Home

User Manual - David Hoyle

1. beam would dissipate after exiting the column if it encountered atmosphere Also the imaging scheme involves the use of a particle detector that is mounted on the outside of the bottom of the column If the operating voltages are applied to this detector while it is exposed to atmosphere it can be damaged or destroyed Maintaining an adequately low gas pressure requires continuous pumping because gas molecules drift into the evacuated spaces from various sources within the system itself These sources include lubricants electrical cable insulation and atmospheric gases and water vapor that are adsorbed by the interior surfaces of the workchamber when the workchamber is opened for servicing Also whenever a mask plate is brought into the workchamber some adsorbed atmospheric gases and water vapor inevitably enter with it Finally the pressure in the workchamber normally rises when a process gas is introduced for deposition or gas assisted etching Vacuum System Schematic A schematic diagram for the vacuum system is shown in Figure 2A 800 000071 Rev A 56 Principles of Operation N2 Gas Cabinets f Wai Check N2 Source mn X Valve i 1 S 1 FIB Q N O valve A V34 PAA opens when TMP3
2. consists of internal CAN Controller Area Network and Ethernet networks These local networks are configured in a star topology with each station or node connected to a central hub thereby reducing system cabling Each major electrical subsystem resides on its own CAN or Ethernet node and each node has a dedicated processor that allows local decision making to take place Because the electrical subsystems are logically isolated from one another each subsystem can be upgraded without disrupting the entire system 800 000071 Rev A 34 Principles of Operation CAN Controller Area Network a communications protocol Ethernet a LAN communications protocol RS232 a communications protocol Front End MUI Manual User Interface AKA Knob Panel Laser Mask Front End mask loader load port module robot PC Loader BC 100 Beam Control scan control and acquisition Ethernet Ethernet Hub Control CAN PC BC 100 Detector USB Rack CAN elie lt ai s Network plitter Module Joystick Flood Sty Gas Gun Box MUI CAN knob panel ee eee CAN onverter RS232 CAN Network a M Br2 Gas Module Box Robo Servo DC CAN j l Supply System Module XeF2 H20 Gas Box Gas Box Work RS232 v RS 232 RS232 FIB Chamber gr
3. etc The ionized gas molecules being deficient in electrons and therefore positively charged are attracted to a negative electrode Their arrival at that electrode draws electron current from the power supply providing the negative potential and this electron current provides the measure of gas pressure in the envelope The CCIG works on the same ion collection principle that the HCIG employs except that instead of using electron bombardment to ionize the gas molecules it uses a strong magnetic field It is thus similar to an ion pump see Jon Pumps on page 60 Vacuum Pumps Turbopump A turbomolecular pump usually referred to as the turbopump pumps through a large opening in the bottom of the workchamber to obtain and maintain high vacuum in the chamber A turbopump is essentially a high speed fan with turbine like blades that rotate at about 40 000 rpm However it cannot achieve or sustain this rotational speed unless the gas pressure to which it is exposed is already relatively low Therefore it must be used in combination with an auxiliary vacuum pump called the foreline pump 4 that is connected to its exhaust outlet through a vacuum hose called the foreline See Auxiliary Pumps below The standard turbopump on the Accura 850 system is a water cooled Varian model 000HT Auxiliary Pumps The system is equipped with two auxiliary vacuum pumps a roughing pump and a foreline pump that are capable of achieving rough
4. executed in the computer is created Sometimes the term class is also used for the instantiated object Using COM objects or classes and their methods as well as associated data are compiled into binary executable modules These executables take the form of files with a dynamic link library DLL or EXE file name suffix A module can contain more than one class 800 000071 Rev A 30 Principles of Operation DLL A DLLisa Dynamic Linked Library a subprogram initiated by the main program only when needed to perform a specific task thereby saving space in memory For example as long as a user of Microsoft Word is editing a document the printer DLL file does not need to be loaded into RAM Random Access Memory If the user decides to print the document the Word application causes the printer DLL file to be loaded and run DLL files that support the operation of specific devices are known as device drivers A DLL file is often given a dll file name suffix DLL files are dynamically linked with the program that uses them during program execution rather than being compiled with the main program The set of such files or the DLL is somewhat comparable to the library routines provided with programming languages such as C and C User Interface At the top of the software structure is the graphical user interface GUT through which the user operates the system The UI layer consists of an EXE framework page controls and man
5. mask to be brought directly under the center line of the 3 For the same reason it is essential that when a faraday reading is taken the stage be positioned so that the beam passes through an opening in the grid If it strikes any part of the grid a significant amount of secondary electron emission from the affected area will escape upward thereby causing an erroneously high reading 800 000071 Rev A 54 Principles of Operation ion beam column that point may in fact be as far as 100 nanometers from the center line on either or both axes when the stage comes to rest The software can compensate for this position error by in effect moving the ion beam column s scan field off center but it must know what the error is This information is provided by laser interferometers or linear encoders for each axis Interferometry Refer to Figure 20 which illustrates how interferometry works on one axis Simply ion column reflector 2 reflector 1 interferometer axis drive mechanical power motor transmission 1 pulse 100 nm 1 pulse 5nm Figure 20 Stage Position Tracking put the interferometer directs a laser beam along the axis to a mirror mounted on the side of the stage reflector 1 and compares the wave length of the reflected light with that of the transmitted light While the stage is moving along the axis toward the interferometer the phenomenon known as the Doppler effect causes the wave length of the reflecte
