadrupole coup
11. and energy analysis Data storage Channel 1 Channel 2 Channel 3 Channel N Coincidence analysis User interface Data storage Figure 3 Functional block diagram of the PAC spectrometer The diagram demon strates the components of the digital PAC spectrometer design Every PMT channel is peer to the other The design comprises no limit to the number of total PAC channels While an acquisition is in progress and whenever it is convenient the Host also analyses the recorded time and energy information from the Slaves The data is exchanged between the Host and the Slaves in a Local Area Network LAN The new TDPAC spectrometer design approaches the acquisition of time of arrival and energy information from radiating probes more directly The idea is to record only the PMT signal with as little electronic equipment as possible because each additional component is known to contribute parametric errors to the measured data Compared with conventional systems the new setup design is much simpler The complete adjustment of a traditional PAC spectrometer is rather difficult because of its complexity With the new approach the PMT signal is directly input to the digitizer card which eliminates the need for external amplifier or signal conditioner units Functions which were accomplished through electronic hardware components before are now implemented as part of the PAC software 31 3 2 Handshake Bus 3 HARDWARE 3 2 Handshake Bus The op
12. as I M k1 01 HI Mi 9 I My ko 02 Ha I M 10 where H and H are the transition inducing Hamiltonians The first braket repre sents the probability to find the nucleus in the state 7 MY right after the detection of y in direction k and polarisation c The second braket represents the probability to detect Y under the angle Z ka k The explicit calculation over all possible states gives rise to the angular correlation function 2 W ka ke gt DL mil H m m Hi m 11 mi my 01 02 m The sublevels m m m form a complete set of eigenstates for the interme diate state J Therefore the summation in Eq 11 is carried out incoherently as all the mixed states have to be taken into account In order to simplify the expressions the lengthy terms M have been substituted by m Rearranging the indices yields 11 2 2 Perturbation of Angular Correlation 2 PAC THEORY to the expression W Fi ko I I m H3 ms my Halm I ml Haji mil Hi m 12 m m mf Mi With the definition of the density operator p gt Gm M m one can define p k Im H p Hi m and p k2 m H os H2 m 13 Then Eq 12 takes on the form W k ka gt m p k2 m mlok m m m Tr o F2 p h1 14 This form is the most general expression for the angular correlation function In the density matrix formalism p k and p k describe the system immediately a
13. be produced by the existence of valence electrons electrons in the conduction band and other defects caused by vacancies and dopants PAC is used mainly in the research of solid state materials as a tool with high sensitivity to characterize local host properties and thus relatively low numbers of radioactive probe atoms are necessary Conventional PAC equipment that is necessary to accomplish the detection of corre lated events that are for the most only nano seconds apart is usually very complex and the setup is very time consuming A PAC instrument is often tuned to accommo date experiments that use a specific isotope to which the timing resolution and the energy window settings were calibrated before Isotope exchanges are virtually not possible without recalibrating the whole instrument Further the tremendous amount of electronic circuits necessary for operation also limits the deployment of recording channels So far it was not possible to add an arbitrary number of recording channels to conventional PAC instruments without gaining complexity in the electronics by orders of magnitude In conventional PAC spectrometers Time to Amplitude Converters measure the time interval between pulses to their start and stop inputs and generate an analog output pulse proportional to the measured time This analog pulse is then converted into digital form so that a computer is able to process the information One negative aspect of this quasi analog timing
14. connector Signal Output PMT2 Anode Output HV Input Analog Multiplexer 3 Channel Select Input Sync Pulse Input PMT Input Busy Line Output HV Power Supply Power Supph D SUB 25 connector Signal Output PMT3 Anode Output HV Input Analog Multiplexer 4 Channel Select Input Sync Pulse Input PMT Input Busy Line Output HV Power Supply Power Suppl D SUB 25 connector Signal Output Anode Output HV Input Channel Select Input Sync Pulse Input PMT Input Busy Line Output HV Power Supply Ethernet Switch CATS Ports SATE Port 4 CATS ei CATS bart Port 1 Slave Computer 1 CATS Analog Multiplexer 1 Power Supph Power Supply Handshake Distributor Parallel port Host Computer D SUB 25 connector D SUB 25 connector Parallel port D SUB 25 connector D SUB 25 connector DP110 digitizer card Channel Select Input Channel Select Output 1 Sync Pulse Input Channel Select Output 2 Ethernet Adapter Analog Input Signal Output PMT Input Channel Select Output 3 RJ45 jack External Clock input Channel Select Output 4 Busy Line Output Ethernet Adapter Sync Pulse Output 1 Sync Pulse Output 2 RUSS jack Syne Pulse Output 3 Syne Pulse Output 4 Busy Line Input 1 Busy Line Input 2 Busy Line Input 3 Busy Line Input 4 Clock Distributor Clock O
15. d 5cm Detector Detector rate 10 s time d Figure 17 Activity of the c In Cd in Fe sample over time as recorded over a time span of 456 hours The initial activity was measured to be around 0 315 MBq The exponential fit yielded a half life of 2 84 days Figure 17 shows a plot of the detected sample activity over time At the beginning of the experiment the detected activity of sample c was determined to be 0 315 MBq The dashed line represents an exponential fit to the data points The half life of the fit was determined to be t a 2 84 0 02 which is in good correspondence to the value given for In Cd in Table 1 According to the measurements above the average trigger rate of each detector channel was 11 4 kHz at the beginning of the experiment Thus the DSP boards recorded one sequence of 8000 segments on average in approximately 8000 87 7 us 0 7 ms The sequence files contain an absolute Unix timestamp that signifies when the data was collected The Unix timestamp gives information about the recording cycle period teycle The cycle time teycie that has elapsed during the recording of one sequence however exposes a variance with respect to the online time tontine This variance is the dead time or offline time tof fine of the spectrometer Figure 18 shows the duty cycle plotted as a function of the detected sample activity The duty cycle was obtained from the ratio of the online time toniine and the cycle
16. diagonalized It can further be shown that Qi is trace free tr Q 0 The forth term of 42 is the quadrupole energy term It can be expressed as z gt VaaQaa 44 where Vaa is the tensor of the electric field gradient EFG The quadrupole energy results from the alignment of the nucleus quadrupole moment with an external electric field gradient As in this work only the quadrupole interactions are of interest the other terms are disregarded in further considerations The Vaa is a symmetric tensor with tr Vaa 0 Typically Vaa is expressed through the asymmetry parameter 7 and Viz by Vea Vyy n n can be a value between 0 lt 7 lt 1 provided the principal axis is defined such that Viz gt Vyy gt Vrx 7 0 is the axially symmetric case where Vys Vy with the z axis as the EFGs symmetry axis Vaa and Qaa are tensors in cartesian coordinates They were derived by using the classical approach The quantum mechanical approach to the solution above is similar A thorough introduction of electric field gradient interaction has been written by Lu 15 The Coulomb energy of a charge distribution in an external potential is described by the Hamiltonian for electrostatic interaction ma 46 ei The quantities with index 7 are the charges and coordinates of the nucleus and the quantities with index 7 are those of its environment This is the quantum mechanical analogue to 37 For r gt r the d
17. of the data as well Transients whose integral and shape were identified as being less than zero or oddly shaped were discarded The percentage of discarded events is on average under 1 The recorded activity for a single channel is given by the number of trigger signals registered within a time frame However the measured detector activity does not reflect the original sample activity as only a fraction of the samples decay is detected The solid angle fraction for the four detector setup used in this work can approximately be expressed as fsa 4 Ap As This ratio represents the area of the four scintillator crystals 4 Ap on the surface of a sphere Ag The radius of the sphere is equivalent to the detector distance with the sample in the center For this experiment a scintillator crystal diameter 3 81 cm and a detector distance d 5 cm yield to the solid angle fraction fs 14 5 Hence it can be assumed that the original sample activity is by an amount of 1 fsa 6 9 higher than the detected one The average activity was determined from a set of 100 sequences that were picked out of the pool of sequence files at various times For 100 sequences the online time differences t trast Ufirse Of the first and last online recorded segment were taken and the average activity calculated as 100 ans i A 8000 Ai Sy 101 t m 101 n n 1 online 71 6 1 Performance 6 RESULTS Oct 28 21 24 52 t n 2 84 0 02 d
18. on Hyperfine Interactions amp XVII International Symposium on Nuclear Quadrupole Interactions Bonn 23 27 August 2004 Jens R der Christian H Herden John A Gardner Klaus D Becker Micheal Uhrma cher and Hans Hofs ss Actual Concepts of Digital PAC Spectroscopy Poster XIV International Conference on Hyperfine Interactions amp XVIII International Sym posium on Nuclear Quadrupole Interactions Cataratas del Iguazu 5 7 August 2007 Abstract Perturbed Angular Correlaction spectroscopy is a hyperfine interaction method that gathers information regarding the environment of a radioactive probe The method is based on the measurement of the anisotropic correlation between directions of emis sions of two gamma rays during the decay of the radioactive nuclei It utilizes the information collected by photomultiplier tubes which are placed around the sample The life times of the intermediate nuclear states are in general very short Therefore the detection system is bound to observe coinciding occurences preferably within fractions of a nanoseconds In order to support the Perturbed Angular Correlac tion methods need for Time of arrival detection and energy discrimination a versatile system was built using Digital Signal Processing equipment Each channel uses its own programmable signal processing board in an off the shelf computing unit The peer design of the system allows the expansion to basically any number of channels T
19. options A APPENDIX Number of channels participating in measurement Default is c1234 all channels Example c24 only channels 2 and 4 h Print this help k lt n gt 0 do not send acquisition termination signal 1 send acquisition termination signal Default is k0 p lt n gt Set process priority see man process for details q Quiet mode tm Test mode will create pulses every 2 us to test performance of spectrometer Commands ct Create single sync pulse Triggers STROBE signal cc Create sync pulse continously cs Switch channel to SYNC Deasserts CHANNEL_SELECT cp Switch channel to PMT Asserts CHANNEL_SELECT 100 A APPENDIX A 4 Configuration files and Command line options A 4 4 Pachost conf pachost config file Author Christian Herden Date O7 Sep 2004 IP addresses for computers for each PAC channel ip1 192 168 1 1 ip2 192 168 1 2 ip3 192 168 1 3 ip4 192 168 1 4 Port to open on PAC channel hostPort 15000 Number of PMT channels nbrChannels 4 Energy downscale factor energyScale 2 Destination project path projectPath data 2007 10 17 dummymessung Number of default segments per sequence nbrSegments 8000 Number of default samples per segment nbrSamples 400 Number of default starting sequence startSequence 1 Number of default ending sequence endSequence 200000 Name of pacmod device name parportDevice dev paci 101 A 4 Configuration files and Command line option
20. probe In Cd This calculation was performed using the Monte Carlo method where the effective angles were assumed for 1 5 diameter scintillators under variable distances The sample dimensions were assumed to be a point sample In addition Figure 14 o exhibits effective anisotropy coefficients that were obtained from measurements on polycrys talline samples during the development of the new PAC spectrometer The measured anisotropy is mostly below the theoretically evaluated values The sole considera tion of the effect of the finite detector size attenuation does not explain the result sufficient enough It must be assumed that the effects of the remaining attenuation factors contribute to the reduction of the effective anisotropy The good resolution of the new PAC spectrometers timing mechanism may lead to the assumption that the finite time resolution effect is relatively small and thus y 1 As these results were obtained from solid samples with finite size the effects of Compton scattering and Self sample adsorption cannot be neglected The Aa2 rj for In Cd in solution in Figure 14 A were obtained using a diluted InCl3 sample Therefore the attenuation effect caused by Compton scattering and self sample adsorption can be mostly ruled out The magnitude of the Aaa rr for the diluted sample were measured to be comparable to the Ago of the solid samples 5 3 Timing Tests Correct timing calibration is essential
21. s The effective anisotrophy was determined as A22eff 0 143 This value varies in general with the detector distance which was d 7 cm in this experiment The obtained core parameters S n from this least square error fit are so 0 218 s 0 256 S 0 403 s 0 123 From the above results the quadrupole frequency wg w 6 for I 5 2 can be derived The quadrupole coupling constant wg is defined for the axially symmet ric case 7 0 by Eq 67 For the quadrupole coupling constant one obtains Wg w 6 5 79 Mrad s for half integer nuclear spin With the quadrupole mo ment Qs5 2 0 77 10 78 m for Cd the magnitude of the electric field gradient element is V 1 98 107 V m These results are in good agreement with the quadrupole coupling constants wo 5 84 1 Mrad s or wg 5 73 4 Mrad s found in the lit erature 48 49 Measurements were performed on In Cd in Sn hosts for various pressures at room temperature 51 Although the published PAC spectra show no texture effects in the core parameters the actual quadrupole frequencies agree very well with the ones found with this spectrometer 6 2 2 Sample b In Cd in Cadmium Silicon Phosphide The quadrupole interaction of the chalcopyrite semiconductor CdSiP was investi gated with the new PAC spectrometer setup The electric field gradient in this type ANBIVCY compound is known to be axially symmetric 7 0 when the sample was annealed 34
22. spectrometer Rev Sci Instrum 58 9 1987 1694 1698 G S Collins 1981 unpublished in ref 19 p 172 H T Su Perturbed Angular Correlation Spectroscopy of Oxide Ceramics at High Tem peratures PhD thesis Oregon State University 1990 25ff A B hm Measuring y ray emission time differences using direct photomultiplier tube transient recordings Master thesis Oregon State University Fachhochschule Ravens burg Weingarten 1997 E Metin Characterization and fitting of BaF scintillator detector transients ob served during y ray detection Master thesis Oregon State University Fachhochschule Ravensburg Weingarten 1997 Acqiris Digitizers User Manual ZM020010C Rev B Switzerland Geneva 2 4 Analog Devices AD8180 AD8182 Switching Multiplexers Datasheet Rev B Nor wood MA U S A ECS Inc International 3951M Series Clock Oscillators Datasheet Ridgeview KS U S A National Semiconductors DS26LS31 Datasheet Ridgeview KS U S A Hamamatsu Photonics R2059 Datasheet Hamamatsu City Japan Ortec 566 TAC Datasheet Oak Ridge TN USA Canberra Industries 2145 TAC Datasheet Meriden CT USA J Axelson Parallel Port Complete Lakeview Research 2000 P J Salzman The Linux Kernel Module Programming Guide Linux Documentation Project www tldp org 2001 Linux Kernel Drivers The Linux Kernel Archives www kernel org L B
23. system is the requirement of timing calibration Prudent adjustments are necessary that relate pulse height to timing resolution Otherwise the acquisition would simply operate out of bounds and deliver unsatisfactory results As a result due to the large amount of maintainance that is needed to keep the PAC system going only a limited group of scientists world wide are deploying this method in research labs And for these reasons PAC spectroscopy is not as commonly used as other nuclear research methods The main objective of this work is to present a versatile time of flight acquisition system that enables its application in time differential PAC TDPAC perturbed an 1 1 INTRODUCTION gular distribution PAD and possibly positron annihilation spectroscopy PAS In principle all of the named methods have the same requirements on the data collection equipment This work presents the realization of a multi channel acquisition system with the help of off the shelf state of the art programmable fast Digital Signal Pro cessing boards Each channel utilizes its own DSP board mounted in a dedicated Personal Computer The peer design of the electronic system allows the addition of any number of channels The benefit of having a larger number of detectors is clearly the increment of statistical information and in addition shorter measuring times This enhances the subsequent use of short lived isotopes A simple electronic handshake bus is lin
24. the classical Larmor frequency and N M M the distance of two M states It describes the precession frequency of the nuclear angular moment T under the influence of a B field The magnetic flux B causes the splitting of the degenerate energy levels For magnetic interaction the perturbation factor GN t 26 2 PAC THEORY 2 6 Static Magnetic Dipole Interaction Zeeman Quadrupole ch Jour see ob er in Figure 2 Energy scheme showing the degeneracy of an intermediate state I 5 2 On the left are pictured the equidistant energy levels due to magnetic dipole in teraction with their characteristic Larmor frequency w On the right the removal of the degeneracy due to quadrupole interaction is presented with the typical PAC frequencies w1 wa and the sum of them ws is calculated by inserting Eq 77 into the Eq 30 The obtained expression for the perturbation factor GAN t e Nwrt 78 is modulated by integer multiples of the Larmor frequency The indices N and k underlie restrictions of Eq 24 Because N lt 2 the maximum observable harmonic of the Larmor frequency for J 5 2 is N 5 For the polycrystalline sample case the index k in Eq 35 is further restricted to kmar 2 and k even Since also N lt k the perturbation factor for the polycrystalline sample case can only contain the Larmor frequency wz and its second harmonic 2w With Eq 78 the magnetic angular correlation function t
25. the correct operation of the recording apparatus very well All measurements show the repro ducibility of the PAC frequencies in samples a d with the new spectrometer setup The few deviations in the results may mainly occur due to differences in the sample preparation methods However the results were all determined within fea sible error margins A careful determination of the relationship between anisotropy and distance was not main subject of interest in the thesis The dependence of the effective anisotropy to the detector distance deviates strongly from the theoretical model The few data points obtained through various test measurements express only roughly the tendency of an increasing anisotropy with increasing detector dis tance Within the appointed distances they never quite reach the magnitude of the predicted value Therefore it must be concluded that additional attenuation values must contribute to the reduction of the effective anisotropy It cannot be ruled out that flaws during development of the new data reduction methods were responsible for inaccurate results Further it is not even clear that the ratio between the effective and the absolute anisotropy can be calculated that well As demonstrated by the test experiments the timing resolution of less than 400 ps of the digital recorderboards is outstanding in comparison to conventionally used analog Time to Amplitude converters In the future the timing resolution may eve