6. object or class instance is what runs in the computer The methods of each instance provide computer instructions and the class object characteristics provide relevant data The user can communicate with objects and they can communicate with each other via well defined interfaces called messages The concepts and rules used in object oriented programming provide these important benefits 800 000071 Rev A 29 Principles of Operation e The concept of a data class makes it possible to define subclasses of data objects that share some or all of the main class characteristics This property of OOP known as inheritance forces a more thorough data analysis reduces development time and ensures more accurate coding e Since a class defines only the data it needs to be concerned with when an instance of that class an object is run the code will not be able to accidentally access other program data This characteristic of data hiding provides greater system security and avoids unintended data corruption e The definition of a class is reusable not only by the program for which it is initially created but also by other object oriented programs and for this reason can be more easily distributed for use in networks e The concept of data classes allows a programmer to create any new data type that is not already defined in the language itself COM Microsoft s Component Object Model COM is a way for software compone
7. pattern described above is used for image acquisition with one low dose short dwell time frame producing each image To obtain the milling effect many frames each delivering a relatively high dose long dwell time are superimposed To obtain deposition thousands of low dose frames are superimposed while the deposition chemical flows onto the specimen When thousands of normally rastered frames are superimposed in one location the minute dosage that the beam delivers to the mask outside the raster as it is blanked travels into occultation at the end of each scan line has a cumulative effect This results in spurious milling or deposition an effect known in either case as blanking tails For this reason serpentine scanning is preferable whenever large numbers of frames are to be superimposed because it decreases the number of times the beam is blanked by a factor equal to the number of frames times the number of scan lines per frame Regardless of which scanning mode is chosen the beam is always blanked after the last scan line of each frame while it is being deflected back to the start of the first scan line of the next frame The total time between the end of one scan line and the beginning of the next is called the retrace time This time consists of the actual deflection time and a programmable delay The the entire cycle time between the start of one frame and the start of the next is known as the refresh time Refresh time is necessa
8. that lights one dot or pixel on the CRT to a particular gray level However the image data may be translated by the software in various ways before it becomes the RGB signal that illuminates pixels on the CRT Imaging Modes The FIB system uses any of three techniques for capturing secondary emission from the specimen electron imaging mode ion imaging mode and a variation of electron imaging mode called multiplex imaging mode In the electron imaging modes the particle detector is electrically biased to attract and capture electrons in ion mode it is biased to attract and capture positive ions Electron imaging mode often produces clearer images than ion imaging mode because the secondary electron yield from most materials is greater than the secondary ion yield and the imaging signal 800 000071 Rev A 45 Principles of Operation also tends to be less noisy Also a lower ion dose can be used for image acquisition However if the area being imaged is electrically nonconductive being either an insulator e g quartz or a small island of conductive material e g chrome surrounded by insulating material it quickly develops a positive electrostatic charge This is due to the loss of electrons through secondary emission and to some extent by the implantation of positive gallium ions This positive charge can prevent secondary electrons from escaping from the surface of the specimen thus rendering the mask invisible to the imaging p
9. these imaging modes the retrace time is the same and the transition time between dwell points is negligible However for multiplex imaging mode the transition time between dwell points and the retrace time are extended to allow charge neutralization and therefore become significant factors Accura 850 systems provide three imaging speeds slow medium and fast The overall image size see below is also a factor in image acquisition time Image Size The term image size is generally used in reference to the area of the J computer screen that is occupied by the image produced by a raster The application all 256 program enables the user to select through an array of graphics such as those shown here at right four progressively larger image sizes ranging from 256 by 256 to 1024 512 by 1024 The numbers refer to the number of pixels on the computer screen hence image size as well as to the number of dwell points in the image rasters that produce 768 the images there is a one to one correspondence The numbers are not related to the dimensions of the area being scanned to produce the image which is determined by ul 1024 the field of view selection However they do relate to resolution the larger the image size selected for a given field of view the greater the dwell point density and therefore the greater the resolution and apparent magnification Pixel Averaging Pixel averaging involves sampling the signal from the detector for a
10. to communicate with controllers and actuators brain and brain and brawn boards at each network node CAN or Ethernet 800 000071 Rev A 32 Principles of Operation FIB System User Application aie APP Application Server Server Process interface between UI layer and Hardware Server Object Model Layer OML high level control of overall tool behavior Behavior Layer BHV controls interfaces between subsystems Hardware Server Process Model Layer MDL controls subsystem behavior Hardware Abstraction Layer HAL communicates with hardware controllers and actuators CAN Ethernet USB Figure 8 Software System Organization All the objects within the hardware server are implemented as bricks Bricks control distinct software subsystems programs within the XT100 These subsystems control specific hardware components The behavior layer within the hardware server manages the interaction of the different subsystems 800 000071 Rev A 33 Principles of Operation A Brick is a Microsoft COM Component Object Model component implemented as a DLL A DLL is a Dynamic Linked Library a subprogram initiated by the main program only when needed to perform a specific task thereby saving space in memory A group of Bricks is managed by a BrickBox and each BrickBox runs in a separate Windows NT process The configuration of the distribution of Bricks over BrickBoxes is stored in
11. vacuum pressures of 1 3 Pa or 10 2 torr or lower These pumps both of which run continuously are used as follows 4 The foreline pump may also be referred to as the forepump or the backing pump 800 000071 Rev A 59 Principles of Operation e During system pumpdown when vacuum is being restored after the workchamber has been vented to atmospheric pressure both pumps work in parallel to evacuate the entire system to rough vacuum so that the turbopump can then be switched on e While the turbopump is running the forepump runs maintains adequately low pressure in the turbopump s foreline e When a mask has been placed in the vented loadlock and the loading sequence is to begin the roughing pump evacuates the loadlock so that the chamber door V6 can be opened without resulting in a major pressure surge in the workchamber See page 57 for more information about the loadlock The standard foreline pump for the XT100 system is the Alcatel model ACP 20G The standard roughing pump is the Alcatel Drytel model 34C Both are dry pumps meaning that they do not use liquid lubrication pump oil for their moving parts The auxiliary pumps produce mechanical vibrations and therefore must reside outside the process module enclosure to minimize the transmission of those vibrations to the workchamber They may be installed behind the FIB system as shown in Figure 22 or on subflooring below it A foreline isolator is in
12. 0 000071 Rev A 61 Principles of Operation Clear Repairs Clear defects missing absorber material are repaired by carbon deposition the layering of a carbonaceous film onto the mask The source of the film is a hydrocarbon solvent commonly known as styrene or vinyl benzene A small amount of this chemical is stored in a cooled and insulated cylinder in a dedicated chemical cabinet The cylinder is cooled by a solid state heat transfer device with an air cooled heat sink to about 10 C below ambient to minimize condensation of the vapor elsewhere in the delivery apparatus The delivery valve is opened and the vapor is flowed onto mask in the vicinity of the defect while it is being rastered by the ion beam Much of it dissipates into the workchamber and is removed by the vacuum pumps but a monolayer is adsorbed by the surface of the mask As the ions pass through this monolayer they decelerate rapidly which results in an energy transfer that dissociates the gas molecules and fixes the carbonaceous material to the surface of the mask Many repetitions of the raster are conducted to build up successive layers of the material until the desired thickness is obtained Thus the rastering technique for carbon deposition involves many iterations of a raster that precisely administers a relatively low ion dose if the dose is too low the material will not adhere if it is too high the ion milling effect will dominate Serpentine scanning is typically use
13. Left Gas Box Doors Figure 5 Electronic Packaging Top View 800 000071 Rev A 26 Principles of Operation UI Cabinet Wing Panel Left Right Electronics Cabinet is Electronics Front End behind this panel Cabinet Light Tower Styrene Gas Box Alarm Panel UI Monitor Br2 Gas Box Load Port Module AKA SMIF Pod MUI Assembly XeF2 Gas Box Video Printer Heo Ges Box Control PC Figure 6 Electronic Packaging Front View 800 000071 Rev A 27 Principles of Operation Right Process ken Electronics Module Electronics Cabinet Cabinet Styrene Gas Box FIB Optics Flood Gun Flood Gun Controller Controller Detector Rack lon Column 4 TMP3 IGP IGP Cont Cont Cont as l BC 100 l Acht DC Supply 4 b S Frozen Bromine EOCU DC Supply gt 5 Controller XeF2 Gag Robot Servo DC Supply Box FIB Gun Supply UPS Battery Race Facility Power Distribution Unit Work Chamber Motion Chassis Vacuum Panel Figure 7 Electronic Packaging Rear View The knobs and switches on the knob panel communicate with the computer via the XT100 local area network The software determines the states of the various knobs and switches by interrogating the knob panel interface The alphanumeric keyboard is connected directly to the computer in the conventional manner 800 000071 Rev A 28 Principles of Operation Software Architecture Figure 8 is aconceptual diagram of the l
14. Principles of Operation System Overview Figure 4 shows a block diagram of an XT100 system The repair process begins when the operator loads a mask carrier also known as an RSP or Reticle SMIF Pod into the load port module The operator initiates the automated sample loading process by clicking the Load button on the Load Unload dialog box in the software user interface A robotic system transfers the mask from the load port module to the loadlock at the front of the work chamber Next the loadlock door closes and the loadlock is pumped down to low pressure so as not to introduce abnormally high pressures in the workchamber After the loadlock is pumped down the mask is transferred via another robotic transfer system to the mask holder on the sample positioning stage An automated clamping positioning system on the mask holder aligns the mask and the mask is then ready to be positioned for imaging and repair procedures The XT100 system is comprised of several subsystems each of which has its own software subsystem power connections communications node and hardware controller The hardware and cabling for these subsystems are mounted in racks in the left and right electronic cabinets and are normally not accessible to the system operator Figure 5 shows a top view of the major system components Figure 6 and figure 7 show the location of the modules in the right and left electronics cabinets 800 000071 Rev A 24 Principles of Operat
15. agers UI layer managers Refer to the XT 00 Program Reference Manual part number 800 0000XX for a description of that program The Application Server The objects within the application server provide an interface between the UI and the hardware server These objects in this server are built using MS COM and are designed to be very generic in the functionality they expose For example the Image Analysis component is designed to expose only generic image analysis functionality with no XT100 contextual information The application server contains the following objects e Recipe Page Manager e Coordinate Lock Manager e Defect Filter Manager e Image Analysis Manager e Repair Page Manager e Chamber Schematic Manager e Manual Navigation Manager e MUI Controller e User Access Manager The Object Model Layer The Object Model Layer serves as the interface between the application server and the hardware server It consists of objects that are implemented as COM components 800 000071 Rev A 31 Principles of Operation The Hardware Server The hardware server consists of programs that run more or less autonomously to maintain continuous control of certain major FIB subsystems The hardware server also provides the interface with the XT100 hardware It consists of the following three layers BHY Behavior layer to control interfaces between subsystems MDL Model layer to control subsystem behavior HAL Hardware abstraction level