26. the intermediate state I M The nucleus behaves like a non spherical charge distribution Its energy level depends on the quantity of nuclear and external charges and how this charge distribution is oriented with respect to the external EFG The EFG is produced by an electrostatical potential 1 ej en Be 36 Kr 36 This field originates mostly by external negative charges such as by electrons sur rounding the nucleus The classical description of the energy of a nuclear charge distribution py 7 in the potential 7 is the Coulomb energy Ea oxar 37 v The potential amp r can be developed into a Taylor series around the center of the nucleus 7 10 7 gt Fe ver _ z 2 7Vo z YPE 5 20 Ly PAE 7 Ox 2 7 02 0 NI Without loss of generality it can be assumed that the nucleus is located at X 0 The first term expresses the contribution from a point charge at the center of the nucleus to the electrostatic potential 9 z 0 The second term contains the electric potential Ey Vd z 0 It describes the interaction between a dipole moment of the nucleus and the electric field The third term is called the quadrupole term In fact it is a combination of a monopole term and the pure quadrupole term The monopole term is related to the average radius of the nucleus and is responsible for 18 2 PAC THEORY 2 5 Static Electric Quadrupole Interaction the isotope shift in
27. time dependent anisotropy Aa2 eff G22 t sl 6 2 PAC measurements 6 RESULTS n 0 24 0 0 1 86 4 Ay Fd Fourier intensity 0 10 20 30 40 50 60 70 00 0 5 1 0 1 5 t ns o 10 rad s Figure 22 Time dependent anisotropy A 2effG22 t and the corresponding Fourier transform for Hf Ta in Hf metal The data point resolution is 0 25 ns channel The sample shows an asymmetry of 7 0 24 The obtained fundamental PAC frequency at room temperature is w 309 2 Mrad s at room temperature and its corresponding Fourier transform are displayed in Figure 22 The PAC spectrum for sample d exhibits a nonaxial symmetric EFG It was fitted to Eq 103 with a single fraction f 1 The core parameters s2 7 were held invariant Since the core parameters are a function of the asymmetry parameter their theoretical values were taken from published tables 54 The three PAC frequencies spacing were attributed to an asymmetry parameter n 0 24 The fundamental frequency yielded to a quadrupole coupling constant wg 51 5 0 3 Mrad s The quadrupole moment of Hf Ta intermediate level is Q5 2 2 35 107 m Using Eq 67 an EFG component V 5 71 10 V m can be derived from the above values The PAC spectrum also exhibits an attenuation of the anisotropy which essentially is visible as broadening of the quadrupole frequencies in the Fourier spectrum This is typically attributed to the presenc
28. to the conclusion that the signal propagation times per channel are not equal Variations in cable lengths or tolerances in electronic components can cause a phase shift in the signal The knowledge of the these delay times however allows one to account for these times when computing the count ratio 94 All results have been summarized in Table 5 The gaussian profile in Figure 15 b shows the prompt peak after smoothening the clock granularity For smoothening the granularity an artificial jitter was added onto the timestamps The jitter function is the result of the Linux kernel randomizer with an amplitude of 100 This technique introduces an additional broadening of the width of the prompt peak by an additional 100 ps The result is a smoother gaussian 66 5 EXPERIMENTS 5 3 Timing Tests Parameter Average time resolution Atpwam ae 82 Average clock ee v Avg signal delay Le Avg signal delay time ps Table 5 Average timing resolution of the spectrometers electronic components and parametric time deviation in to of the acquisition system involving Channel 1 as a Start left column and as a Stop right column distribution under the sacrifice of some time resolution which for many experiments is negligible These results determine the maximum achievable time resolution of the PAC setup The average timing resolution of the measuring system lies well below Ins This result gives also evidence that the independent D
29. to the classical description the terms of order l 0 are the monopole terms l 1 dipole terms and l 2 the quadrupole terms Higher order terms are generally not regarded as their 3see Green s functions of the Laplace operator 21 2 5 Static Electric Quadrupole Interaction 2 PAC THEORY contributions are negligible Practically it is enough to assume the first three terms of the Hamiltonian Her 4nqgV 1 sel 2 1 QOV 4 2 te DO 53 m 2 The lowest order term contains V the Coulomb potential The order l 1 term vanishes due to the requirement of parity conservation under a space reflection The only term that needs to be retained in the electrostatic interaction is the quadrupole term The quadrupole interaction Hamiltonian is therefore An Hor er V 54 V is the quantum mechanical analogue to the classical electric field gradient oper ator Vaa The spherical components V can be expressed in terms of the cartesian derivatives Vzz V With the Coulomb potential 36 the tensor components of ee Vaa can be derived as follows or Ci 2 2 _ ey a per Ir nz PB RO 55 From 51 the relation with the spherical derivatives can be identified as 2 1 5 V Vee 0 4 V2 3 cos 6 1 1 5 A 3 Ne an iVyz 56 1 5 V3 1 ae Vyy i2Vzy 22 2 PAC THEORY 2 5 Static Electric Quadrupole Interaction From Eq 44 it is known that the non diagonal element
30. 0 8191 0 8191 Print this help Sample interval in units of s Default is ile 9 1 ns sample interval Set chi square value for sync pulse analysis Default is k200 94 A APPENDIX n lt n gt O lt n gt p lt n gt of s lt n gt S lt n gt th lt n gt to lt n gt ts lt n gt tw lt n gt v lt n gt w lt n gt A 4 Configuration files and Command line options 1 good sync detection 0 save always Save waveforms and segments only if good sync was detected Default is n1 Base line offset in Volts Default is 00 09 90mV Set process priority see man process for details Record a single sequence at once without binding to TCP socket and exit immediately Number of samples per segment Default is s1000 Number of segments per run Default is s4000 trigger level in of full scale value Default is t130 1 timeout when no events recorded 0 no timeout Default is tol 1 negative 0 positive Trigger slope Default is ts1 1 trigger detected 0 no trigger detected Save waveforms and segments only when trigger was detected Default is tw1 Full scale range in Volts Default is v0 2 200mV 0 save only segment entries after energy window check 1 save waveform as ASCII data along with segment entries 2 do not perform energy window check on data prior to saving b save in binary format Default is w0 95 A 4 Configuration files and Command lin
31. 07 REFERENCES REFERENCES 49 50 51 52 53 54 H Haas and D A Shirley Nuclear quadrupole interaction studies by perturbed angular correlations J Chem Phys 58 8 1973 3339 3355 W B Pearson The Crystal Chemistry and Physics of Metals and Alloys Wiley Interscience New York 1972 J A H da Jornada E R Fraga R P Livi and F C Zawislak Pressure Dependence of the Electric Field Gradient on Cd in Zn and Sn hosts Hyperfine Interactions 4 1978 589 593 M Winter Webelements Scholar Edition University of Sheffield www webelements com 2007 V B Brudanin PAC spectrometer for condensed matter investigation Nucl Inst Meth A 547 2 3 2005 389 399 M A Alves The Dynamics of Oxygen Vacancies in Zirconia An Analysis of PAC Data PhD thesis Oregon State University June 2003 108 Lebenslauf Name Geburtsdatum Geburtsort Eltern Staatsangeh rigkeit Familienstand Schulische Bildung 1978 1981 1981 1987 1987 1990 Mai 1993 Ausbildung 1990 1993 Mai 1993 Studium 1993 2000 Juni 1996 Januar 1999 Dezember 1999 April 2000 Arbeitsverhaltnisse August 2000 Marz 2001 seit April 2001 Christian Heinrich Otto Herden 04 Februar 1971 Ibbenb ren Deutschland Brigitte Herden geb Stiegemeier und Werner Herden deutsch verheiratet zwei Kinder Grundschule Westerkappeln Realschule Wester
32. 1 Time dependent anisotropy A22 effG22 t and the corresponding Fourier transform for In Cd in a iron metal foil at room temperature The data point resolution is 0 25 ns channel The sample shows magnetic interaction on two sites f and f The two obtained Larmor frequencies are w 558 8 Mrad s and wja 384 8 Mrad s The first fraction f represents Cd placed on substitutional defect free sites of Fe in the Fe host lattice 45 The second site may occur from remaining defects in the host lattice due to the ion implantation process The second fraction fa is very much suppressed in comparison to the main fraction It was reported that the presence of a second site results from nn defects however metallic samples generally show negligible radiation damage after low dose implantation treatment 46 The fairly high frequency PAC spectrum could be very well resolved with the new spectrometer setup over a wide time span The Larmor frequency w could be very well reproduced and deviates within an amount of 0 9 compared to the quadrupole coupling constant wg 553 4 Mrad s found in 37 43 6 2 4 Sample d Hf Ta in Hafnium metal As a hexagonal closed packed hcp group IV metal Hafnium is a noncubic metal In contrast to diamagnetic cubic metals a noncubic metal already exhibits quadrupole interaction if a probe atom occupies a substitutional lattice site with an unperturbed surrounding due to the intrinsic EFG The obtained
33. 1 202 0 876 1 284 1 This matrix shows for instance that detector 2 has overall the worst Start efficiency compared to all other detectors All non diagonal elements in column 2 are below an efficiency of 1 This represents that the efficiency of detector 2 as being a Start is lower than all the other detectors All l 2 are larger in magnitude than ka By definition of the multiplication of all anti diagonal elements must yield to 1 which is the case here 83 7 CONCLUSION 7 Conclusion The experiments conducted on the new spectrometer design verify that the acquisition system operates more than sufficiently for the intended purpose Major differences to conventional PAC spectrometers can be described as follows The setup cannot be classified into Slow Fast or Fast Fast systems anymore The information about the time of arrival and y ray energy is always recorded simultaneously There are no restrictions of which channel can act as the Start and which channel as the cor responding Stop All channels are peer to each other The small dead time allows the spectrometer to record any Y ray that enters a detector creating a snapshot of the nuclear decay over time Any information about the radioactive probe seen by a detector is thus stored on the harddrive An experiment can virtually be reconducted at any other time allowing successive alterations to parameters The results from the experiments conducted on several samples validate
34. 10 MHz timing signal The clock input is driven by a terminated 50 2 line It requires a voltage swing of at least 1V to operate properly The baseline DC offset is programmable The digitizers can be operated in a sequential storage mode In this sequence mode the digitizer s memories are divided into N segments Each segment contains n sample points With a sample interval t I ns and n 1000 data points per segment each segment yields a time period of 1000 ns On reception of a trigger event the sampling operation is started for a period of 1000 ns The digitizer will be ready to receive the next trigger within the next following t lt 500 ns 22 Taken the dead time tp segment size n and sampling rate into account the theoretical resulting maximum detector reception rate Rpetector yields at least 1 1 666 kHz 80 n ti tp 1000 10 9 s 500 10 9 s R Detector Figure 5 shows a plot of subsequent segments containing recorded pulses As already mentioned the amount of memory is limited to 8 MByte Therefore is the maximum number of segments S t N 8 MByte Ben Segments 81 1000 samples The acquisition stops when the maximum number of segments has been recorded The Pacslave process is responsible for downloading the entire memory buffer into 4The present DSP configuration forbids to split the memory into more than 8000 Segments This is a limitation that the DSP hardware puts up upon the system Future generations m
35. 1NO IA INI IX 1404 TAT Wad LSOH 92 A 3 Schematic diagrams A APPENDIX d8Z 00 90 SBBZ GT T red T T nasus zu JBAUP 49019 nad za9qunN usunDog NOUE ISOH 371LIL USPISH VENSUYI Aq umeig u pJ H LENSUYO 900Z F00z JWBUAdOD YSN YO sIIIEAIOD ENI e sr ELNO NJ anod ZNI zLNO ZNO Na LLANO 1LNO 8 NEWLSBE SI3I 93 A 4 Configuration files and Command line options A APPENDIX A 4 Configuration files and Command line options A 4 1 Pacslave Usage pacslave options Options a lt n gt b lt n gt c lt n gt cfd d lt n gt e lt n gt 1 anode 0 dynode PMT output polarity 1 negative anode signal 0 positive dynode signal Default is al 1 Bind to socket 0 No socket Open TCP socket for network communication Default is b1 Number of channels participating in measurement Default is c1234 all channels Example c24 only channels 2 and 4 Engage constant fraction discriminator Delay time of trigger in units of s Trigger point is shifted before or after signal crossing threshold Default is d 20e 9 20 ns Down scaler for energy integral Default is e4 Example e5 would divide energy integral by a factor of 5 f lt lowerstart gt lt upperstart gt lt lowerstop gt lt upperstop gt i lt n gt k lt n gt Energy windows thresholds Only save events with energies within lower and upper limits for start and stop Default is f
36. 40 3 HARDWARE 3 10 Photomultipliers Sample range V amp Aw N BoA 191 0 200 400 600 800 1000 timet ns Figure 7 Single transient pulse as recorded by the DP110 digitizer board In this sample segment the trigger delay is set to 20 107 s The DC base line is set to 115 0 mV The shaded area of 400 sample points signifies the minimum portion necessary for resonably good energy determination The left ordinate represents the sample points in signed 8 bit The right ordinate represents the sample points in mV and therefore appears as a negative pulse at the output Figure 7 shows a single transient pulse as it was recorded from the PMT anode For the conversion of higher energy Y rays into light pulses a scintillator crystal is mounted in front of the optical window of the PMT The two most common types of scintillator materials are BaF and Titanium doped Nal BaF detectors are used more often instead of NaI because of their higher timing resolution eventhough their energy resolution is 30 less than for Nal The PMTs used in this work are equipped with cylindrical BaF crystals with a size of 1 5 in diameter and length The tube s of type Hamamatsu R2059 operating voltage is rated between 2 0 kV and 3 0 kV 26 41 4 SOFTWARE A Software Software applications are the most vital part of this PAC spectrometer The spec trometer is controlled and operated by Host and Slave software applications The H
37. Conceptual Design and Setup of a Fully Digital Recording Apparatus for the Application in Perturbed Angular Correlation Spectroscopy Von der Fakult t f r Lebenswissenschaften der Technischen Universit t Carolo Wilhelmina zu Braunschweig zur Erlangung des Grades eines Doktors der Naturwissenschaften Dr rer nat genehmigte Dissertation von Christian H O Herden aus Ibbenb ren 1 Referentin oder Referent Professor Dr Klaus Dieter Becker 2 Referentin oder Referent Professor Dr Fred Jochen Litterst eingereicht am 12 12 2007 m ndliche Pr fung Disputation am 29 02 2008 Druckjahr 2008 Vorver ffentlichungen der Dissertation Teilergebnisse aus dieser Arbeit wurden mit Genehmigung der Fakult t f r Lebenswis senschaften vertreten durch den Mentor der Mentorin der Arbeit in folgenden Beitr gen vorab ver ffentlicht Publikationen Jens R der Christian H Herden John A Gardner and Klaus D Becker Fully Digital Time Differential Perturbed Angular Correlation TDPAC Spectrometer eingereicht 2007 an Nuclear Instruments and Methods Section A NIMA S 07 00089 Christian H Herden Mauro A Alves Klaus D Becker and John A Gardner A New Generation TDPAC Spectrometer Hyperfine Interactions 159 1 4 2004 379 383 Tagungsbeitrage Christian H Herden Mauro A Alves Klaus D Becker and John A Gardner A New Generation TDPAC Spectrometer Poster XIII International Conference
38. DC 50 Trigger slope negative Trigger delay s 20 10 PMT signal Anode Listening IP port 15000 Project path to be determined upon start Individual setting Description IP address 192 168 0 lt number of slave pc gt Trigger threshold Depending on PMT type and power setting Energy windows Individual if applicable Typically energy windows are set to filter y and 9 Table 4 A list of Pacslave s default settings isotopes decay which is available for analysis later at any given time The overwhelm ingly large amount of data has to be reduced in following data reduction steps For instance a typical PAC experiment for 8 hours with a 4 detector setups and 40 10 registered events per second gathers roughly 60 GB of data The Cocheck process applies Start and Stop windows and searches for segments that coincide with time and energy Random coincidences that do not fall within these windows are discarded for memory conservation purposes The raw coincidence data of the new collection method can contain count rates in any arbitrary time interval resolution These time interval widths are primarily limited by the DSP hardware Depending on the half life of the decaying process and strength of perturbation the time interval width are typically be set somewhere between 100 58 5 EXPERIMENTS 5 2 Data Processing N 10 t t i 250 200 150 100 50 0 50 100 150 200 250 t ns Figure 12 Time spectrum
39. SP clocks are acually running coherently It can thus be assumed with confidence that the synchronization circuit is working sufficiently Six coincidental distributions of three channel pairings are displayed in Figure 16 All combinations were measured with respect to Channel 1 The first group consists of three distributions solid black which represent the combinations where the times tamps of Channel 2 3 and 4 were subtracted from the timestamps of Channel 1 The second group consists of three distributions light gray where the timestamps of Channel 1 were subtracted from those of Channel 2 3 and 4 Within every group the center of the three coincidental distributions are displaced by an average amount of to12 0 13 and to 14 For the second group the time shifts are to21 to 31 and to 41 respectively Values for the average signal delays are given in Table 5 It is easy to see that the two groups of distributions are symmetrical around t 0 ps The knowledge of the time shifts between the detector combinations is important for the correct computation of the count rates Eq 89 67 5 4 PAC Sample Preparation 5 EXPERIMENTS Start 1 Stop 2 Start 1 Stop 3 Start 1 Stop 4 by yi i id Pri a u i Si li i ik pk 1 INI 2500 2000 1500 1000 500 0 500 1000 1500 2000 2500 t ps Figure 16 Time deviation in tp of the acquisition system involving Channel 1 as a Start black and as a Stop