16. ayered structure of the FIB system software Commands pass downward from the user to the FIB system hardware and status information passes upward from the hardware to the user The Operating System The foundation for the entire software system is Microsoft s Windows NT operating system The operating system is a collection of software that controls the basic operation of the computer and that allows the execution of application programs by allocating memory space passing information to and from hardware devices and performing other related tasks As an operating system Windows NT includes device driver programs for operating the standard peripheral devices with which the computer is supplied disk units communications ports and so forth In addition to these device drivers the FIB system manufacturer augments the operating system with device drivers for moving information to and from the special hardware interfaces such as the vacuum controller that are unique to the FIB system The Application Program At the top of the software structure is the application program whose primary function is to present the graphical user interface GUD through which the user operates the system Refer to the XT 00 Program Reference Manual part number 800 0000XX for a description of that program Software Design Principles The XT100 software architecture is designed to be extendible upgradeable and easy to maintain It consists of reusable software compo
17. can be placed at any point within the range of deflection Also whereas the voltage applied to a blanking deflector is either zero or some predetermined fixed value the voltages applied to a deflection assembly are scaled according to where the beam is to be placed During imaging milling or deposition the ion beam is deflected back and forth along a series of parallel paths in what is called a raster pattern see Rastering below To achieve this various combinations of ramped voltages are applied to the deflection assembly to move the beam along the appropriate paths Typically the area that a raster has to cover is smaller than the deflector s maximum scan range When this is the case the raster producing scan signals that are applied to the deflector can be biased so that they offset the center of the raster from the center of the deflection range an effect known as panning The pan knobs on the knob panel allow the user to pan the scan field when imaging with small fields of view and the application software also uses the panning effect to correct for stage positioning error when it brings work sites on the mask under the ion column Rastering Rastering means scanning the beam such that it traces a pattern of closely spaced parallel lines as shown in Figure 13 One iteration of this raster scanning pattern is called a frame It is also referred to as a raster refer to the Glossary for further discussion of raster Each of the scan
18. ctor Thus the extractor voltage is the main determinant of the rate of ion emission The 800 000071 Rev A 37 Principles of Operation high voltage server attempts to use the lowest extractor voltage that provides a stable rate of emission Although the extractor voltage is the main determinant of ion emission it is not the only one Various phenomena conspire to cause the rate of ion emission to vary slightly over time To correct for these ion source 0to2kV 0 to 30 kV l suppressor acceleration extractor amp spray aperture end Do n 20 to 20 kV f stigmator z upper octopole V limiting Ooo ooo rrr apertures ea Ze ion gun isolation valve VGI blanking deflectors e c blanking aperture yi C z lens 2 deflection assembly lower octopole al stage Figure 10 lon Column Architecture 800 000071 Rev A 38 Principles of Operation variations in the interest of providing a constant beam current at the specimen an element known as the suppressor is used Because of its proximity to the source a relatively small and therefore easily controlled positive potential on the suppressor counteracts the extraction voltage and thereby reduces the amount of emission The regulation technique then is to set the extractor voltage for more emission than is required and to hold the emission at a desired setpoint by dynamically adjusting the suppressor voltage The HV server monitors t
19. current density or to a much lesser degree the spot size for a given beam limiting aperture a virtual aperture can be implemented through the application software When the user selects a virtual aperture a specified hardware aperture is selected but the software sets the Lens 1 focus voltage according to an alternate lens table Stigmation The upper octopole also known as the stigmator is used to introduce correction for astigmatism which is the failure of the column to make the ions converge uniformly to a circular spot Astigmatism can be expected to develop as the beam enters Lens 1 because despite the passively collimating effect of the spray aperture see above not all the ions entering the lens will be on parallel paths nor will they be evenly distributed about the central axis The focusing effect of the lens is therefore not uniformly applied and the ions do not become uniformly convergent Figure 12 a attempts to illustrate the nature of the problem by representing the beam as having become elliptical in cross section instead of circular as itemerges from Lens 1 the result of having become more convergent in y than in x Note that the dots representing the ion paths are not evenly distributed Thus the problem is not simply a matter of the cross sectional shape of the beam being non circular which could be corrected by a limiting aperture but involves the spacing of individual ion paths and Lens 2 would be expected to ampli
20. d light to appear shorter than the wave length of the transmitted light Conversely while the stage is moving away from the interferometer the apparent wave length of the reflected light appears longer than that of the transmitted light Interference patterns resulting when the two light beams are combined provide detectable pulses that can be directly correlated with distance moved The digital counters that register stage position are reset to zero when the stage is at the Stage Zero reference position and then are incremented by pulses produced by the interferometers as the stage moves away from the reference position and decremented as it moves back 800 000071 Rev A 55 Principles of Operation A reference mirror mounted in a fixed position in proximity to the ion beam column allows the interferometer to detect and compensate for relative motion between itself and the stage that results from thermal expansion or contraction of the workchamber The net effect is that the interferometers indicate motion in increments of 5 nanometers The arrangement illustrated in Figure 20 is duplicated at 90 for the other axis The mirror for each axis is wide enough and carefully aligned so that motion on one axis is undetectable by the interferometer for the other axis Linear Encoders If an interferometer indicates an error condition for example an interruption of the laser beam by a foreign object stage motion feedback to operate the posi
21. d to minimize the blanking tail effect see Rastering on page 43 for related information Opaque Repairs Opaque defects unwanted absorber material are repaired by milling sputter etching the material away with the ion beam The relatively massive and highly accelerated gallium ions bombard the material breaking molecular and atomic bonds and scattering the loosened particles more or less randomly in all directions Most of the debris is propelled away from the worksite and is removed from the workchamber by the vacuum pumping apparatus although without the use of process chemicals a certain amount can be expected to be redeposited on the mask in the area being scanned Chrome GAE The chrome GAE gas assisted etching repair process involves the introduction of a process gas consisting of a mixture bromine and water vapor to the defect area while it is scanned by the beam This gas reacts with the sputtered chrome to form molecules that are volatile in that they tend not to stick to whatever surface they might strike but move about in the workchamber until they are removed by the vacuum pumps This reduces the amount of material that is redeposited on the mask The presence of the gas also reduces the sputtering effect of the ion beam on the quartz glass substrate The net effect is an enhanced selectivity of chrome removal and a reduction in the ion dose required for a given repair This dose reduction is significant because the relativel
22. drivers other controls operate by requesting actions through the stage positioning server Stage motion for the purpose of positioning defect sites on the mask under the ion column is normally controlled by the application software through the positioning server based upon inspection data provided for the mask Stage Zero The stage s position or address is registered by hardware counters as its displacement in x and in y from mechanical switches that define the right rear limit of its travel see Figure 15 Thus the stage s position address increases as it moves away from this reference position and decreases as it moves toward it The reference position is referred to as Stage Zero The stage 800 000071 Rev A 49 Principles of Operation can be sent to this position zeroed whenever the indication of its current position becomes suspect for example when system power is restored after having been interrupted Stage Labels The addresses of various stage positions are recorded in files called stage labels so that the positioning server software can send the stage to any of them when requested to do so by the application program Included are stage labels for the load position in which the stage is docked such that a mask can be moved between it and the loadlock as well as stage labels for placing the geometric centers of masks of various sizes are under the ion column and stage labels for placing diagnostic fixtures see page 52 o
23. e l Beum CAN acl Hys Robot node Figure 9 Communications The Controller Area Network The majority of the electrical subsystems of the XT100 communicate via a CAN Controller Area Network CAN is based on the so called broadcast communication mechanism This broadcast communication uses a message oriented transmission protocol It does not define stations and station addresses it only defines messages Each message is identified with a message identifier A message identifier has to be unique within the entire network and it defines not only the content but also the priority of the message This priority is important when several stations are competing 800 000071 Rev A 35 Principles of Operation for bus access The message with the lowest node ID wins bus access in the XT100 the PC has the lowest node ID Losing nodes keep trying until the bus is available A high degree of system and configuration flexibility is achieved as a result of this content oriented addressing scheme It is very easy to add stations to a existing CAN network without making any hardware or software modifications to the existing stations as long as the new stations are receivers This concept of modular electronics allows multiple reception as well as the synchronization of distributed processes Data required by several stations can be transmitted via the network making it unnecessary for each station to know who
24. eam current than the beam blanking aperture element because the walls of the cup usually capture all of the secondary particles that are emitted in the cup as a result of ion bombardment Although a set of elements is installed 800 000071 Rev A 53 Principles of Operation ion faraday electron faraday silicon grid temperature sensor Figure 19 Diagnostic Fixtures on the Stage in the column in proximity to the beam blanking aperture element to suppress secondary electron emission while trapping positive secondaries some electrons or other negative particles may still escape to other parts of the column thus drawing extra electron current through the metering circuit and yielding beam current readings that are slightly high 3 However in general the blanking aperture is a preferable metering point for beam current because obtaining a reading does not require the stage to be repositioned The Electron Faraday The electron faraday measures the electron output from the flood gun It is physically and operationally similar to the ion faraday see above but it measures electron current from the cup to electrical ground instead of vice versa and it does not have the grid structure across its opening Stage Position Tracking The piezoelectric drive motors for the FIB system s stage can position the stage to within 100 nanometers of an x y address called for by the software Thus when the software calls for a particular point on the
25. ely exhibit a relatively narrow energy spread they are all travelling at about the same velocity and therefore are more suitable for focusing by Lens 1 Lens 1 Lens 1 is a three element structure in which each of the elements is operated at a different electrical potential The electrostatic field gradients produced by the differences in potential causes the paths of the ions near the edges of the beam to bend inward toward the center thereby making the beam convergent The uppermost element of Lens 1 is operated at the same potential as the extractor which means that the magnitude of the extraction voltage changes the focusing effect of the lens and the lowermost element of Lens 1 is grounded Assuming that the extraction voltage remains constant which is normally the case it is the potential that is applied to the middle element of the lens that is increased or decreased to vary the focusing effect of the lens Lens 1 is used to set the beam current by determining the diameter of the beam as it arrives at the beam limiting aperture element If the diameter to which the lens focuses the beam is equal to or less than the diameter of the beam limiting aperture the current density the number of ions per unit area that pass through the aperture is maximum If Lens 1 is used to defocus the beam so that its diameter exceeds the size of the aperture the current density downstream from the aperture is diminished Figure 11 illustrates the princi