40. The value of the anisotropy A contains nuclear parameters and was determined by other experiments in which their nature in nuclear physics were of primary interest For In Cd and Hf Ta those values are given in Tables 1 and 2 The experimentally measured anisotropy A22 ff however varies from the true value Several sources for causing this deviation have been listed as 19 e Finite detector size attenuation Ya e Finite time resolution p e Compton scattering Ye e Sample self absorption ya The effective anisotropy can coarsely be expressed by the true Ag by A2 eff YaYoYeYaAn2 97 The finite detector size attenuation Ya is a source that can be easily reconstructed theoretically The count ratio 94 is generally computed under the assumption of 62 5 EXPERIMENTS 5 3 Timing Tests a point source and infinte detector distance In practice the ideal condition is never fulfilled If the finite size of the sample and the finite detector distance is taken into account the angular correlation function 35 needs to be replaced by one that is a spatial average over the effective angles spanned by the source and the detector By inserting the term 93 into Eq 94 the R t becomes Ra 242C20 Pa 180 P3 90 34 Azaz P 180 2 Pa 90 inet 98 where Yq is the theoretical correction factor for the finite detector size effect Figure 14 shows the result of a calculation of y A22 for the radioactive
41. Uhrmacher and H Hofs ss Actual Concepts of Digital PAC Spectroscopy XIV International Conference on Hyperfine Interactions Iguazu Brazil August 2007 Poster Fortlaufende Projekte NSF DMI 0422218 SBIR Phase II Accessible Scalable Vector Graphic Authoring and Editing Applications Rolle Beteiligter Wissenschaftler 15 August 2004 31 Juli 2006 Phase IIB 1 August 2006 31 Juli 2008 NIH 2R44EY015968 02 SBIR Phase II High Speed Tactile Graphics and Braille Embosser Rolle Federf hrender Wissenschaftler 1 Juni 2006 31 Mai 2008 Subcontract on NIH 1R21EY016799 01 Penn State Phase 1 Testing Polymer Braille Display Properties Rolle Teilnehmer 1 August 2005 31 Juli 2007 Abgeschlossene Projekte NIH 2R44EY014481 01 SBIR Phase I The Accessible Graphics Reader for Blind People Rolle Beteiligter Wissenschaftler September 30 2002 March 29 2003 NIH RA3EY015968 01 SBIR Phase I High Speed Tactile Graphics and Braille Em bosser Rolle Federf hrender Wissenschaftler 1 September 2004 28 Februar 2005 NSF DMI 0231691 SBIR Phase I Accessible Scalable Vector Graphic Authoring and Editing Applications Rolle Beteiligt 1 Januar 2003 30 September 2003 NIH 2R44EY014481 02 SBIR Phase II The Accessible Graphics Reader for Blind People Rolle Beteiligt 1 Dezember 2003 30 November 2005 NIH 1R43EY017238 01 SBIR Phase I Accessible Graphical Display fo
42. WindowUpper 8191 stopWindowLower 0 stopWindowUpper 8191 CFD delay time in ns cfdDelay 5 98 A APPENDIX A 4 Configuration files and Command line options A 4 3 Pachost Usage pachost lt firstsequence gt lt lastsequence gt nsequence options or pachost lt command gt lt firstsequence gt Sequence number to start measurement with lt lastsequence gt Sequence number to end measurement with lt nsequence gt Number of sequences to record Overwrites lt lastsequence gt with lt firstsequence gt lt nsequence gt Options to configure pacslave over a TCP IP connection e lt n gt Down scaler for energy integral Default is e4 Example e5 divides energy integral by a factor of 5 n lt n gt 1 good sync detection 0 save always Save waveforms and segments only if good sync was detected Default is n1 s lt n gt Number of samples per segment Default is s1000 S lt n gt Number of segments per run Default is s4000 tw lt n gt 1 trigger detected 0 no trigger detected Save waveforms and segments only when trigger was detected Default is tw1 w lt n gt 0 save only segment entries after energy window check 1 save waveform as ASCII data along with segment entries 2 do not perform energy window check on data prior to saving Default is w0 wb Save in binary format Ws Save sync pulse after recording sequence Measurement options c lt n gt 99 A 4 Configuration files and Command line
43. ability to other systems A number of test measurements were performed to verify the proper operation of this new time of flight acquisition system Some questions that need to be answered are whether the new apparatus s timing resolution will be sufficient for short lived inter mediate states 71 lt 10 ns and if the new system will at least reach the performance of conventional PAC spectrometers Reset times in traditional Time to Amplitude Converters are typically between 1 us and 50 us which limits the detection of repete tive signals in the worst case to 25 kHz 27 28 The timing tests on the new PAC 2 1 INTRODUCTION setup will also show the achievable quality of the apparatus s timing resolution These general test measurements were succeeded by real PAC measurements on com pounds that have been studied previously Their properties are well known and un derstood The isotopes In Cd and Hf Ta are the most popular radioactive probe atoms used in PAC spectroscopy and therefore a vast number of publications about compounds with these isotopes are available The PAC samples were prepared according to the published methods to insure the repeatability of the previous results The samples were picked under the aspect of that ease to identify spectra If the ac quisition system would not perform as expected the data would show evidence of misbehaviour immediately For instance attenuation effects can often be attributed to
44. akes on the form W 6 t bj cos kwrt 79 k Figure 2 shows the splitting of the energy levels of a nucleus with angular momentum 3 for both magnetic and quadrupole interaction The removal of the degeneracy due to magnetic interaction is portrayed on the left side the electric on the right The splitting of the energies under the magnetic effect is also known as the nuclear Zeeman effect 12 27 3 HARDWARE 3 Hardware 3 1 Overview of Setup and Operation Conventional TDPAC spectrometer can basically be distinguished into slow fast and fast fast systems In both systems the timing signal is derived from the anode of the Secondary Electron Multiplier SEM tube The energy signal is typically taken from the PMT dynode The dynode signal is amplified and shaped by Linear Amplifier LA A scintillator crystal attached to the SEM s optical cathode window will con vert the high energy y into ultraviolet light The flashes enter through the optical window and strike out electrons from the negative cathode Those electrons are be ing multiplied over several cascades until the avalanche reaches the positive anode A combination of scintillator crystal and SEM is also referred to as a Photo Multiplier Tube PMT Typically in slow fast systems Nal crystals are being used for better energy but less timing resolution In fast fast systems BaF crystals are being used for faster timing but less energy resolution In slow fast systems the t
45. all core parameters that 52 1 complies Secondly the integrity of the quadrupole frequencies w3 w wa has to be regarded For magnetic interaction the perturbation function can contain the fundamental Lar mor frequency wz as well as higher harmonics k wz The core parameters bo for magnetic interaction yield more or less to the statistical information about the orien tation of the magnetic flux The resulting anisotropies obtained for samples a d shall be discussed in the following 6 2 1 Sample a In Cd in 3 Sn metal Tin is a non cubic highly crystalline metal Below 13 2 C it exists as a Sn which has cubic crystal structure At room temperature the crystal structure is tetragonal 50 Due to tetragonal crystal structure the V component of the electric field gradient is non zero In general is it difficult to predict where the radioactive probe is integrated 76 6 RESULTS 6 2 PAC measurements Fourier intensity 0 100 200 300 400 0 0 0 1 0 2 t ns o 10 rad s Figure 19 Time dependent anisotropy A22 effG22 t and the corresponding Fourier transform for In Cd in Sn metal The data point resolution is 0 25 ns channel into the host lattice The characteristics of the obtained time dependent anisotropy A22 effG22 t for sample a were f
46. an 3 01 41 58 2006 TEMPERATURE 38 SEGMENT 0 5834283940 0 439 1415 50 4 SOFTWARE 4 3 Host applications SEGMENT 1 9791467276 913 775 1755 SEGMENT 2 12821855718 961 217 1172 SEGMENT 3 15789855866 948 365 1320 SEGMENT 4 where the interpretation of the labels is WAVEFORM lt pac channel gt lt sequence gt lt segments gt DATE lt unix time gt lt date and time string gt TEMPERATURE lt degree C gt SEGMENT lt n gt lt timestamp gt lt energy gt lt horPos gt lt cfd value gt The first line designates the file type The WAVEFORM label signifies that this is a sequence file It has three arguments The first lt pac channel gt entry is a zero based index of the PAC channel number The second entry is the sequence number The third argument is the total number of segments this file contains The DATE label represents the Unix time as an integer value and as a string TEMPERATURE gives the current temperature of the DSP board in units of C The parameters that were obtained from the analysis are kept under the SEGMENT label This is mainly a 64 bit wide timestamp in units of ps The energy integral is an arbitrary value In the first SEGMENT line the energy value has a special meaning The timestamp identifies a valid synchronization pulse when it is equal to zero A non zero value represents an invalid entry 4 3 Host applications 4 3 1 Pachost The host application Pachost is a helper process w
47. an example we observe the decay scheme 2 3 in In Cd as dis 5 2 1 Theory of Angular Correlation 2 PAC THEORY played in Table 1 For all transistions the parity never changes m m 1 The multipolarity order of the first transition is of order l 1 which accounts for pure dipole radiation character In general M 1 transitions versus E l transitions can be neglected because transition probabilities are reduced by orders of magnitude 6 Whereas E 1 versus M are usually of comparable magnitude Therefore the character for y in In Cd is predominantly M 1 E 2 The second transition of order l 2 has quadrupole character E 2 In the near vicinity around the nucleus the y radiation is exposed to multipole fields In the far field a y ray s nature is an electromagnetic wave and thus can be described by a set of coupled wave equations which originate from a generalized vector potential A that satisfy Maxwell s equations The transverse electric and magnetic solutions of Maxwell s equations in regard to a general multipole vector potential can be expressed by 1 1 B j kr n LY 0 pje t E i l J kr ES r p l CY x5r B 3 1 r l gt E j kr LY 0 pje Bi i Vx E M l jkr T 1 0 9 l Ee l M where j kr are spherical Bessel functions Yp 0 p is the spherical harmonic func tion and L ih r x V is the angular momentum operator With help of the
48. arrival 29 3 1 Overview of Setup and Operation 3 HARDWARE independent of its signal height The time delay between Start and Stop events is processed in a Time to Amplitude Converter TAC Sorting into channels is per formed by Analog to Digital Conversion ADC of the TAC output signal The ADC can be read out by a microcomputer at a later time Conventional PAC spectrometer designs require a significant amount of custom electronic circuits and complex wiring of all their components The goal within the context of this work was to keep the design of anew TDPAC spec trometer as simple and less laborious as possible Like in fast fast systems both the timing and energy information are available at the same time But unlike conventional equipment this information is gathered from only the detector s anode output A great deal of flexibility is introduced by using personal computer systems in combination with state of the art Digital Signal Processing DSP technology The PMT s anode output is being sampled digitally at high speed at all times The main focus during the development of this new spectrometer setup was the reduction of the complexity of the electronic components Over the last few years DSP technology has evolved and analog to digital recording speeds have increased tremendously Recorderboards with sampling rates of 10 samples per second or more have become available Thus they are applicable for recording fast transient PMT s
49. bling me to continue to work on the PAC project even from a remote distance Then I also would like to express my acknowledgements to the following people I worked with during the last couple of years Dipl Chem Jens R der for his engagement in this project and his help in promoting the work in Braunschweig even without my presence For sharing his knowledge in chemistry science and being a resourceful partner in technical conversations I appre ciated the vivid discussions about security issues in modern information technology and conspiracy theories Dr Mauro A Alves for sharing his thoughts about PAC theory with me performing some of the original PAC measurements with the new spectrometer setup and for preparing and printing the conference materials Dr Roland Platzer for introducing me to John Gardner in the first place and also for his input and assistance on the issues regarding photomultiplier tubes Dr Robert L Rasera for discussing the issues about data reduction methods used in PAC spectroscopy and suggestions for publishers Dr Matthew O Zacate for the preparation of the Tin sample and discussing the results Dr Ken S Krane for supplying the activated Hafnium sample I also acknowledge the DFG for financially supporting this work under the grant application SPP 1136 Substitutionseffekte in ionischen Festk rpern Finally I want to thank my wife Gaby for her encouragement and support during the creation of thi
50. can be observed The nucleus decays from a single excited state J M 7 into the ground state My Tf I M and 7 are the quantum numbers of the quantum mechanical states of the system The decay process involves two successively occuring particles 71 72 denoted through the cascade I Ma m gt I M r 2 Ir My ap I M T is an intermediate level with a finite lifetime ry In quantum mechanical systems the characteristics of the levels are described by their well defined angular momentum I and parity m Transitions are treated as changes in quantum numbers h m m for y and l2 M2 T2 for y2 In this work only those transitions which result in the emission of y radiation are of interest During a transition from a higher state into lower state typically y radiation is emitted with respect to the conservation of momentum and energy Therefore the emitted y radiation is classified according to the changes in nuclear parity and angular momentum The parities 71 2 of the y rays are related to the level parities by Term Mean 1 The angular momentums l 2 can have a range of values given by fg I lt h lt t il I I sh lt I 2 The y radiation is classified in E l radiation by m 1 and M l radiation by m 1 For different orders of the multipolarities Io 1 2 3 the lower modes of radiation can be summarized as Parity change T 1 l 2 l 3 yes E 1 M 2 E 3 no M 1 E 2 M 3 As
51. ch MUX to synchronize x Handshake r u Raise CHANNEL_SELECT CHANNEL SELECT high Yes Y Start digitizers y Slaves busy Handshake Raise BUSY BUSY low y No Analysis and storage of waveform data Y Release synchronization pulse y Updating parameters Y Switch MUX to PMTs Figure 10 Flow chart for the PAC processes during the recording cycles of the PAC spectrometer Depicted in the left column is the program flow for the Pachost process on the right the program flow for Pacslave The arrows in the center denote the inter communication messaging between those two processes 52 4 SOFTWARE 4 3 Host applications The first argument specifies the number of the first sequence file The second ar gument specifies the number of the last sequence file The measurement will stop immediately after the last sequence file was written by Pacslave Pachost will exit after completion of the last recording cycle 4 3 2 Cocheck For each PAC Slave exists a set of sequence files containing segments with times tamp and energy information The sequence files must be analyzed for coincidences Cocheck is the next instance in the series of PAC helper applications that puts the recorded data through energy and coincidence filters All events are divided into Start and Stop events according to the energy window settings The energy values are arbi trary integral numbers that are usually sorted into arrays which th
52. ction of PAC fitting libraries 86 A APPENDIX A Appendix A 1 Parallel port pin assignment Pin No SPP signal Direction Register Hardware D 25 inverted STROBE Control DATAO Data DATA1 Data DATA2 Data DATA3 Data DATA4 Data DATA5 Data DATA6 Data DATA7 Data ACKN Status BUSY Status PAPER_OUT Status SELECT Status AUTOFEED Control FAULT Status INIT Control SELECT_IN Control Ground ZZZ ZZZ ZZ Z x COO OOOO oO 8 3 4 5 6 T 8 9 10 11 F meme Aa Ww N Z O rm On mm ID Table 7 Pin assignment of the D Type 25 pin standard parallel port connector 87 A 2 Parallel port I O registers A APPENDIX A 2 Parallel port I O registers o En port Write only Nam Read Write Property Base 1 DEE port Read only BUSY i ACKN i Paper Out i Select In j Error i Irq i Reserved i Reserved nn 2 a port Read Write Unused j Unused j Enable Bi Dir port j Enable Irq i Select printer i nit printer j Auto linefeed i Strobe Table 8 Bit assignment of the Standard Parallel Port I O register 29 88 A APPENDIX A 3 Schematic diagrams A 3 Schematic diagrams A 3 1 Wiring Plan Figure A 3 1 Overview of the wiring between the spectrometer components A 3 2 Analog Multiplexer Module Figure A 3 2 Schematic diagram of the analog multiplexer module The center piece is an integrated high speed multiplexer circuit U4 with 2 analog inputs Its input IN2 is dir
53. ded after the analysis This saves quite a bit of harddisk space Typically the harddisk space of a slave is filling up according to the trigger repetition at a rate of several thousands per seconds 3 7 Analog multiplexer module The center piece of the analog multiplexer module is an AD8180 from Analog Devices The AD8180 is a high speed 2 to 1 integrated analog multiplexer circuit The func tional block diagram is shown in Figure 6 It is designed as a high speed analog signal switch and offers a 3 dB signal bandwith for frequencies greater than 750 MHz The crosstalk between the two channels is specified to be as low as 80 dB 23 Switching happens upon the change of the logic input level at its SELECT input When SE LECT is logic low the common output AOUT equals the analog input IN1 otherwise AOUT equals IN2 The MUX is necessary for synchronizing the timestamps of all Slaves This is achieved by switching the MUX s first input IN1 to its output AOUT As shown in Figure A 3 2 in Appendix A 3 IN1 is connected to the synchronization module output The two resistors R7 and R11 and the capacitance C6 form a deriva tive element which transforms the incoming logic pulse edge from U2 into a short pulse Due to extreme precise timing requirements the pulse edge is required to be as steep as possible Therefore the choice of the relatively small pull down resistor value R11 25 Q and the small capacitance C6 47 pF seemed appropriate Duri
54. e Hamiltonians H and H are Mal Hy m and m Ho m 17 These amplitudes describe the probability of the population of the spin states For no extranuclear perturbation the intermediate spin levels are m m The temporal change of the intermediate states through the time evolution operator can be formally t ma Kim mi A t Ima 18 described as The expression in parentheses represents the unit operator a superposition of all pure intermediate eigenstates The angular correlation equation for the perturbed case rises from the combination of the probability amplitudes of the first transition and the perturbed probability amplitude of the second transition The matrix elements are ma Hi m and m HoA t Ma 19 The probability density function of the emission of the two successive transitions in directions ky and ko yields to the expression 2 Whi kot X J mAH ma mal Hi ms 20 my mi Ma Inserting Eq 18 into Eq 20 one can calculate a more explicit expression of W k ke t With the absolute square value executed and some summation terms 13 2 2 Perturbation of Angular Correlation 2 PAC THEORY rearranged the probability density can be rewritten as W ki ka t gt 17 p Ha mo mij A t Ima mal Him 5 1 1 M Ff Mi Ma Ma Mb My m Holm ml A m mal H m 21 The impact of the perturbation displays itself only in the two separated bra
55. e of a nonstatic fluctuating EFG The effective anisotropy 22eff was determined as Aaa erp 0 22 for a detector distance of d 10 cm The PAC frequencies are in good agreement with the quadrupole coupling constants wg 50 9 7 Mrad s and asymmetry 7 0 24 previously published for 99 at 8 Ta in HfZr alloy 39 42 82 6 RESULTS 6 3 Detector Start Stop Efficiencies 6 3 Detector Start Stop Efficiencies So far the analog portion of this fully digital acquisition system has not yet been discussed much The analog part of this apparatus consists mostly of the PMT and their high voltage power supplies As mentioned earlier there are several unpre dictable sources for unwanted effects in PMTs Thermal electrons randomly excite and multiply causing unwanted trigger events which affect the efficiency of the sys tem Physical damage to the PMTs or their shields are additional sources for detector activity For practical reasons it is important to know how well the PMT operate in this otherwise very deterministic environment The detector Start Stop efficiencies have been extracted using the raw count rates N 6 t obtained by the measurements mentioned in the previous sections The count rates were combined as described in Section 5 2 according to Eq 91 and 92 For those measurements the elements of for the current set of detectors were determined as 1 0 722 1 075 0 836 139 1 1 484 1 145 105 Kij g 0 948 0 677 1 0 784
56. e options A APPENDIX 96 A APPENDIX A 4 Configuration files and Command line options A 4 2 Pacslave conf pacslave config file Author Christian Herden Date O7 Sep 2004 Channel number This Slave is Ch 1 channelNbr 1 Relative or absolute path to data This is where pacslave will store sequence files Do not attach to this path segPath home data lt project name gt Name attachment of energy histogram file outBinArray energy dat Reference file for identification of sync pulse syncPulse syncpulse dat Default number of segments nbrSegments 8000 Default number of samples nbrSamples 400 Allow host IP address hostIP any Specify host port hostPort 15000 Bind pacslave program to the host This options does not execute any measurement unless the host connects bindToSocket 1 Anode or dynode signal of PMT anodeSignal 1 Default trigger slope 1 negative 0 positive trigSlope 1 97 A 4 Configuration files and Command line options A APPENDIX Full scale in V 256 bit resolution over a default of 200mV fullScale 0 2 Vertical offset 90mV vert ffset 0 09 Trigger level in of full scale 25 trigLevel 26 trigLevel 25 Energy scaler energyScale 4 Input coupling 3 500hm DC 1 1MOhm DC inputCoupling 3 trigger delay time in seconds delayTime 20e 9 Energy window limitations are applied prior to storage startWindowLower 0 start
57. ectly connected with the PMT signal output IN1 is input to the synchronization pulse which is directed through the derivative element C6 R7 R11 from the logic trigger input at U2 A 3 3 Synchronization and Channel Select Module Figure A 3 3 Schematic diagram of the Synchronization and Channel Select fanout module Two quadruple integrated 100 Q line driver circuits provide the necessary impedance conversion from the logic TTL output to the BNC output A 3 4 Reference clock module Figure A 3 4 Schematic diagram of the Reference clock driver module One inte grated quad 100 Q line driver converts the oscillators TTL output to a 50 Q coaxial output 89 A 3 Schematic diagrams A APPENDIX Parallel port Slave Computer 2 D SUB 25 connector DP110 digitizer card Analog Input External Clock Input Ethernet Adapter RJ45 jack Slave Computer 3 Parallel port D SUB 25 connector DP110 digitizer card Analog Input External Clock Input Ethernet Adapter RJ4S jack Slave Computer 4 Parallel port D SUB 25 connector DP110 digitizer card Analog Input External Clock Input Ethernet Adapter RJ45 jack PMT 4 G56 Analog Multiplexer 2 RGSB HV Power Supply Power Suppi D SUB 25
58. ectrometer setup it has been replaced by pulse analysis routines implemented into Pacslave The source code was developed to utilize the anode signal of the PMT to perform energy and time extraction The anode signal can be separated into two fractions the fast and the slow component 20 21 It was found that the slope of the fast component of the anode signal had a rise time of about 3 ns and a decay time of 4 5 ns and could be expressed through a Gaussian distribution The shape of the fast component was found to be extremely suitable for the CFD technique Figure 8 46 4 SOFTWARE 4 2 Slave applications displays in detail how the technique is applied to the waveform The PMT pulse is recorded with a pre delay of tg 20 107 s This moves the trigger point well into the recording window so that the CFD technique can be applied The fast component is first inverted divided in half and added to the original fast component data by an offset of less than the FWHM of the fast component The zero crossing point tefa is independent of the original signal height tf can now be correctly appraised under consideration of the timestamp ts the horizontal position parameter tyorPos and the zero crossing point tefa by the formula lof tts thorPos tefa 85 The horizontal position parameter tnorPos is a correction factor for the DSP horizontal parameters This value is provided by the DSP hardware Trigger events usually occur asynchronou
59. eed of 133 MByte s it takes approximately ttrans 0 024 s to download the waveform data into the PCs memory One can conclude that the spectrometers dead time is as low as the digitizer hardware and the PCI bus speed would possibly allow it In this experiment the segment length was set to 400 ns The energy integral was calculated from 400 sample points A relative dissolution of energy of 16 5 was acheived Using a longer segment length was thought to increase the energy reso lution More sample points might have put more weight on the integral But no significant improvement could be detected even up to 1000 sample points per seg ment Depending on the calibration of the photomultiplier tubes the anode signal decays within the first 400 500 ns Choosing a larger segment length does not gain significantly more information about the energy For the Cocheck process the energy window width was set to twice the FWHM of the energy as shown in Figure 11 in Section 4 3 2 as the red and green shaded areas The total amount of coincidences gathered in this experiment was 183 7 10 in 12 detector combinations and a co incidence window with of 8 us Thereof the coincidence efficiency results in 3 6 The starting coincidence rate was 1 5 kCo s for an activity of 0 315 MBq and en 73 6 1 Performance 6 RESULTS Duty cycle Sample activity MBq Figure 18 The spectrometers duty cycle over the detected sample activity Data point
60. en reflect the energy spectrum Figure 11 shows three typical energy spectra for Hf Ta In Cd and 22Na as recorded by the PAC spectrometer The energy windows for Start and Stop are depicted as shaded areas Segments with energies that fall beyond the energy windows are discarded The remaining segments are put through a conincidence perceiving routine which searches for coincidences in every PAC Slave combination Options and the configuration file are listed in Appendix A 4 5 The coincidence algorithm is a nested loop that cycles through all Start Stop per mutations to create the time spectra D t where i is the index of the Start channel and 7 the index of the Stop channel The algorithm will skip combinations of i j Therefore the total number of time spectra created is m n n 1 with n num ber of PAC channels For a 4 channel PAC spectrometer this yields to 12 different time spectra or 30 time spectra for a 6 channel setup With N average number of segments the complexity of the algorithm can be classified as O NY The common timing reference between all PAC Slaves is the timestamp of the first segment In general the timestamps of the DSP cards have no relationship with respect to each other Though running coherently the DSP internal clocks are running independently The reason for the Host inserting the synchronization pulse is the provision of a common timebase Every Slave is recording the synchronization pulse fir
61. ent Dopants and Native Defects in Group IV Oxides Indium and Cadmium in Ceria and Zirconia PhD Thesis Oregon State University 1997 M O Zacate by correspondence with author R L Rasera T Butz A Vasquez H Ernst G K Shenoy B D Dunlap R C Reno and G Schmidt Strength symmetry and distribution of electric quadrupole interactions at Ta impurities in hafnium zirconium alloys J Phys F 8 1978 1579 1589 R Hanada Oxidation of iron studied by PAC and CEMS Hyperfine Interactions 120 121 1999 539 544 P Wodniecki B Wodniecka A Kulinska K P Lieb M Neubauer and M Uhrma cher Indium solubility in iron studied with perturbed angular correlations Hyperfine Interactions 120 121 1999 433 437 M Neubauer K P Lieb P Schaaf and M Uhrmacher Ion beam mixing of Ag Fe and In Fe layers studied by hyperfine techniques Phys Rev B 53 1996 10237 10243 M Uhrmacher A Kulinska Yu V Baldokhin V V Tcherdyntsev S D Kaloshkin A Maddalena and G Principi Hyperfine Study on Mechanically Alloyed Fe Mn Systems Hyperfine Interactions 136 137 2001 327 332 J R der C Herden J A Gardner and K D Becker A Fully Digital Time Differ ential Perturbed Angular Correlation TDPAC Spectrometer submitted to NIMA S 07 00089 J Christiansen Hyperfine Interactions of Radioactive Nuclei Springer Verlag Berlin Heidelberg New York Tokyo 1983 273ff 1
62. er and the next sampling clock varies randomly in time The true time reference lies within a fraction of the sampling clock The horizontal position parameter permits a very precise positioning of the acquired trace and is necessary for the correct interpretation of the CFD s resulting parameter The most precise trigger time is determined when the sampling clock is at maximum resolution For the DP110 this time interval ist 1 ns 3 5 Host computer The host computer is a standard state of the art Personal Computer PC The host system is equipped with an ASUS motherboard Parallel port serial port USB PCI a 3 2 GHz Intel hyperthreaded microprocessor 512 MB of RAM a 1 Gbit network card and 80 GB harddrive The main function of the host computer s hardware is to control all participating slave computers This is done through a series of status lines that lead from each slave computer to the host computer s parallel port as the parallel port offers some convenient way to use its logic inputs Further since all machines are linked together by a local network switch the host is able to submit commands to and query settings from the slave over a TCP connection The host computer switches all analog MUX boards to the sync pulse input channel Then it sends a signal to each slave to prepare their recorderboards for a measurement cycle by switching its outgoing status lines It waits until all slave computers respond with an acknowledge sig
63. erage background count is very important for the determination of the count ratio function 94 In general 75 6 2 PAC measurements 6 RESULTS the time spectra need to be both background and time corrected According to Eq 89 the background corrected N t is obtained by subtraction of the background count The random background data is available for a large time span so it can be easily fitted to a linear approximation With the help of the channel delay times from Table 5 the time corrected N t can be determined The delay times can be used to correct the time axis to positive or negative times respectively Thus the count ratio 94 can be built from the obtained time and background corrected time spectra N t The characteristics of the obtained count ratios R t were all fitted according to Eq 34 for electric quadrupoles and Eq 79 for magnetic dipoles Practical more general and simplified versions of Eqs 34 and 79 suitable for fitting the data under the assumption of a static non fluctuating EFG or intrinsic B field and I 5 2 are Rolt 3 Aeff fiX San tn 103 i 1 n 0 kmaz Ru t Aeff ss basco 104 i 1 k 0 even while the fractions f were introduced considering the fact that different sites in the lattice may be occupied by the nucleus Depending on the type of interaction multiple constraints can be inflicted on the fitting function Firstly for quadrupole interaction it is essential for
64. eration of the PAC spectrometer is sectioned into recording cycles During each cycle the DSP boards are triggered by pulses of the PMT anode When the DSP memory has filled up the Slaves are forced to transfer the recorded waveforms into the main memory of the Slave A new recording cycle is initiated immediately while the waveforms are being analyzed by the Slave A smooth operation is arranged by the Host over a handshake bus system A simple control network was created using the parallel port s in and output capabilities The Slaves need to report their state of recording to the Host computer and the Host has to inform the Slaves to prepare their recorderboards The principle of handshaking is illustrated in Figure 4 During the operation all Slaves are supervised and controlled by the Host The Host cannot initiate another recording cycle as long as one of the Slaves is reporting BUSY The Host starts a cycle by deasserting the CHANNEL_SELECT line This action disconnects the PMT from the digitizer input All Slaves prepare their DP110 for the upcoming recording cycle It also informs the Slaves to prepare for the next recording cycle As long as CHANNEL_SELECT is held low no PMT signal are routed through By asserting their BUSY lines each Slave will notify the host that the preparation of its DSP board is complete The Host will release a synchronization pulse after the last Slave has reported not being BUSY The Host asserts the CHANNEL_SELECT li
65. ergmann C Schafer and H Niedrig Lehrbuch der Experimentalphysik Optik 3 deGruyter 1993 2004 S Unterricker M Dietrich G Bohm L Pasemann A Moller R Vianden M Deicher R Magerle A Burchard and W Pfeiffer Quadrupole interaction in chalcopyrite structure semiconductors Hyperfine Interactions C 1 1996 238 241 S Unterricker M Dietrich A M ller R Vianden M Deicher R Magerle W Pfeif fer G B hm and L Pasemann PAC investigation of ternary semiconductors with chalcopyrite structure Cryst Res Technol 31 1996 761 768 106 REFERENCES REFERENCES 35 36 45 46 47 48 S Unterricker T Butz and W Tr ger TDPAC investigations of the Cd quadrupole interaction in ternary chalcopyrite semiconductors Hyperfine Interac tions 62 4 1991 373 377 S Unterricker and F Schneider The EFG lattice structure dependence of Cd on A sites in A II B IV C V 2 semiconductors Hyperfine Interactions 15 16 1983 827 830 S Unterricker V Samokhvalov F Schneider M Dietrich and THE ISOLDE COL LABORATION Magnetic Hyperfine Interaction of a Cubic Defect in a Iron Hy perfine Interactions 158 1 4 2005 229 233 A M ller Doktorarbeit TU Bergakademie Freiberg 1997 R Vianden Electric field gradients in metals Hyperfine Interactions 16 1983 1081 1120 M O Zacate A Microscopic Study of the Interaction Between Alioval
66. evice driver 2 222 22 4 1 2 Using the device driver from the console 4 1 3 Using the device driver in a C C application 4 2 Slave applications 1 3 2 2 4 Sie ae da 4 2 1 Pacslave wi ge Mn a ae fa eat a a end 4 2 2 Pacslave options 2 3 2 ee ein are 42 3 Pacslave file format 4 a 0 82 ease ew a 4 3 Host applieationss e lt Ka 2 SE ae ath ei re Fe 43 1 Pachost uses eos 28 Ber Ae Bent na CONTENTS CONTENTS CONTENTS 232 Woche suis anne oP oof Fs are ee area 53 4 3 3 Vocheck file format 2 2 82 eit ee bn Se ra 55 4 4 Network File System 4 au Sig eee nn Bee re 56 5 Experiments 57 5 1 Host and Slave Preparation 2 020004 57 92 MATa Procesin a a ea eee hr 57 pio EINER Test re eed ee ee Se en ae eh 63 5 4 PAC Sample Preparation Am ac hw gach ht Bok Me ate rad 68 6 Results 71 6 1 Performancen euer Oe a a ral 6 2 PAC measurements 044 3 6499 44 8444952469 hai RS 75 6 2 1 Sample a In Cd E 8 Sn metal wh 2 lew eee eee 76 6 2 2 Sample b In Cd in Cadmium Silicon Phosphide 78 6 2 3 Sample c n Cd in Iron toil a s 2er othe et 80 6 2 4 Sample d Hf Ta in Hafnium metal 81 6 3 Detector Start Stop Efficiencies rar sn 8 a bee ae ES 83 7 Conclusion 85 A Appendix 87 A 1 Parallel port pin assignment ser ra SR oe ES 87 A 2 Parallel port I O registers 1 24 22 2 22 See Bee Prada 88 A 3 Schematic diagrams Akg hk ee ee Bae ee hae ees 89 cord
67. f the spectrometer but also to certify that the synchronization mechanism is working accordingly During the measurement the timestamps were recorded Later the coincidence check process subtracted all the recorded timestamps for all permutations of channels i and j Figure 15 a and 15 b show two prompt peaks that represent the achievable timing resolution The data points in Figure 15 a show a gaussian distribution and also exibit an interference pattern which results from the superposition of two timestamp counters driven by two nearly coherent running clocks This behaviour can be explained for mally like a double slit experiment where two coherent wave fronts interfere with each other 32 Each clock can be adjudicated a frequency v and a phase y The 64 5 EXPERIMENTS 8 3 Timing Tests a N a u 2200 2000 1800 1600 1400 1200 1000 t 10 b N a u 2200 2000 1800 1600 1400 1200 1000 t 10s Figure 15 These two prompt peaks represent the achievable timing resolution of the bare electronic acquisition system between two PAC channels a shows the peak without the application of CFD and jitter The data points o in a were fitted according to Eq 100 with Atpw m 2368 ps b shows the prompt peak with both jitter and CFD applied 65 5 3 Timing Tests 5 EXPERIMENTS granularity of the DSP internal clocks are specified by the manufacturer at 80 ps The inter
68. for PAC spectroscopy For the new hardware it is not necessary to perform any timing calibration as it is known from conven tional spectrometers The spectrometer depends on the reliability of the DSP clocks The maximum time resolution of the electronic system depends on the stability of the DSP clocks Therefore it is necessary to have knowledge of the quality of the 63 5 3 Timing Tests 5 EXPERIMENTS 0 04 O Extracted R der Herden ADiluted sample Alves 0 06 R der Herden Jaeger MC Point Sample 0 08 Su Collins 0 10 lt 0 12 0 14 0 16 0 18 Detector distance d cm Figure 14 Effective anisotropy A2 ff in relationship with the detector distance The data points marked as o represent the obtained effective anisotropies A22 ff from various measurements on polycrystalline samples The data points marked as A were obtained from anisotropy measurements with diluted samples The dashed line just guides the eye As a comparison the plot shows two results from Monte Carlo calculations done by 16 and and empirical equation by 18 synchronization system The DSP clocks are synchronized by a 10 MHz external clock signal which is distributed from a common source to all channels In order to test the performance of the timing circuit the output of an external frequency gen erator was hooked up to all the MUX s PMT inputs The frequency generator was primarily used to test the performance o
69. fter the emission of 7 and Ya No perturbations of the intermediate sublevels were regarded in this case The following section will discuss extranuclear effects on the substates 2 2 Perturbation of Angular Correlation The derivation of PAC theory presented in this section follows the introduction of perturbation yy angular correlation of Schatz and Weidinger 3 In general the theory describes the interaction of an intermediate nuclear state of a yy cascade with a perturbing potential The intermediate state m has a finite life time Ty During this time the nucleus is exposed to hyperfine fields causing a realignment of the magnetic substates One can describe the perturbation by a unitary time evolution operator A t that expresses the quadrupole interaction on the intermediate substates m with time It satisfies the time dependent Schrodinger equation nS A t HorAt 15 Hg is the perturbing Hamiltonian and is independent of time for static interaction In this case one solution of Eq 15 the change of eigenstates is expressed by the 12 2 PAC THEORY 2 2 Perturbation of Angular Correlation complex phase A t eiet 16 The population of the m states depends on the initial transition The interac tion of the time evolution operator A t on the intermediate states m causes a re population into states m A t m from which the final transition into the ground states occur The matrix elements of th