26. field and into occultation Although its motion is very rapid it still delivers a finite ion dose along the way Depending upon the application this may result in a cumulative effect known as blanking tails See Rastering below for related information 800 000071 Rev A 42 Principles of Operation Lens 2 Whereas Lens is used to adjust the beam current density Lens 2 is used to focus the beam on the specimen This is the lens that the user controls with the Focus knob on the knob panel or the software presets Lens 2 applies the same principle as Lens 1 to cause the ions to converge as they travel toward the focal plane the plane in which the ion trajectories cross and therefore the plane in which the spot size is smallest The Deflection Assembly The deflection assembly includes an octopole that is similar to the stigmator but is operated differently In the stigmator opposing pairs of deflectors are used to exert lateral push push or pull pull forces on the beam as necessary to correct for astigmatism Once the appropriate combination of correctional voltages is established it can be maintained until some event that changes the beam s astigmatism occurs In the deflection assembly opposing pairs of deflectors are used dynamically to exert lateral push pull forces to deflect the beam from the central axis Thus they work somewhat like the blanking deflectors except that with four pairs of deflectors the beam
27. fy the effect The stigmator consists of eight radially oriented electrostatic deflection elements Voltages of equal magnitude and polarity are applied to opposite elements so as to produce opposing forces that tend to reshape and collimate the beam Figure 12 b attempts to illustrate this effect and Figure 12 c attempts to illustrate the result The stigmator is used empirically to obtain the most sharply focused spot for the column as a whole and its correction therefore represents some combination of undoing the astigmatic effect of Lens 1 and anticipating the astigmatic effect of Lens 2 Minor differences in the forces exerted by a given voltage on opposite poles of the stigmator due to manufacturing tolerances are corrected by introducing voltage offsets through a service procedure known as the quad balance calibration Because the stigmator is located above the beam limiting aperture it can also used to deflect the beam slightly to center it on the aperture an effect that is known variously as fine align upper shift and beam align Blanking Given the amount of electrical power that is required to produce ion emission and the complexity of maintaining stable emission the ion beam cannot be rapidly switched on and off at the source Instead the combination of the blanking deflectors and the blanking aperture is used to achieve this switching effect as follows 800 000071 Rev A 41 Principles of Operation a cross sectiona
28. he rate of ion emission and performs this regulation function automatically The monitoring technique works as follows Many of the ions that leave the source follow paths that do not take them through the annular orifice in the extractor the spray aperture but in fact land on and are absorbed by the extractor electrode and the metal surrounding the orifice in the spray aperture Typically the gallium ions are singly charged deficient in one electron and each ion absorbed by the extractor is neutralized by an electron supplied by the power supply that develops the extractor voltage Thus there is a correlation between the electron current flow in the extractor circuit which is relatively easy to meter and the rate of ion emission from the source The current generated in the extractor circuit is known variously as the extractor current and the extraction current The HV server program adjusts the suppressor voltage to maintain an extractor current setpoint which is typically about 2 microamperes high enough for a stable supply of ions that will provide the desired beam current for the work at hand but low enough for reasonable source life A liquid gallium source operated at 3 microamperes can be expected to provide stable emission for at least 1000 hours A spray aperture is at the bottom of the extractor and it allows only those ions that are travelling more or less parallel with and close to the central axis to pass These ions collectiv
29. icron centers It is used as an imaging target for various system alignments When the mask plate is registered on the stage its top surface is in the same horizontal plane as the top surface of the silicon grid 800 000071 Rev A 52 Principles of Operation Photograph of Mask Carrier my nomenclature for the RSP or Reticle SMIF Pod which protects the mask plate reticle against abrasion and particulate contamination while it is being transported to and from the FIB system Figure 18 The Mask Carrier The lon Faraday The ion faraday is a small cup like structure used to measure the ion beam current being delivered by the column It provides an alternative to using the beam blanking aperture element page 41 to measure the beam current Also the ion faraday has across its opening a metal grid structure that serves as a convenient imaging target for performing various alignment tasks such as stigmation The grid is about 3 mm in diameter the bars in the grid are roughly 30 to 50 microns wide and the spaces between the bars are about 110 to 135 microns wide To measure the beam current with the ion faraday the stage is positioned so the beam can be directed through an opening in the grid and onto the floor of the cup The cup is connected to electrical ground through an electron current metering circuit in essentially the same manner as is the blanking aperture The ion faraday cup may provide a slightly more accurate measure of b
30. ion SMIF Standard Mechanical Interface MCP Microchannel Plate Mask Carrier AKA RSP Chemical Cabinets Gas Nozzle Funnel Front End mask loading Mask Holder Load Port Module AKA SMIF Pod Flood Gun Workchamber _ Gage MCP Stage positionin Be g Chamber Mask Robot Transfer Loadlock Robot Turbo Pump Transfer Robot Electronics Figure 4 System Block Diagram side view 800 000071 Rev A Reticle SMIF Pod 25 Principles of Operation The system computer is a workstation running Microsoft s Windows NT operating system The application program presents a graphical user interface GUI on the monitor interprets input from the mouse the knob panel and the keyboard and issues the appropriate hardware commands via a local area network to the various hardware control interfaces located in the electronics modules See Software Architecture below for related information Access to front and Access to rear of rear of electronics electronics cabinet cabinet UI Cabinet UI Wing Panel Electronics pain away Cabinet acuum Fane Right Front End Workchamber Load Lock Load Port Module r I I l l I l ee d L Facility Service Panel Monitor Electronics Gas Boxes Process Module Cabinet