70. g This allows a calibration of the energy windows on the fly For the future it is conceivable that these energy windows are being located automatically by software when the type of nucleus is chosen by the operator Calibration parameters for each radioactive probe nuclei can be pulled out of a system database when needed There have also been advances in DSP technology since the first set of recorder boards for this work were acquired More recent versions feature more on board memory and higher sampling rates The use of more memory would reduce dead times Improvements are also expected from using a more recent generation computer with fast PCI Bus technology and higher processing speed The slave computers remain even during recording mainly idle Utilizing their computing power in a parallel computing environment the coincidence check process could improve the throughput for high end data analysis The development of additional helper applications is in progress at the present time For the operator in regards to Ease Of Use a World Wide Web interface will provide access to the spectrometer It may allow to perform all the necessary adjustments One can follow the course of the experiment remotely at any time as long as the spec trometers host computer is connected to the internet The rather complex modality of extracting PAC frequencies from the data will be possible through a graphically based web interface supported by a universal colle
71. g form by inserting Eq 16 maA ma mA ma mul rf e e mi n n m e 7a n n X mijn mn man m n je Per Fan 25 n n It is easy to see that Eq 25 contains a complex phase which incarnates a splitting of the energy levels The degeneracy of the intermediate sub levels is hence broken up The energy eigenvalues can be written as Em hwg 8m I I 1 26 For half integer J there are 2 1 2 I 1 for integer I doubly degenerate m levels 2 3 Absence of Perturbation The influence of the extranuclear perturbation is completely described by the per turbation factor The magnitude is contained in the Hamilton operator Hg for quadrupole interaction which determines A t from Eq 16 For a vanishing per turbation Hg 0 the time evolution operator reduces to the identity The matrix elements from Eq 23 vanish for all states a b and therefore is N N N With the orthogonality relation of the 3 j symbols 16 I I I I 1 mami mM TMa m Ma 2 the perturbation factor then becomes Ge t Ok ko ON N2 28 Thus the angular correlation function 22 reduces to kmax W ki k2 Ar Ar Prlcos0 29 k 0 keven 15 2 4 Polycrystalline Samples 2 PAC THEORY which is now independent of time 2 4 Polycrystalline Samples Angular correlation experiments are usually carried out on polycrystalline samples Polycrystalline samples or powder samples consists of a
72. ge level swing of The digitizer s external clock input requires at least an order of 1 V voltage level swing The external clock input is terminated internally by a 50 Q resistor The clock source must be able to drive the 50 Q impedance line to provide the required voltage swing Therefore the oscillator output is amplified by an integrated line driver circuit Its output signal is passed through a capacitor to remove any DC offset in the clock signal The capacitors value matches the cable impedance at a nominal frequency of 39 3 9 Synchronization and Channel Select fanout module 3 HARDWARE 10 MHz Figure A 3 4 in Appendix A 3 shows the schematic diagram of the common clock driver design A quartz oscillator from ECS 3951 series with a nominal frequency of 10 000 MHz serves as the clock source It is rated with a frequency stability of 50 ppm 24 On the first prototype board the output is split up into four paths which leads to the input of the high speed line driver integrated circuit The type of line driver used here is a DS26LS31 from National Semiconductors It is a common type line driver used in several high speed interface applications such as RS422 RS458 or Ethernet 801 3 25 3 9 Synchronization and Channel Select fanout module The Synchronization and Channel Select SCS fanout module is the bridge from the Host s parallel port to the Slave s analog Multiplexer module It conditions the hand shake and synchronization
73. grey The center of the prompt peaks represent the time delay between two channels within the electronic system The average width of the peaks are Atrwam 365 ps 5 4 PAC Sample Preparation PAC samples utilizing In Cd in chloride solution and Hf Ta were prepared for testing purposes The test samples were chosen because the compounds were studied with PAC spectroscopy before and results are listed in numerous publications 33 34 38 40 44 43 42 39 Thus their PAC frequencies are well known and can be used as a reference to the results obtained with this new spectrometer setup The following list of samples were deployed in the first test runs 68 5 EXPERIMENTS 5 4 PAC Sample Preparation a In Cd in Sn metal b In Cd in CdSiP c In Cd d Hf Ta in Hf metal in Fe foil and Sample a was provided by M O Zacate Department of Physics Washington State University A few drops of InClz solution were added to 3 g tin powder in a melting dish At preparation time the sample activity was calibrated to 160 wCi The sample was then molten in an oven for 2 h at a temperature well above the melting point of 231 C During this time most of the chloride evaporated out of the sample The sample was then quenched and shaped to a sphere in a polishing process The sample was received 2 days after preparation and the activity was measured to be around 73 pCi Sample b was obtained from Dr S Unte
74. he photomultipliers output signal are being processed digitally and the energy and the Time of arrival information are placed on a storage medium for later evaluation This fully digital recording system holds an average time resolution of about 400 pi coseconds thanks to the high speed acquisition boards that are being utilized in this design The new system provides advantages with respect to sample handling and spectrometer calibration with respect to conventional setups It competes also well in performance and many experimental difficulties related to electronics have been eliminated The spectrometers acquisition cycles are being controlled by software applications allowing a wide range of flexiblity in an easy reconfigurable system Acknowledgements I wish to take this opportunity to thank all the people who supported me in my work and foremost inspired me with their ideas and knowledge First of all I would like to thank Prof Dr John A Gardner for leading me onto the path to PAC spectroscopy in the field of solid state physics and nuclear science Also for sharing his thoughts with me and being a steady source for terrific ideas and not at last for reviewing my thesis towards the end Further I would like to thank Prof Dr Klaus D Becker for giving me the opportunity to use my work on PAC spectroscopy as a subject for my doctoral dissertation for the provision of lab space and equipment supporting my efforts at all times and for ena
75. hich controls the operational flow of all Slaves during a recording cycle When Pachost is invoked it connects itself with all Slaves over the local area network It evaluates its configuration and sets up the Slaves With the beginning of a new measurement Pachost will deliver the project settings to the Slaves such as the project name and date and time when the measurement was started etc The final step is to initialize the handshake bus before assuming operation The diagram in Figure 10 b depicts the application flow of the host process Pachost requires the following arguments at minimum pachost lt first sequence gt lt last sequence gt options 51 4 3 Host applications 4 SOFTWARE a b Y Parsing configuration Parsing configuration files Parsing command line arguments Parsing command line arguments Y AA Bu Exchange configuration Open TCP connection with Slavey 4 Create TCP socket over TCP IP layer Y y Initialize handshake bus signals Initialize handshake bus signals Lower BUSY E Last sequence No Y Je Lower CHANNEL_SELECT Handshake Next sequence for next sequence CHANNEL_SELECT low Y Exchange parameters Yes with Slaves h d Readout waveforms h d Slaves busy Handshake Lower BUSY BUSY high No Y Swit
76. ient for the trigger time analysis The configuration for each PAC channel is stored in the file pacslave conf The file is located in the project folders etc directory Pacslave can be given a set of command line options Any option specified on startup overrides its default value The set of command line options and the content of the configuration file can be found under Appendix A 4 1 The following is an example of how to start Pacslave with an input gain of 0 2 V full scale an offset of 0 09 V a negative trigger slope and a trigger level at 50 mV the command line would look like this pacslave a1 v0 2 00 09 ts1 t140 s400 S8000 The number of segments per recording cycle is set to 8000 Each segments holds 400 sample points 4 2 3 Pacslave file format This is a discription of the file format used by Pacslave to store the fitting param eters The file name is composed of the PAC channel number nen and a 6 digit sequence number nse plus the extension seg A 4 channel spectrometer would pro duce chronologically related files with names 1000100 seq 2000100 seq 3000100 seq and 4000100 seq where nseg 100 In general sequence files contain ASCII text Labels are introduced by sign at the beginning of every line A label represents the beginning of a series of arguments Arguments can be any kind of number types or string types The header of the 1000100 seq file looks like WAVEFORM O 100 7041 DATE 1136248918 Tue J
77. ight not have this limitation 34 3 HARDWARE 3 4 Digital Signal Processing Boards Voltage U mV 0 12 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Segment N Figure 5 This graphic shows a series of subsequent segments from the PC s memory Each segment here contains n 1000 sample points the PC s main memory and analysing the waveforms While the DSP boards are recording they are considered to be online The time they are online is defined as the time toniine that elapses until all 8000 segments are filled up A dead time toffline exists during the data transfer to the PC mainboard The time for one recording cycle is teycle tontine tof fline The detector activity can be expressed as the number of trigger events per online cycle Thus the detector activity is nn 82 tontine teycle bof fline where m is the number of segments The duty cycle the ratio between the online time tontine and the cycle time teycie of the DSP boards are dependend of the detector activity A The duty cycle can be expressed as follows om tonline tof fline 83 t mn cycle toffline where the ratio m torfiine can be expected to be a constant factor over the course of multiple recording cycles 35 3 4 Digital Signal Processing Boards 3 HARDWARE 3 4 1 Scope settings The input channel provides a fully programmable input amplifier The probe input settings are selectable from 50 mV to 5 V full scale At a full scale ra
78. ignals Photomultiplier pulse widths are typically in the order of several nanoseconds The acquisition of several thousands of PMT signals takes place entirely inside the digitizer memory buffer The attached computer system extracts timing and energy information and stores them The actual evaluation of Start and Stop events can take place at any later time as all the events are captured on a storage device Due to the tremendous amount of data flow it is basically just the storage space determining the length of the snapshot that can be taken of the decaying sample An overview of the new spectrometer setup is shown in Figure 3 Every PAC channel consists of an off the shelf personal computer sytem Slave Each one is equipped with a DSP board The DSP inputs are connected to a custom build Analog Multiplexer board MUX All internal clocks of the DSP units are synchro nized by a central 10 MHz reference clock network using the external clock input of the DSP board Note that this design is basically not restricted to a certain number of channels Due to its peer design the apparatus can theoretically be extended to an arbitrary number of channels without adding any electronic complexity to the system The Host computer system is the supervising instance in this setup Through an external handshaking network it monitors each Slave s state and directs their actions 30 3 HARDWARE 3 1 Overview of Setup and Operation Signal processing Time
79. ime and energy events are available at different times The energy determination is regarded as a slow process because the shape of the dynode pulse for Nal detectors is broader The energy information is more precise when integrated because a Nal scintillator produces more visible photons for a given gamma than BaF does In fast fast systems the time and energy information are practically available at the same time The fast component of a BaF detector s anode pulse enhances the detection of the signal s arrival time The much smaller slow component contributes less information when integrated This yields in comparison to Nal scintillators to an inferior energy resolution The energy information is available faster due to shorter integration times The energy discrimination is performed subsequent to the amplification by twin Single Channel Analyzer SCA The energy information is sorted into Start and Stop events according to the energy windows One SCA has its window set on the lower energy y and the other on the higher energy The size of the windows are usually sufficiently small so no overlap can occur The anode signal is usually artificially delayed by electronic circuitry until the energy information from the slow circuit is available With the arrival of a Start event a tim ing window is opened and associated with a Stop event The timing signal is typically shaped by a Constant Fraction Discriminator CFD which delivers a time of
80. information about general project settings such as project path and the Slave computer IP addresses Pachost opens up the network ports to the listening Pacslave processes and negotiates the preparation for the measurement with them When the acquisition cycle begins the Pachost process supervises the measurement and the Pacslave processes constantly log the energy spectra and time of arrival into the specified project path Both the energy spectra and timestamps are available during the measurement so that the Cocheck process can be started any time during the acquisition cycle The Cocheck process generates the energy spectra from the gathered information The operator can adjust the Start and Stop energy windows according to each channel s energy spectrum individually Cocheck will search for coinciding timestamps and will build the count rate spectra with the help of the Start and Stop window settings Note that no calibration of any electronic component is necessary 5 2 Data Processing During the performance of a PAC experiment the Pacslave processes collect y ray information during the decay of the isotope The time of arrival and energy of the y radiation is stored on the storage medium At this time neither energy nor time discrimination is performed on the raw data Basically a snapshot is taken of the 57 5 2 Data Processing 5 EXPERIMENTS Global setting unit Value Full scale mV 200 Offset mV 90 Input coupling Q
81. ion fh 456 Number of sequences 4 168000 Acquired Y rays 5 1 10 Acquired coincidences 183 7 10 Beginning trigger rate kHz 4 11 4 1 5 2 FWHM window 10 5 open window Beginning coincidence rate kCo s Table 6 Summarized results for the c In Cd in Fe sample experiment 6 2 PAC measurements The PAC samples introduced in Section 5 4 were measured in the new PAC setup This section will present the results obtained for samples a d The results will be compared with data denoted in the literature In order to verify the system functionality these measurements on well known systems were done for validation purposes All measurements were carried out at room temperature and no external magnetic or electrical fields were present Because of variations in sample activities and detector distance the duration of the experiments varied between 7 hours and 3 days The Cocheck process accumulated the time spectra into a set of 12 files denoted by dij dat These files contain 12 single time spectra D t plus the random background count Coincidences were collected in time slots with a resolution of 250 ps The coincidene spectra consists of 8000 of such slots Coincidences were collected for a wide span of positive times as well as negative times The prompt peak to is located in the center of the spectrum The large time span enables the exact determination of the average background count B The knowledge of the exact av
82. ion times tofCorrect1 0 tofCorrect2 0 tofCorrect3 0 tofCorrect4 0 Artifical jitter to smoothen timestamp 1 no jitter gt 1 randomized integer is added to timestamp in units of ps jitter 1 Solid angle fraction Based on detector distance from sample d 5 cm and scintillator crystal diameter of sc 1 5 solidAngleFraction 0 036 detectorDistance 0 05 crystalDiameter 0 0381 Save time spectra only within this boundaries 1 0 amp r 7999 saves all leftIndex 0 rightIndex 7999 104 REFERENCES REFERENCES References 1 J D Jackson Classical Electrodynamics New York London Sydney Toronto 1962 2 H Frauenfelder and R M Steffen Angular Distribution of Nuclear Radiations Al 10 11 12 13 14 15 16 pha Beta and Gamma Ray Spectroscopy Volume 2 ed K Siegbahn North Holland Amsterdam 1965 997 1198 G Schatz and A Weidinger Nukleare Festk rperphyisk Teubner Studienb cher Physik Stuttgart 1997 K S Krane Los Alamos Scientific Laboraroty Report LA 4677 1971 A J Ferguson Angular Correlation Methods in Gamma Ray Spectroscopy Amster dam 1965 H C Brinkmann Quantenmechanik der Multipolstrahlung Noordhoff International Pub 1965 R R Kinsey et al The NUDAT PCNUDAT Program for Nuclear Data 9th Inter national Symposium of Capture Gamma Ray Spectroscopy and Related Topics Bu dapest Hungary
83. is segment is disregarded The energy can be determined by integrating the signal over the entire length of the segment The tail of the pulse the slow component typically decays within several hundred nanoseconds Its integral contains most of the energy information 47 4 2 Slave applications 4 SOFTWARE 0 U t A R 50 min U t 100 0 5 10 15 20 25 30 35 40 45 50 t ns Delayed U UCet VOLL HN HH HH gt zero crossing ty x U UM 00000000000000000044 0 5 10 15 20 25 30 35 40 45 50 t ns Figure 8 Illustration of the software based CFD technique applied on a transient waveform The slope of the BaF detector s fast component is used for the corrections on the time of arrival The zero crossing point tefa is independent of the signal height 48 4 SOFTWARE 4 2 Slave applications Zero crossing t ns 0 100 200 300 400 500 600 E keV Figure 9 Results of the CFD technique applied on the three energy peaks of Hf Ta On average lower energies appear to be detected at later times than higher energies The data points represent the zero crossing times t ra by which the actual timestamps have to be corrected For practical reasons the segment size is set 400 samples thus giving the maximum number of recordable number of segments of 8000 Furthermore by using Eq 80 this yields to a maximum possible repetition rate of 833 kHz Test runs have shown that this theoretica
84. isotopes with different radii 11 It is according to Poisson s law aD Ve Vi 39 Ox 0 E0 Thus the source for the monopole field is the nuclear charge q f pyn r dr eZ with the charge distribution located about x 0 To simplify things the monopole term is separated from the quadrupole term in 38 by adding and subtracting 39 7 6 The partial derivatives have been replaced by a The separated quadrupole term is now trace free Under these considerations the Taylor series 38 can be rewritten into the form 2 3 3 n Be 1 1 r P r Qo rEo a 2 P 3 2 Pij lriry zou 40 Inserting 39 into the third term of 40 gives 3 a i 1 P T do FE ar Ado amp Burn 7765 41 aj The electrostatic energy of the nucleus can now be calculated by inserting 40 into 37 The Coulomb energy of the nucleus looks like Ea do putin v pwliyrEvar v 1 Ado on dr Eo mt P rirj r dr PE qo PEo Advark ts gt gt Dt 42 tj Here is ry the average radius of the nucleus By definition is p the classical dipole moment and Qu PiMan n 1 6 aV 43 e 19 2 5 Static Electric Quadrupole Interaction 2 PAC THEORY the classical quadrupole moment t With the selection of a suitable coordinate sys tem the integral of the non diagonal elements is equal to zero Q O jz and the quadrupole moment tensor is therefore
85. istance between the nucleus and the external 1The dimension of the electrodynamic quadrupole moment term is As m In many quantum mechanical publications the quadrupole moment Q is defined by QY e Q where the quadrupole moment is reduced by the elemtary charge The sum of the diagonal terms contain gt 3r r 0 20 2 PAC THEORY 2 5 Static Electric Quadrupole Interaction potential producing charges is large enough 46 can be rewritten as Ejer Hg DI Sl cos 6 47 ij l 0 cos 0 un is the angle spanned by r andr With the help of 6 and the addition illri theorem for spherical harmonic functions the above formula can be expressed through the spherical harmonic functions Do en 2 y 0i pi y 0507 48 tj Reordering the summation terms and using the complex conjugation of Y 1 Y we can rewrite 48 into the form La n 1 m E eiri Y 0 i Pi pe ej pry ra 0j pj 49 A more compact form of 49 can be achieved when the two summations over i and j are replaced by two spherical tensor operators Q and V of rank I 13 With the definition of the scalar product of two tensor operators l MR S DOR 50 m I and Ci yy Ove Deny Ti 8 Yi vn Rat 0j pj 51 id 49 becomes An Hyp x ee 52 1 0 The Hamiltonian is now expressed by the product of two factors a nuclear factor and an external field factor The summation is of order l In analogy
86. itted according to Eq 103 The orientation of the Sn metal grains is mostly randomly distributed within the quenched metal The core parameters Sgn however may deviate from the theoretical proportions due to non random texture effects This can occur when there are just a few large grains in the sample A correlation has been seen between low melting temperature and increased non random texture effects 41 Figure 19 shows a plot of the fit function prototype Eq 103 and the time dependent anisotropy as measured for sample a with a timing resolution of 250 ps The corresponding Fourier spectrum on the right renders the intensity relation of the coefficients S n On average the maximum acquired coincidence count for all detector pairings was 9000 around to The statistical coincidence information per time slot decreases with higher timing resolution for experiments with equal duration The acquisition hardware in this work is 2 4 times superior in comparison to conventional analog PAC setups which typically yield between 0 5 and 1 ns per slot 17 53 Lower statistical information per time slot results typically in larger errors bars for each data point Sample a showed a quadrupole interaction with an asymmetry factor 7 0 The occupation of Cd in one site was observed as is fi 1 The three PAC frequencies TT 6 2 PAC measurements 6 RESULTS wn were found to be in the ratio 1 2 3 with the fundamental frequency w 34 76 Mrad
87. kappeln Fachgymnasium Technik Osnabriick Allgemeine Hochschulreife Deutsche Bundespost Telekom Kommunikationselektroniker Studium der Physik an der Universitat Osnabr ck Diplomvorpr fung in Physik Diplomarbeit in der Arbeitsgruppe Kr tzig Angewandte Optik Diplomhauptpr fung in Physik Wissenschaftlicher Mitarbeiter am Institut f r Physikalische und Theoretische Chemie der TU Braunschweig Computer Hardware Ingenieur bei ViewPlus Technologies Corvallis Oregon USA 109 Ver ffentlichungen D Kip M Wesner C Herden and V Shandarov Interaction of spatial photore fractive solitons in a planar waveguide Appl Phys B 68 1999 971 974 M Wesner C Herden D Kip and P Moretti Photorefractive steady state solitons up to telecommunication wavelengths in planar SBN waveguides Opt Commun 188 2001 69 76 M Wesner C Herden and D Kip Electrical fixing of waveguide channels in strontium barium niobate crystals Appl Phys B 72 2001 733 736 M Wesner C Herden R Pankrath D Kip and P Moretti Temporal development of photorefractive solitons up to telecommunication wavelengths in SBN Phys Rev E 64 2001 36613 36617 D Kip C Herden and M Wesner All Optical Signal Routing Using Interaction of Mutual Incoherent Spatial Solitons Ferroelectrics 274 2002 135 142 C Herden M Alves K Becker and J Gardner A new generation TDPAC spec trometer
88. ked to each channel and the main controlling unit As a very cost effective solution the relatively few handshake logic lines were established through the in and output pins of the Standard Parallel Port that every PC mainboard is equipped with While the hardware complexity and costs were kept to a minimum the system on the other hand required the development of a vast amount of software applications During the course of the development a framework was created that combines all vital applications and drivers Among them are the low level device driver pacmod o and the manufacturer supplied driver acqrs1 o sup porting the handshake bus and the digitizer hardware The high level applications such as Pacslave Pachost were instantiated as part of the real time inter process ac quisition framework For a N detector instrument there is one instance of Pacslave processes running on each of the N slave computers behind each detector The main controlling process Pachost is running on the main host computer and is monitoring each one of the slave computers All processes are connected through the fast hand shake bus and the Ethernet network switch The helper application Cocheck was developed for coincidence filtering and first stage data reduction purposes Filters for energy selection are applied also in this instance These filters are customizable in configuration files at any time All applications were written in plain C C under the objective of port
89. kets where the time evolution operator A t appears A more complicated and lenghty calculation brings Eq 21 into a more accessible form kmaz W ki kat gt 3 An 71 Ake 12 Ci t k k2 N1 N2 1 i Ta aT 90 YA 62 2 22 1 2 with the separated perturbation factor Beh XO amt 2k 1 2k2 1 Ma Mmh Mp Mi er a Seinen 23 mi Ma N m m N k and k are the wave vectors of the two successive y rays Ap y1 and A 72 are anisotropy coefficients The parameter Ag 71 Ar l l Zi I is a number which depends only on the multipolarity of the emitted radiation l and and the spins of the nuclear states involved in the transition The parameter Ax y2 Ars la 15 Lp T is similar to the one before but involving the multipolarities and spins of the second transition Tabulated values can be found in 4 5 The index k is limited by the relation O lt k lt kmar min 27 1 l l l 24 where kmar is the highest index representing the highest involved spin state The summation contains only even values of k because it is restricted to directional y y angular correlation not measuring the circular polarization of the emitted y rays One can find a set of eigenvectors n for that a diagonalized form of the perturbing 14 2 PAC THEORY 2 3 Absence of Perturbation Hamiltonian Hg exists The scalar products in the perturbation factor 23 can be rewritten into the followin
90. l reception rates can be achieved by the electronic system In practice the limiting factor is by far the PMT When the incident photon flux is too high the anode saturates and the trigger slope cannot be identified sufficiently After the fitting process is completed for all segments the results the energy inte gral number the timestamp and the CFD factors are stored in sequence files A discription of the sequence files is given in Section 4 2 3 With every new sequence the sequence counter is increased The file name reflects the sequence counter value A set of files with the same sequence number possesses a chronological relationship The synchronization pulse is typically the first segment recorded by the DSP board Because of its unique shape it can be easily identified as a valid reference point Therefore the first entry in a sequence file represents this reference point Within a set of sequences the Cocheck process will later look for coinciding events See Section 49 4 2 Slave applications 4 SOFTWARE 4 3 2 for more details 4 2 2 Pacslave options In order to work with the DSP cards correctly Pacslave must be provided with the right set of options Each PAC channel has its individual set of parameters The full scale range Vrgr default value is 200 mV with a vertical offset of Voss 90 mV The trigger threshold must be adjusted individually according to the PMT gain A trigger delay of tg 20 10 s proofed to be suffic
91. large number of randomly oriented microcrystals It is possible to express the perturbation factor by averaging over all possible orientations of microcrystals This has been done by 19 The matrix elements of Eq 25 need to be integrated over the solid angles that account for all possible orientations This leads to the mean perturbation factor ice t ka k2 N N gt a ER er 30 m m which is diagonal in N and k The s coefficients can now be separated into a so called hard core term which is time independent and time dependent terms that represent the different split sub levels of the perturbation The hard core term originates from all states n n The time dependent terms are generated from all states n n The perturbation factor for a powder source takes on the form Gilt X skh S Shim cos Em En t h 31 m mzm Using Eq 26 and introducing a new index n m m 2 for half integer I n m m for integer I the core parameters can be written as sk stk OnJm2 m 2 2 half integer I 32 sh gt 3 ae integer T 33 Thus Eq 31 can be simplified as Gurt sk gt s cos wnt 34 n gt 0 The angular frequency w is equivalent to smallest non vanishing quadrupole fre quency 6wg for half integer J 3wg for integer J There is no perturbation for t 0 16 2 PAC THEORY 2 4 Polycrystalline Samples thus G t 0 1 This conc
92. le folders are shared among the Host and the Slaves over the Network File System The Slaves export their data directories to the NFS The Host gets access to them by mounting the exported data directory into its own file system On the Host side each mount point is constructed as mnt pac n data projectname where n is the index of the PAC channel and projectname the path to the measurements data directory respectively 56 5 EXPERIMENTS 5 Experiments 5 1 Host and Slave Preparation Before an acquisition cycle can start the spectrometer elements have to be config ured This involves mainly the preparation of the configuration files for the Pacslave Pachost and Cocheck processes Every measurement requires its specific set of config uration files The Pacslave process uses the provided information in file pacslave conf to prepare the DSP cards accordingly It receives information about electrical prop erties of the PMTs that are specific for this kind of experiment Additional project information such as project path and network port is also provided Table 4 contains asummary of the default options for the Pacslave process Once the Pacslave process is started on each of the Slave computers it listens on a network port for the Pachost process to connect When the Pachost process is executed the project configuration file pachost conf is loaded The configuration file was setup specifically for the measurement and con tains
93. lely by the strength of the third component Vz of the electric field gradient tensor and the asymmetry parameter 7 The lattice environment in both metal alloys produces an electric field gradient component V that is non zero The asymmetry parameters can vary between 0 lt 7 lt 1 For the above cases the magnetic domains and the domains with quadrupole interaction are assumed to be oriented randomly within the sample 1 INTRODUCTION and therefore the perturbation function W 0 t can be treated with the approximation for polycrystalline samples Due to possible non random texture effects however hardcore parameters may not reflect their theoretical predicted values Additional measurements were performed on In Cd in CdSiP2 a semiconductor with chalcopyrite structure The environment of the In Cd on substitutional Cd sites produces a symmetric electric field gradient V 4 0 7 0 at room tempera ture Unlike the above mentioned metal alloys this sample in the state of tiny grains represents a perfect polycrystalline sample The hardcore parameters are expected to reflect the values predicted by the theory 2 PAC THEORY 2 Introduction to Perturbed Angular Correlation Spectroscopy 2 1 Theory of Angular Correlation Radiations from an atomic or nuclear system are not random in their directions The method used in PAC spectroscopy works because a pattern in the decay of nuclear states over metastable to ground states
94. life Ty 18 1 us 10 8 ns oo Quadrupole moment Q 2 35 6 b none Magnetic moment u 3 29 3 un 2 3705 7 un Gamma energy 133 02 1 keV 482 18 9 keV Multipolarity E 2 M 1 E 2 Table 2 Characteristics of isotope Hf Ta Values taken from 7 8 10 2 PAC THEORY 2 1 Theory of Angular Correlation Generally the nucleus is not known to be in a definite quantum state One can therefore assume that the initial magnetic substates are equally populated With the detection of y in direction k a selection is made out of a pool of eigenstates of the quantum mechanical system The nucleus is now in an intermediate state with some relative in general unequal population of 2 1 states Upon decaying to the ground state the detection of the second 9 ray in direction ko is weighted with a certain probability k and ky are wave vectors in carthesian coordinates as opposed to 0 and in polar coordinates Calculations can be done to determine the probablity of the observation of y and Ya with respect to their polar angles The angular correlation function W ka ko predicts the probability of finding a Ya in direction ko subsequent to the detection of a 7 in direction k In a quantum mechanical formalism the initial intermediate and final states can be described in the following ket notation L Mi Z M and Ir My 8 For the two transitions the matrix elements of the transition amplitudes can be ex pressed
95. line PAC experiments involving probes with an J 5 2 intermediate state the angular correlation function 35 can be written in the more simplified form W 6 t 1 Ay Goo t Po cos 8 93 For In Cd and 8 Hf Ta is A44 lt Ago and hence A44 can be neglected for practical reasons In order to extract the time dependent anisotropy A22G22 t from the count rate spectra the following formulae ray 2 N 180 t Ba Be 7 R t AwG a ON 94 71221122 1 4R t 61 5 2 Data Processing 5 EXPERIMENTS can be used with the four N 180 t and eight N 90 t spectra The time dependent anisotropy Ag2Go t is here equated with the count ratio function R t In polycrys talline samples with random nuclei orientation exists no preference in orientation with respect to the detector arrangement The four 180 count rates and the eight 90 count rates can be grouped together to gain a single expression as follows Na Ne NaN paN a 180 8 95 TN N 180 t Kok kakz 1 N 90 t Nos Ni2No1 Na3N30N 21 N10N32 ya VEOF IFES Net W 90 t 96 TN Koki K2k3 This artifice makes use of all available data at once increasing the statistical informa tion in the combined count rates N 0 t All the information about the interaction between the nucleus and the extranuclear fields are only contained in Gaa t In practice the count ratio is fitted to an expression which represents the perturbation function
96. ling constants were obtained from a least square error fit on the data according to Eq 103 with a single fraction f 1 The theoretical predicted values of the core parameters s for a polycrystalline sample with an axially symmetric EFG were held fixed during the process The theoretical core values for I 5 2 are 2 Sx 1 5 er S2 2 7 S 1 7 The least square error fit complies very well with the obtained data The effective anisotropy was identified as A22 eff 0 109 for a detector distance of d 5 cm The 79 6 2 PAC measurements 6 RESULTS fundamental PAC frequency resulted in w 107 2 Mrad s giving rise to a quadrupole coupling constant of wo 17 87 Mrad s for half integer spin and an EFG component of V 6 11 107 V m n 0 The quadrupole frequencies and components of the EFG could be reproduced with an accuracy of about 4 and comply very well with the quadrupole coupling constant vg 110 1 9 MHz wo 17 5 4 Mrad s found in 33 34 6 2 3 Sample c In Cd in Iron foil Electric quadrupole interaction is usually abondoned in materials with body centered bcc cubic structure such as a iron metal 52 Ion implantation induced defects in the host lattice can lead to break up of the symmetry and leaves the probe atom in an environment which causes the EFG to be non zero Defects are usually healed when the irradiated sample is annealed the original symmetry is restored and the quadrupole in
97. ludes that the sum of all core parameters is gt sk 1 One feature of the hard core term is that the angular correlation is not completely zero for a powder sample even if the time dependent perturbation vanishes The simplified angular correlation function can hence be rewritten as W 0 t 5 Arr GNN t P cos 0 35 k 0 even This expression is only dependent of the angle 0 Z k k and the time t The angular correlation function consists of the time dependent anisotropy AkkGkk t and the angular term P cos 0 For the two most commonly used PAC probes In Cd and 8 Hf Ta the nuclear spin is J 5 2 which yields to kmar 4 For these elements the anisotropy coefficients Apo A22 and A44 are involved For the zero order term the Legendre polynomial Po cos and the Aoo are equal to 1 Since Ay lt Ag for both In Cd and Hf Ta probes they are usually neglected in the analysis of the correlation function for practical reasons 17 2 5 Static Electric Quadrupole Interaction 2 PAC THEORY 2 5 Static Electric Quadrupole Interaction The electric quadrupole interaction is the interaction of the nuclear quadrupole mo ment with an electric field gradient EFG at the nucleus The EFG is characteristic of the charge distribution around the nucleus It is usually described by a tensor Vaa containing the symmetry of the charge distribution The presense of an EFG causes a splitting of the degenerate eigenstates of
98. n increase when digitizers with sampling rates of 2 GHz or more are being employed However at this time the timing resolution limiting factor can clearly be attributed to the present y ray detector technology 85 7 CONCLUSION The new PAC spectrometer has advantages in many aspects The design allows the expansion to basically any number of detector channels The complexity of the com ponent arrangement does not pose any limit on channel numbers For convenient mathematical reasons conventional PAC detector arrangements of 4 or 6 detectors are typically chosen with angles of 90 and 180 on all three spatial axes With gaining spectra under any angles between detectors within one experiment conven tional analysis and data reduction methods are not necessarily applicable anymore Therefore new data reduction methods need to be developed for arbitrary detector setups The entire system is controlled by software which can in most cases be easily ad justed to certain needs The calibration of the spectrometer is comparatively simple A manual timing calibration of the system is basically not applicable due to the precise clock generators on the digital recorderboards Because these sample clocks are programmable the spectrometer timing resolution can be adjusted individually by software The y ray spectra are recorded continuously Start and Stop windows can be set according to the spectra that are being stored while a measurement is runnin