31. ion column the ion source has a needle like tip coated with liquid metal gallium A current regulated power supply heats the source slightly in normal operation to maintain the liquidity of the gallium It also allows the source to be heated to a higher than normal temperature when necessary to eliminate impurities and to otherwise condition it for emission The source is held at a very high positive potential relative to ground This potential is called the acceleration voltage because when the ions are extracted from the source it is the force exerted by this difference in electrical potential that causes them to accelerate through the column to reach ground potential The higher the acceleration voltage the faster the ions are travelling as they exit the column and the greater the energy they impart to the specimen The high voltage server normally brings the acceleration voltage gradually to 30 kV when requested by the application program to power up the ion column Ion emission from the source occurs primarily because of the difference in electrical potential between the ion source and the nearby extractor electrode The resultant force draws the liquid gallium downward into a characteristically pointed shape known as a Taylor Cone the gallium atoms at the tip of the cone become excited such that they begin to lose electrons thus becoming by definition positive ions and having assumed a positive charge are drawn from the source to the extra
32. is off GRVi 1 v32 Qvas 1 Q IaPig QH TMPs PVP3 LS OP syrene 8 n2 P A as hae fy GIVi lt GWi var check LLPG sou Q Vave IGPic H T S ZK Q BPvi LLVS wea CLV LLVS he vss r amp v36 NY 1 1 V46 i l amp Q V66 Loadlock va3 BT4 v56 b i i es 3 Q2 ee tah LLIV j vet Check N Vave Ghani one LLVV V64 1 WRG K 4 TMP1 TMP2 1 v62 Q i v65 i ow QH S Q xeF2 Na e EREA E EER Purge ur Leche inlet Q CRV LLBV Q Vave LLRV LLSRV i 8 PVP1 QO FLPG 1 V54 1 Z 8s l i PVP2 1O 5 eon RLPG Valves GIVi Gun Isolation Valve FIB GV Vi Gun Vent Valve FIB GRVi Gun Roughing Valve FIB gun BPVi By Pass Valve FIB column Pumps Gauges d oe IGPig lon Getter Pump FIB Gun RLPG Pirani Vacuum Gauge Roughing eM K column poata Ayo FIB not used IGPic lon Getter Pump FIB Column RLPG Pirani Vacuum Gauge Loadlock GAY Chad nat Ba h i i CHPG Pirani Vacuum Gauge Chamber not used av va Chan har Vent Wak ae PVP1 Pre Vacuum Pump Chamber FLPG Pirani Vacuum Gauge TMP1 Tey Pai a i aye Sid PVP2 Pre Vacuum Pump Loadlock WRG Wide Range Gauge These atin Toad Ven not used PVP3 Pre Vacuum Pump FIB Gun HCIG Hot Cathode lon Gauge not used ane ws ahs ea EY i CCIG Cold Cathode lon Gauge not used TA E dig event ve i alve TMP1 Turbo Molecular Pump Chamber BTx Chemical Cabinet Baratron Gauge SRN B E k Soi Sag Fes TMP2 Turbo Molecular Pump Load
33. ketone PEEK and support the mask approximately 2mm in from the side and rear edges Small drive motors are then employed to bank the mask squarely against three rotating banking pins that are also manufactured from PEEK The banking mechanism which is mounted underneath the mask holder applies a light force to the lower right corner of the mask pushing it left against two alignment pins and forward against the remaining pin The banking mechanism consists of a 12mm diameter cam driven by an 8mm diameter DC motor The cam directs a pivoting v block riding on a low profile ball slide which applies a spring force to the mask and holds it firmly in place When the mask is in position it is recognized by a presence sensor as being properly in place and the stage is then free to move Figure 16 shows the mask holder and Figure 17 shows a mask positioned on the mask holder The Mask Carrier RSP A mask carrier also referred to as an RSP Reticle SMIF Pod is a square pan like carrier that holds a mask plate of a particular size The mask carrier protects the mask plate or reticle against abrasion and particulate contamination while it is being transported to and from the FIB system Loading amp Unloading To load a mask the user places an RSP Reticle SMIF Pod at the top of the Load Port Module also referred to as the SMIF pod The user then opens the Load Unload dialog box by clicking Load Unload on the Menu Name menu Next the user
34. l representation of the b corrective forces exerted by the stigmator beam before stigmation c astigmatism reduced Figure 12 Stigmation With no potential applied to the blanking deflectors the beam passes undisturbed through a large orifice in the blanking aperture element This is the unblanked state When voltages of opposite polarities are applied to the blanking deflectors the beam ions being positively charged are repelled by the positive deflector and attracted by the negative deflector the net effect being that the beam misses the orifice and is absorbed by the aperture element This effect is known as blanking the beam When the column is powered up but not in actual use the beam is blanked The amount of beam current that the column is delivering to the mask is an important factor in the milling and deposition processes The blanking aperture element provides one means of measuring the beam current in the column downstream from Lens 1 and the beam limiting aperture The blanking aperture element is connected to electrical ground through a DC metering circuit While the beam is blanked the potential created by the arrival of positive ions draws electron current from ground through this metering circuit The beam current can also be measured with the ion faraday see page 53 When the beam is deflected onto the blanking aperture by the blanking elements it does not instantaneously disappear it moves out of the normal scan
35. le the imaging mode is multiplex and it is blanked its emission suppressed while the imaging mode is electron Thus in general the operation of the flood gun is automatic and operator intervention is not required Although if the user images a mask in electron mode or multiplex mode and then proceeds with milling or deposition without switching to ion imaging mode the flood gun will remain blanked and no charge neutralization will take place Depending upon the circumstances this may or may not be desirable The flood gun assembly consists of an electron gun and a cylindrical structure sometimes called the snout which incorporates an electrostatic lens and a vertical deflection element The electron gun itself comprises several electrodes refer to Figure 14 e Filament a cathodic electrode consisting of either a small tungsten disk or a short length of tungsten wire A direct current the filament current is passed through it to produce thermionic emission of electrons The power supply that is the source of the filament current is floated on biased by a negative potential produced by a power supply called the cathode supply such that the filament itself is negative with respect to ground The electron current that flows from this cathode supply to ground is called the cathode current and it serves as a measure of the output of the gun e Anode Electrons emitted by the filament are accelerated toward and exit through a grounded an