99. nal Immediately after all Slaves have reported being armed it releases the sync pulse and switches the analog MUX boards to the PMT input channel From now on the host monitors the status of the Slaves until all of them are finished recording This cycle repeats continuously until the measurement is finished 37 3 6 Slave computers 3 HARDWARE 3 6 Slave computers At this point of development of the PAC spectrometer there are 4 slave computers involved in the aquisition of event data For symmetrical reasons 4 identical Personal Computer configurations were chosen Each system features a state of the art moth erboard with a 2 8 GHz AMD Athlon microprocessor 512 MB of RAM a 100 Mbit network adapter and 80 GB harddrive Note that there are practically no limita tions on the harddisk space Each slave is equipped with an Acqiris DP110 transient recorder board in one of its 32 bit PCI slots A PAC slave computer is responsible for configuring the recorderboards such as setting up the input gain trigger threshold and trigger slope memory usage sample intervals etc The Slaves download the data from the recorderboards once a recording cycle is finished Their main purpose is to analyse the recorded waveform data primarily extracting information such as photon energy and time of arrival Instead of storing the entire waveform only a few param eters are permanently stored on the harddrive For practical reasons the waveforms are generally discar
100. nal clock frequencies shall thus be assumed at v 1 80 pst The synchro nization intervals are given by the external 10 MHz clock When a coinciding event is recorded between two channels a snapshot of the two timestamp counters is taken A slight divergence of the clock frequencies v and phases y during the 100 ns period must be regarded therefore the internal clock frequencies v can in general be slighty out of tune The superposition of two frequencies v and vs results in a fast and slow beat oscillation V fast v 12 2 gt Yslow v V2 2 99 Are the two frequencies close enough together the slow oscillation almost vanishes and Vast can be regarded as Vfast Vi V2 When the phenomenon of interference between two coherent sources are described mathematically by the superposition the orem the result is usually modulated by a sinusiodial term 32 The experimental evidence suggested that the data points were fitted according to the empirical expres sion A ft er 1 cos Orv fase t 10 y t 100 Several measurements between all detector combinations were performed with results that yield to an average time resolution of Atpwram V2ln2 w 365 82 ps The minimum profile width ever achieved was Atpw m 236 8 ps From the fits the average clock granularity could be determined as Vyas 1 76 ps7 The center of the prompt peak tg is shifted to negative times at t 1698 ps This gives rise
101. nd definition PACMOD_TRIGGER Create sync pulse on m PACMOD_SYNC Switch MUX to sync channel BT Deasserts DATAO PACMOD_PMT Switch MUX to anode channel a ee OS PACMOD_DATA_SET Set bit in parallel ports ESS pata e PACMOD_DATA_RESET Reset bit in parallel ports Table 3 List of command declarations which are currently implemented in the pacmod kernel module source of Pacslave is to manage the DSP hardware It prepares the DSP acquisition cycles as commanded by the Host The chart in Figure 10 a shows the program flow of Pacslave The DSP cards are supported by programming libraries which have been provided by the manufacturer The libraries give access to the DSP card settings and are utilized by Pacslave Pacslave analyzes the formerly acquired data parallel to the digitizer s recording cycle so that no overhead time is demanded for the analysis task It accumulates a time and energy stamp from each segment and stores this information on the harddisk The timestamp information is more or less already provided by the DSP boards trigger mechanism in form of a 64 bit integer variable in units of 10 s However the precise time of flight ti is often depending of the signal height of the incoming pulse A technique called constant fraction discrimination delivers the trigger time independent of the signal height In conventional PAC setups this technique is applied to the PMT signal by external electronic hardware In this sp
102. ne immediately This action connects the PMT with the digitizer input The Slaves will now report being BUSY until the onboard memory of the DSP has been filled up with transient pulses After all Slaves have transfered the waveform data from the DSP to the PC memory they will deassert their BUSY lines The Host will then reinitiate the next recording cycle 3 3 Standard parallel port The signals of the Standard Parallel Port SPP are divided into 3 groups Data 8 outputs Status 5 input and Control 4 outputs Control and Status lines were originally designed for interface control and handshaking purposes Eight data bits provide data from the computer to a device Each control and data bit can be individually set or reset Data Control and Status bits are accessible through an I O port The most common base address is 0x378 In some systems this address can also be either 0x278 or Ox3bc Usually the SPP base address can be assigned manually in the BIOS In SPP mode the data port is write only Writing a byte to the base 32 3 HARDWARE 3 4 Digital Signal Processing Boards _Hostsignalname Dir SPP pin SPP pin Slave signal name SYNC_PULSE OUT STROBE i i CHANNEL_SELECT OUT DATAO ro JACKN IN CHANNEL SELECT ot DATAI OUT 2 5 Figure 4 Handshake bus system timing diagram for a Slave Host combination The Slave signalizes to the Host that it has finished with this recording cycle at
103. ne up in the ratio of 1 2 3 25 2 6 Static Magnetic Dipole Interaction 2 PAC THEORY 2 6 Static Magnetic Dipole Interaction A charged rotating body has a magnetic moment ii when it exhibits an angular moment J The ratio between these two quantities is defined as p ri 73 where y is the gyromagnetic ratio Classically the gyromagnetic ratio y 54 is defined as a charge and a mass distributed uniformly Protons neutrons and many nuclei however carry nuclear spin that gives rise to a gyromagnetic ratio as above The ratio is conventionally written in terms of proton mass and elementary charge as Sh Sg 74 y Cao 74 The factor g is the Lande g factor which is a dimensionless quantity that characterizes the magnetic moment and the gyromagnetic ratio of the particle and uy is the nuclear magneton If the magnetic nuclear dipole moment ji is exposed to a magnetic flux density B then the interaction energy is ASAT eB 75 Classically the magnetic energy depends on the angle between ji and B In the quantum mechanical world exists a direction quantization that forces only certain adjustments for u If the B field is parallel to the z axis the magnitude of the magnetic energy of the angular momentum state I M is B I M L M M 7B AM 76 As can be seen in Eq 76 the magnetic energy is linear in M Therefore the energy levels Eu M Eu M M M 7B h Nhwz 77 are equispaced where wy is
104. ng 38 3 HARDWARE 3 8 Reference clock distributor Figure 6 Functional block diagram of the 2 channel analog multiplexer circuit AD8180 from Analog Devices which is used on the MUX module the synchronization state the host holds the SELECT input signal low Then the host initiates the TRIGGER pulse and immediately switches the MUX to the second input IN2 where from now on PMT pulses are getting to the SIGOUT output This procedure ensures that the digitizers record their first pulse from which is known it has been generated from a single source Since the pulse arrived at all digitizer units at a common time the first timestamp can be regarded as a reference to following timestamps In addition to switching the multiplexer channels the SELECT signal is also con nected to the Slave s parallel port ACKN input X1 Pin 10 Upon the transition from high to low the slave is also notified through the parallel ports interrupt that the host has put the spectrometer into the synchronization state This interrupt event is received by the Linux kernel driver A more detailed description can be found under section 4 3 8 Reference clock distributor A common clock source is required to operate the external clock input of the four digitizers In order to stay synchronized throughout the measurement the digitizers require a fixed frequency of 10 MHz The output signal of a high precision quartz oscillator can be used for that purpose The volta
105. nge Vrsr 200 mV the recorded value of 128 would correspond to a voltage of 100 mV 0 meaning 0 V and 127 a voltage of nearly 100 mV The quantization increment would result in 784 N Some care should be taken when selecting a voltage range to allow the signal to be recorded at a maximum scale A variable offset voltage Vo is programmable in the range of 2 V The programmable offset can be regarded as a negative offset to the full scale range Vrsr ranges from VesrR VrsrR 84 Vo lt Vrsr lt Vo Signals outside of Vrsr will be clipped and should be regarded as erroneous Some considerations have to be taken into account when choosing the optimal scope settings The digitizer analog input must match the impedance of a RG58 cable which connects to the MUX s output These cables are typically rated with an impedance of Zo 50 Q In general fast transient signals could cause signal reflections which propagate back into the cable when cable impedances do not match up Those re flections can cause multiple triggers which could lead to wrong timing information To avoid possible signal reflections the digitizer input must be operated with its 50 Q termination turned on As a result signal levels tend to be fairly low Thus a reason able sensitive input gain must be selected Signal levels from the PMT are typically of a few mV Hence an appropriate full scale input is programmed at Vrsr 0 2 V The vertical offset is setu
106. of a detector pair observing the Start and Stop Y s of an intermediate nuclear state For negative times the random background count is evenly distributed At to the coincidence count rate rises instantly For positive times the coincidence peak decays according to an exponential function with the half life of the intermediate state For larger positive times only the random background remains ps and 1 ns Theoretically there is also no limit in size of the observable time window In this work the spectrometer was set up to search for coincidences within an 8 us window Figure 12 shows a typical time spectrum The prompt peak for each detector pair i j is typically located about the center of the spectrum to Using the wide data range available left and right of o j it is possible to determine the average background B very accurately In addition the high time interval resolution supports the determination of the exact location of to The knowledge of both the horizontal location of the time spectra and the average background is important in the process of obtaining the perturbation function Given a sample with activity N every detector i registers a sample activity propor tional to N Q N with N e T No and 7 is the decay time of the nucleus describes the single detector response efficiency and Q is the fraction of the solid angle Not every y ray that enters a detector represents necessarily a Start event Also
107. ory of Angular Correlation 2 PAC THEORY m 1 1 rae cos 0 3 1 5 sin 0 cos 0 ql 3 cos 6 4 cos 0 Figure 1 Illustrations of the normalized radiation characteristics F m for pure dipole and quadrupole radiation and their mathematical representations in the table below Illustrations of the normalized angular distribution functions for pure dipole and quadrupole radiation and their corresponding mathematical expressions are given in Figure 1 2 PAC THEORY 2 1 Theory of Angular Correlation Parent isotope In Cd T1 2 8047 d lm leg 171 keV M 1 E 2 s 85ns 245 keV E 2 m or 2 8047 d EC Ur State Spink T Energy Half life Ty Quadrupole moment Q Magnetic moment u Gamma energy Multipolarity Initial 120 ps i 171 28 3 keV M 1 E 2 E 416 70 5 keV 245 40 2 keV 85 0 7 ns Ay 0 18 A 44 0 002 A 4 0 204 Ayp 0 001 Final 1 2 0 0 keV oO none 0 5948861 8 un Table 1 Characteristics of the isotope In Cd Values taken from 7 8 1 Theory of Angular Correlation 2 PAC THEORY 2 Parent isotope Hf Ta T 42 39 d 18lpr 42 39 d 1 2 p 1817 132 3 2 136keV 133keV A 0 288 5 t 108ns A 4 0 076 A54 0 062 346 keV A y 0 318 482 keV g t 40 ps 136 keV q T State Initial Intermediate Final SpinParity 3 3 T Energy 615 17 1 keV 482 18 9 keV 0 0 keV Half
108. ost is the controlling instance of the setup The Host demands control over the measurement and carries out all supervising tasks Slave applications are distributed processes which are performing helper tasks on the Slave computers Both Host and Slave applications are using the computer hardware resources An application running on a Linux system can gain access to the computer s hardware through a device driver The device driver intervenes with the computer s hardware resources A device driver in the world of Linux is called a kernel module 4 1 Parallel port device driver The kernel module pacmod o claims access over the parallel port hardware It handles two character devices thus giving access for user level applications to use the parallel port On the system it provides the two device names dev pac0 and dev pacl Read and write operations can be performed to either one of these device names Reading from the device retrieves a status word about the logic state of the parallel port input pins Writing to the device sets the parallel port output pins The kernel module pacmod o is installed on all Host and Slave computers 4 1 1 Installing the device driver Two nodes must be created before it is safe to install the module Under Linux a device node is identified by a unique major resource identifier A minor id helps to distinguish between a multiple number of device names such as dev pac0 and dev pacl More information about ke
109. ow is marked through the shaded green area the Stop window is marked through the shaded red area The 511 keV peak of the Sodium sample is marked by both the Start and Stop window This peak appears through a e e annihilation process in Sodium where two y rays are emitted in opposite directions at the same time The detection by two opposite PMT tubes can be used for calibration purposes of the PAC setup 4 3 3 Cocheck file format The time spectra are saved in files dij dat i and j represent permutations of the PAC channel numbers The file format contains a header followed by a two column data field The first colum is the running index of the slot The second one holds the number of coincidence counts The columns are separated by the tab character The header is designed like this DIJ lt i gt lt j gt lt number of data lines gt lt next sequence gt lt last sequence gt lt max count gt lt time scale gt lt diff unix time gt MAX lt max count gt 59 4 4 Network File System 4 SOFTWARE CENTER lt slot of tzero gt TIMESCALE lt time scale gt ps 0 567 1 553 2 532 3 The DIJ label is a designator for file type i and j are zero based indices of the PAC channel combination The third argument gives the total number of slots The MAX CENTER and TIMESCALE labels hold the maximum number of coincidence counts the center slot index and the time base in units of 1071 s 4 4 Network File System Fi
110. p according to the polarity of the PMT output The anode delivers an output signal proportional to the PMT power supply polarity Typically a negative polarized high voltage supply is used to operate the PMT Therefore one can expect to read a negative transient signal at the digitizer s input In order to gain an optimal voltage range for the transient waveform the vertical offset has to be set as close as possible to the positive sampling window An offset Vo 0 09 mV works sufficiently for a Vrsr 0 2 V giving an extra head room for determining signal levels that are out of range The trigger settings determine when the DSP device will start acquiring data The trigger source is the input channel When the input signal crosses a programmable threshold the waveform is recorded into the DSP memory The trigger threshold 36 3 HARDWARE 3 5 Host computer Vinres can be set to anywhere within Vrsr The threshold also determines the trigger sensitivity Small signals and noise can be filtered out With a negative trigger delay time tg the acquisiton of the waveform before the actual trigger occurrence is enabled The DSP hardware provides a timestamp for every trigger event The timestamp is derived from the DSP s internal sampling clock In addition it provides a horizontal positioning parameter as a correction factor for the timestamp Trigger events usually occur asynchronously with the digitizer clock Therefore the time between the trigg
111. r Numerical Data Rolle Beteiligt 1 April 2006 31 Marz 2007 111
112. r in precision and accuracy 22 Acqiris digitizers are available in two industry standards The DP series plugs into any vacant PCI slot on a standard PC motherboard The DC series are CompactPCI compliant and require an appropriate CompactPCI compliant crate They are supported with a vast amount of program ming libraries and interfaces All their functionality and scope settings can be easily programmed through different ways such as LabView or C libraries The PAC spectrometer in this work features four DP110 model digitizers The DP110 33 3 4 Digital Signal Processing Boards 3 HARDWARE series digitizers are a basic edition with one channel probe input and one external clock input The digitizers contain an Analog to Digital conversion system that is capable of recording in real time with speeds ranging from 100 samples up to 10 samples per second thus giving a timing resolution of at least 1 ns The analog bandwidth at full sampling rate is limited to 250 MHz meaning that fast signal components will be suppressed by an amount of gt 3 dB The ADC resolution of the full scale input is 8 bit ranging from values of 128 to 127 Data from the ADC system is stored in the on board acquisition memory The acquisition memory has a capacity of 128 kBytes by default An optional memory expansion to a total of 8 MBytes was installed to gain more digitizer performance The external clock input is used to synchronize all four digitizers to a common
113. re The sample was provided by Dr Ken S Krane at the Oregon State Radiation Center Oregon State University More detailed information 69 5 4 PAC Sample Preparation 5 EXPERIMENTS about the irradiation process and initial sample activity was not available to the author The sample was a recycled piece of irradiated hafnium metal wire of unknown age Due to hafniums fairly long decay time of nearly 1 5 months the sample showed sufficient activity to perform a PAC experiment on 70 6 RESULTS 6 Results 6 1 Performance The performance of the recording apparatus will be discussed in this section Pri mary interest is how the new apparatus performs in comparison to conventional PAC spectrometers In this analysis the data recorded for sample c was used to determine the perfor mance of the apparatus The sample was measured over a time span of 456 hours During this time the spectrometer recorded 5 1 10 segments in 4 168000 sequence files It takes the Cocheck program about 1215 minutes to process all the segments The quotient between process time and real time is 4 4 Thus the Cocheck process is on average 22 5 faster than the recording Every sequence file contains a set of 8000 segments that represent a transient signal recorded by the DSP boards The raw data usually contains all trigger events without preselection through energy win dows Thus all events from Compton scattered photons and thermal electrons are part
114. rnel mode device drivers can be found in 30 31 The following command creates two character devices mknod dev pacO c 254 0 mknod dev paci c 254 1 After these nodes have been created the module can be installed While installing the module two additional kernel parameters can be provided the base address and the interrupt number The most typical values for these values are 0x378 and 7 but in 43 4 1 Parallel port device driver 4 SOFTWARE general they depend on the BIOS settings The following command loads the module into the kernel insmod pacmod o 1ptIOBase 0x378 lptIrq 7 Whether the module has been loaded properly can be checked with lsmod The module can be removed again by the command rmmod pacmod 4 1 2 Using the device driver from the console For testing purposes the device driver can be communicated with by sending com mands over the console using the echo command A set of command strings are described in the Appendix A 4 3 The following two command lines manually switch between the MUX channels echo cp gt dev pacO echo cs gt dev pacO 4 1 3 Using the device driver in a C C application A C C application can use in and output functions from the standard I O library The following example in C shows how a read and a write operation is executed with the kernel driver include lt stdio h gt int main int argc char argv FILE f fopen dev pacO r if f
115. rricker Institut f r Angewandte Physik TU Bergakademie Freiberg as a few splinters from a single grown crystal CdSiP2 is a chalco pyrite structure compound of type AUB YCY The splinters were grinded down manually in a mortar dish into a powder The powder was then given into a quartz glass vial The amount of InClz solution was measured to not exceed the activity of 27 wCi The solution was then given onto a quartz glass plate and then dried in an oven for a short time to vaporize the chloride content The dried In Cd plate was added to the powder in the quartz vial Before sealing the vial a spatula tip of red phosphorus was added This retains the partial pressure of phosphorus inside the vial during the annealing process and prohibits the sample to be reduced from phosphorus The vial was sealed under a vacuum and then heated for 2 3 h at a temperature of 900 C Under this temperature the In Cd will substitute with the Cd sites while the overall structure of the chalcopyrite stays intact A piece of iron foil of the size of a centimeter in square and a thickness of 5 um was use as sample c The radioactive isotope In Cd was implanted at IONAS II Physikalisches Institut Universitat Gottingen with an energy of 400 keV The terminal activity did not exceed 27 uCi due to legal restrictions and transportation purposes The probe element Hf Ta in sample d was obtained by neutron irradiation of a piece of Hafnium metal wi
116. s A APPENDIX A 4 5 Cocheck Usage cocheck lt firstsequence gt lt lastsequence gt options lt firstsequence gt Sequence number to start coincidence check with lt lastsequence gt Sequence number to end coincidence check with Options lt n gt lt lowerstart gt lt upperstart gt lt lowerstop gt lt upperstop gt Specifies energy windows for channel n Default is n 0 8191 0 8191 Open Window Mode C lt configFile gt Path to config file c lt n gt Channels participating in coincidence check Default is c1234 all 4 channels Example c24 only channel 2 and 4 d lt n gt 1 dump found coincidences into CIJ DAT files This option sorts coincidences in Open Window Mode and discards all events that are dispensible Default is dO f lt n gt 1 Fast search algorithm Advanced analysis 0 Slow search algorithm Considered obsolete Default is fl j lt n gt Force jitter in units of ps to be added to timestamp Default is j80 l lt n gt 1 Start coincidence search on top of DIJ DAT content 0 Start coincidence search with empty bin arrays This options loads DIJ DAT content into bin arrays upon start up if selected Default is l1 o lt l gt lt u gt Specific lower lt l gt and upper lt u gt energy window limits for all channels at once 102 A APPENDIX A 4 Configuration files and Command line options p lt prjName gt Specify project name path to perform coincidence search on
117. s are shown for activities between 0 and 0 315 MBq The dashed line is a fit through the data points and represent the development of the duty cycle for higher activities ergy window width of 2 FWHM Table 6 summarizes all the obtained results for this experiment The coincidence rate varies with the width of the energy windows This is under standable because the number of random coincidences increases with the size of the energy windows In open window mode all the possible coinicidences are collected re gardless of the y ray s energy The beginning coincidence rate in open window mode was 10 5 kCo s The experiment showed that the new PAC apparatus did perform well with respect to a conventional PAC spectrometer Sample activities were moderate and did not challenge the acquisition system With an average trigger rate of 11 4 k s this is well below of what the DSP boards can handle In Section 3 4 it was previously discussed how the digitizers are very well capable of triggering at faster speeds Using Eq 80 a theoretical trigger rate of 1110 kHz for segment sizes of 400 samples is possible However tests with pulse generators have shown that the apparatus triggering saturates at rates around 825 kHz 47 74 6 RESULTS 6 2 PAC measurements Parameter unit Value Scintillator crystal diameter m 0 0381 Detector distance m 0 05 Solid angle fraction 14 5 Detected beginning sample activity Bq 0 315 106 Recording durat
118. s of the electric field gradient tensor vanish with the selection of the coordinate system z axis along the principal axis of the EFG so that in this case 56 reduce to 1 85 YW l Vr 0 4 T v 0 57 Lye Veo z 4 gn Va The matrix elements of the quadrupole Hamiltonian 54 are 2 An I M HoilI M CDU MQR MVE 58 m 2 Using the Wigner Eckhart theorem the matrix elements can be calculated as I I 2 I M Q2 I M 1 I MQ 1 M _M m TQ 59 where I Q Z is the reduced matrix element Under the assumption that the matrix element of a known ground state M I is not equal to zero the reduced matrix element can be calculated as I T2 2 IQ ER TQ 60 The conventional definition of the nuclear quadrupole moment 43 can be expressed in spherical coordinates using 55 1 1 1 167 Os op Th o Na 61 Inserting 61 and 60 into 59 delivers the following expression of the matrix ele ments of the quadrupole moment tensor 1 2 n gym DD I I 2 I I2 I M Q I M 1 9 4 Be IAIO 62 23 2 5 Static Electric Quadrupole Interaction 2 PAC THEORY Defining cf M 3M I I 1 UM 42M T M TFM 1 CM f Tim DTEM IFM DTFM 2 one can express the following 3j symbols in terms of cP M 14 I I 2 _4 I M ERO I 2 SEM IRB FM M 1 1 1 fr 1 M Ft eee N FM M 2 2
119. s thesis and for that she has endured her husband being in front of a computer screen many times Thanks to my children Lara and Luke which have beared a sometimes inadvertently behaving father during the first five months of their lives my cousine Nehle for letting me stay at her apartment during my visits in Braunschweig and my parents family and friends for being so challenging and teasing about my profession CONTENTS Contents 1 Introduction 2 PAC theory 2 1 Theory of Angular Correlation 2 2 Perturbation of Angular Correlation 2 3 Absence of Perturbation 2 22 4 4 2 ge Go PANGS 2 4 Polycrystalline Samples 2 5 Static Electric Quadrupole Interaction 2 6 Static Magnetic Dipole Interaction 3 Hardware 3 1 Overview of Setup and Operation 3 2 Handshake Bus 28 200 den ee lee fh otk ee 3 3 Standard parallel port aa ern ease te ae 3 4 Digital Signal Processing Boards 3 4 1 Scope settings ar vn sa ave Ae a ee eS 305 Host compiieer 23 Ene ok 2g nite eee Ete eh Eee land 3 6 Slave computers 2 Gi Bw oe we ee at 3 7 Analog multiplexer module 3 8 Reference clock distributor 3 9 Synchronization and Channel Select fanout module 3 10 Photomultipliers tse oh ho a ete AA a tials 4 Software 4 1 Parallel port device driver 2 4 4 2 282222 dwt u A 4 1 1 Installing the d
120. signals to be transmitted over a RG58 cable The module also features an integrated line driver circuit 25 which transduces the low power out puts of the SPP into four 50 Q impedance lines A resistor in series with the output matches the output impedance to the cable impedance This is required to ensure the signal strength at the receiver is maintained and reflections in the cable are sup pressed The schematic diagram in Figure A 3 3 in Appendix A 3 shows the CHAN NEL_SELECT bus signal driven by DATAO STROBE drives the SYNC_PULSE line Both signals are fanned out into four channels 3 10 Photomultipliers Photomultiplier tubes are extremely sensitive detectors of light in the ultraviolet vis ible and near infrared These detectors multiply the incident light by an amount of as much as 10 Even single photons can be resolved The incident photon strikes a highly negatively charged photosensitive cathode and releases electrons as the con sequence of the photoelectric effect These electrons are directed through a focusing electrode towards a cascade of electron multiplying dynodes The process of secondary electron emission leads to a measurable amount of electrons at the last instance of the cascade All electrons are collected at the positively charged anode The inte gral of the anode signal is proportional to the energy of the incident photon The amount of multiplied electrons rescinds the amount of positive charges at the anode
121. sly with the digitizer clock Therefore the time between the trigger and the next sampling clock varies randomly in time The true time reference lies within a fraction of the sampling clock The horizontal position parameter permits a very precise positioning of the acquired trace and is necessary for the correct interpretation of the CFDs resulting parameter As an example the variations in tefa for the isotope Hf Ta are demonstrated in Figure 9 S1Hf Ta has three distinct energy peaks that are useful to show the relationship of the photon energy and the time of arrival On average the timestamps for lower energies are corrected toward shorter times as they tend to be detected at a later time Higher energies appear to be detected at earlier times and therefore their correction term is towards larger times The CFD technique was applied on transient waveforms of the BaF detector s fast component In addition the waveforms were filtered according to their energy integral Only those y rays which fell into the energy windows were regarded On average the lower energy y rays are detected later than the higher energy y rays The first 136 keV appears to arrive about 800 ps later than the second 384 keV And the third 481 keV arrives even earlier than the second by almost 400 ps The application of the CFD technique also yields to the identification of false trigger events If the shape of the recorded transient deviates from a certain average th
122. st before switching over to the PMT channel If the energy integral of the first segment is of zero value the corresponding timestamp represents the chronological origin tno Thus within a segment subsequent timestamps tny need to be converted into absolute timestamps with respect to tno Coinciding events can hence be found 53 4 3 Host applications 4 SOFTWARE among absolute timestamps The Start and Stop are collected into time spectra D t according to formula Uris Dyl Atijr 1 if 0 lt At lt sizeof Di 87 D At jk otherwise Dij Atijk l where the indices and j are permutations of the PAC Slaves and sizeof D the size of the D j array The factor in Eq 86 shifts the time spectrums center point tp into the mid range of the array Figure 12 in Section 5 2 shows such a typical a time spectrum with the prompt peak located in the center of the graph 54 4 SOFTWARE 4 3 Host applications Intensity a u Intensity a u _ 1 f i A i aaa I i I 0 100 200 300 400 500 600 0 100 200 300 400 500 Energy keV Energy keV Intensity a u L L li 1 li 1 li 1000 1250 750 Energy keV Figure 11 Three plots showing the decay energy spectrum of a Hf Ta b In Cd and c Na The horizontal energy axis is scaled down to match the significant energy peaks The Start wind
123. stance Chronological resolution displayed in prompt peaks Time deviation between six channel pairs Sample activity over 456 hour experiment Duty cycle with respect to sample activity A2 effG22 t for In Od in 8 Sn metal 2 2 2 2 A22 effG22 t for In Cd in Cds Pa teas at ee eh A22 effG22 t for In Cd in a iron metal 2 2 2 2 2 20 A22 effG22 t for POP Hf Ta in Hf metal oaa LIST OF FIGURES LIST OF TABLES LIST OF TABLES List of Tables oo Nn no F WN FP Characteristics of isotope Y In Cd oo 9 Characteristics of isotope Hf Ta 2 22 10 List of command declarations in pacmod 2 2 2 sie an Bar ar 46 Pacslave s default settings 2 2 ur 0 2 Eu OE Sons SE OE Eek 58 Parametric time deviation Yan wet Wet ee ae 67 Results for Y In Cd in Fe experiment 04 75 Pin assignment of the D Type 25 pin standard parallel port connector 87 SPP I O register assignment ad ae Er ae ea 88 VII 1 INTRODUCTION 1 Introduction Perturbed Angular Correlation spectroscopy PAC is one of the many hyperfine methods used in nuclear solid state physics research and supplements methods such as Mossbauer spectroscopy and Nuclear Magnetic Resonance They are all based on the interaction between the spin momentum of specific nuclei and electric and magnetic hyperfine fields caused by the condition of the immediate neighborhood Such fields can
124. t to At marker t t the Host deasserts the CHANNEL_SELECT The Slave arms the DSP board and acknowledges this at t t The Host creates the synchronization pulse at t t3 Shortly after that it asserts the CHANNEL_SELECT at t ty ts symbolizes the same point in the cycle as tg A low dead time is given for ty tp amp ts t4 address causes the pins to adopt either high 5 V or low 0 V The status port is read only The control port is both read and writable Some signals are inverted and some are low active Table 7 in Appendix A 1 shows the pin assignment of the SPP connector 29 The ACKN signal is a special input pin It is capable of being used as an inter rupt source for the Programmable Interrupt Controller PIC A high low transition 5V 0V on the ACKN pin triggers the PIC and causes the microprocessor to ser vice a custom interrupt handler routine At boot time the computer BIOS routes this signal to the PIC This interrupt line is assigned to an integer number between 0 and 15 The typical assignment is interrupt number 7 To some extend the interrupt request number can also be assigned manually in the BIOS 3 4 Digital Signal Processing Boards The center piece of every PAC Slave is a fast transient recorderboard The recorder boards used in this spectrometer setup were manufactured by a swiss company called Acqiris Acqiris digitizers are very reliable and flexible recording devices They are also superio
125. teraction is known to be zero 43 The fact that iron exhibits sponta neous ferromagnetism leads to the presence of internal magnetic fields When exposed to a magnetic field the nucleus magnetic momentum is aligned with the magnetic flux The strength of the local magnetic field is proportional to the split of the nucleus dis crete energy levels This magnetic dipole interaction was measured in sample c A quadrupole interaction was practically non existing The measured time dependent anisotropy A22 effG22 t is presented in Figure 21 The coincidence data was recorded with a timing resolution of 250 ps For magnetic interaction the theory predicts that the time dependent perturbation factor G22 t is modulated with the Larmor fre quency wz For a nuclear spin of J 5 2 and the index k lt 2 as of Eq 78 the perturbation factor G22 t can contain integer multiple terms of the fundamental fre quency wz The obtained data shows clearly magnetic interaction with wz but there is evidence of the presence of a second site with a slightly lower Larmor frequency wr2 The anisotropy data was fitted according to Eq 104 with a two site variation The results for this fit are listed as follows Site fi wz Mrad s 1 91 3 558 8 0 2 2 8 7 384 8 1 1 80 6 RESULTS 6 2 PAC measurements 111 Cd in Fe foil A t Fourier intensity 0 20 40 60 80 100 120 140 160 180 00 05 10 15 t ns amp 10 rad s Figure 2
126. the acquisition systems limited timing resolution If a spectrum taken with this new PAC spectrometer shows attenuation on a sample from which is known that damping effects do not exists then it must be assumed that the apparatus does not work accordingly Measurements were performed on metallic alloys and insulators with In Cd and 81Hf Ta impurities Non cubic metals provide the most suitable environment for the observation of a quadrupole interaction The reservoir of conduction electrons in the metal brings the recoil atom produced by the nuclear reaction quickly into its equilibrium charge state A second advantage of metals over insulators is the tendency of the former to form alloys with impurities such that the impurity atom occupies a substitutional site in the lattice The metal samples in this work were in the state of solid pieces Impurities were introduced through melting and subsequent quenching or ion implantation The metal samples investigated were In Cd in a Fe In Cd in 8 Sn and 81Hf Ta in Hf metal In the cubic fce structure of a Fe the electric field gradi ent is zero In ferromagnetic materials however non random Weiss domains produce intrinsic permanent magnetic fields at room temperature that produce a magnetic dipole interaction with the nucleus For the two non cubic metal alloys Sn and Hf the interaction with the nucleus is known to be of quadrupole nature The quadrupole interaction is characterized so
127. time teycie For low activities tontine is dominant in regard to tof fine which is almost negligible Therefore 12 6 RESULTS 6 1 Performance the duty cycle approaches 100 The duty cycle values were obtained from activities recorded during the experiment Activities beyond 1 MBq were not available for experimental purposes due to lab conditions and safety regulations For a sample activity of 0 315 MBq the duty cycle of the apparatus was reduced to 87 The dashed line in Figure 18 was obtained from a fit according to Eq 83 The fitted function extrapolates the duty cycle beyond the measured values The development of the duty cycle starts out linear at first but is expected to approach 0 for higher activities According to these results the 50 duty cycle point of the apparatus can be estimated at a sample activity of 2 3 MBq The fitting parameter m toffine gains a value for the average dead time of tof fiine 0 14 0 04 s with m 8000 segments This value is well to expect if one assumes the dead time only originates from the relation lof fline trans tarm 102 The transfer time for transfering the raw waveform data over the PCI bus tirans plus the time for reactivating the recorder boards tarm add up to almost that value According to the manufacturers specification the time to arm the boards for the sequence mode is rated around tarm 0 1 s 22 The amount of data to be transfered is 3 2 MByte With the PCI bus at a sp
128. utput 1 Clock Output 2 Clock Output 3 Clock Output 4 Drawn by C Herden Copyright 2004 2007 Christian Herden Corvallis OR USA TITLE New Generation TDPAC Spectrometer Document Number Wiring and Design Plan REV 1 0 Date 21 Nov 2007 Sheet 1 1 90 A 3 Schematic diagrams A APPENDIX T T 3994S ErT ZT 8B 988C EC E EU JaAI3931 asind ous pue Joxajd jnu Boyeuy 4squnn uswn gt og s npou xnW BoTeuy 37LIL YSN YO sileA1od USPISH ueysuyo Aq UMeIG UapsaH UBNSUYD 900Z F0OZ JYBUAdOD INO9IS So62W1 an S082W7 sn pod jajle ed Ajddns samod EL cl LL OL NM nro Glo 081807 A inow FI8YN3 anD 193135 LNI 99 doo dzz wzestgzsd NRG YNI ano PNl ENI 22 WINO ENI N Eno uno Na ZNI LAND ZNI INl DA ON 3SINd ONAS wiestazsa ONQO LOSTASTSNNVHO ENI anod ELNO ZNI ELNO ZLNO Na ZLNO FLNO Na F1NO LLANO YNI 1LNO SIA INI 91 A APPENDIX A 3 Schematic diagrams T T 88US dg8Z 00 90 SBAZ ST T 70 JSAUP pajas aUUBYD pue as nd oug YSN UO Sieno Nd ONAS Nd INAS Nd INAS Nd ONAS MeaquNN U3UNDog yNoue SOH 3ILLL USPJSH uegsuyp Aq umeig UapsaH UBNSUYD 900Z F0OZ UBUKdoD ENI anod ELNO ZNI ELNO ZLNO N ZLNO FLNO Na F1NO LLANO PNI L1NO DIA INI ENI anod ELNO ZNI ELNO ZLNO N ZLNO FLNO Na F1NO LLNDO PNI L

Conceptual Design and Setup of a Fully Digital Recording

Contents

Download Pdf Manuals

Related Search

Related Contents