36. lines constituting the raster scanning pattern is in fact a linear array of discrete points to which the beam is rapidly deflected and at which it pauses or dwells for a predetermined period These points are called dwell points and the time for which the beam dwells at each dwell point is called the dwell time The dwell point spacing also called DAC spacing is the distance between the centers of adjacent points Raster scanning is used for milling deposition and imaging Two raster scanning schemes are implemented normal and serpentine In normal raster scanning the beam proceeds along each scan line in the same direction It is blanked see page 41 at the end of each scan line while the deflection electronics are directed to place it at the opposite end of the next scan line It is then unblanked and deflection proceeds Except for multiplex imaging rasters the beam is not blanked between dwell points multiplex imaging is explained under Imaging Modes on page 45 The act of redirecting the beam from the end of one scan line to the opposite end of the next scan line is called retrace Serpentine scanning involves reversing the scan direction on alternate scan lines thus eliminating retrace and the need to blank the beam while it is taking place The normal raster 800 000071 Rev A 43 Principles of Operation dwell points dwell point spacing frame one iteration of raster pattern Figure 13 Raster Scanning scanning
37. lock fais as ross o A Sa he alve TMP3 Turbo Molecular Pump FIB gun Sensors Loadlock Roughing Valve LLVS Loadlock Vent Sensor LLBV Loadlock Backing Valve Vx1 6 Chemical Cabinet Valves Figure 21 Vacuum System Schematic The Loadlock The loadlock is an air lock that allows the mask to enter and leave the workchamber without requiring the workchamber to be completely vented and re pumped The loadlock opens to atmosphere through an outer door called the loadlock door and it opens to the workchamber through a wide vertically operating gate valve called the chamber door CLV in Figure 21 Venting To minimize adsorption of atmospheric gases when any part of the vacuum system must be vented to atmospheric pressure nitrogen gas N gt is admitted to equalize the internal and external 800 000071 Rev A 57 Principles of Operation pressures This occurs for example whenever the loadlock is vented to atmospheric pressure at the end of the unloading sequence Vacuum Server Controller A software program knownas the vacuum serverruns in the system computer concurrently with the application program and other servers It maintains an awareness of the status of the vacuum system and it processes requests for vacuum actions by the application program This involves translating high level requests for example load a mask plate to the actual sequence of low level commands involved in carrying out the request for e
38. n the stage under the column Registration The positioning of successive masks on the stage must be repeatable in the x y plane so that given the Cartesian coordinates of defect sites found on a mask by an inspection system a predetermined transformation can be applied to obtain the FIB system stage coordinates that will bring the defect sites under the ion column It must also be vertically repeatable because FIB process parameters for mask repair presuppose a fixed relationship between the top surface of the mask the electron funnel and the ion column The repeatable positioning of masks in these respects involves the use of registration surfaces against which a mask is brought to bear The internal or chamber robot places the mask on the mask holder The mask holder uses small drive motors and registration pins to register the mask on the stage in the x y and z planes The Mask Holder The mask holder mechanically accepts a 6 inch square mask or reticle that is transferred from the internal robot The internal robot places the mask onto the mask holder to an accuracy within several thousandths of an inch then retracts to its home position The mask rests on three spherical registration buttons whose z axis height and coplanarity are aligned with a high degree of accuracy This ensures the surface of the mask is properly located in the reference plane The buttons are manufactured from a high performance thermoplastic called polyetherether
39. nents so that changes can be made without recompiling the entire application and system level code To ensure reusability the software source code is written using C and is COM compliant A brief explanation of C and COM follows C C is an object oriented programming OOP language that is organized around objects rather than actions and data rather than logic Historically a program has been viewed as a logical procedure that takes input data processes it and produces output data The programming challenge was seen as how to write the logic not how to define the data Object oriented programming takes the view that the objects that are manipulated are just as important as the logic required to manipulate them Examples of objects range from human beings described by name address etc to buildings and floors whose properties can be described and managed down to the little widgets on the computer desktop such as buttons and scroll bars The first step in OOP is to identify all the objects to be manipulated and how they relate to each other an exercise often known as data modeling After an object has been identified it can be generalized it as aclass of objects and defined by the kind of data it contains and any logic sequences that can manipulate it Each distinct logic sequence is known as a method A real instance of a class is called an object or in some environments an instance of a class The
40. nts to communicate with each other COM is a binary and network standard that allows any two components to communicate regardless of what machine they re running on as long as the machines are connected what operating systems the machines are running as long as it supports COM and what language the components are written in A component is a building block program that is self describing This means that it can be run with a mix of other components and each will be able to understand the capabilities and characteristics of the other components Practically this means that a new application can be built by reusing components already known to exist and without having to compile the application It also makes it relatively easy to distribute different components of an application among different computers in anetwork Microsoft s Distributed Component Object Model DCOM adds interfaces to do this In addition to its self description a component consists of one or more classes that describe objects and the methods or actions that can be performed on an object A class or co class in COM terminology has properties described in an interface or co interface The class and its interface are language neutral Associated with the class are one or more methods and fields that are implemented in a specific language such as C or Java or in a visual programming environment When a class is instantiated an object something real that can be
41. nular anode at the opposite end of the gun e Grid This is a wire mesh electrode interposed between the filament and the anode A negative potential can be switched onto the grid to counteract the acceleration potential between the filament and the anode and the gun can thereby be blanked shut off while preserving a state of filament readiness The electrons pass through the anode and into the snout where they are directed such that when they exit the snout and enter the electrostatic field produced by the various potentials on the particle detector they assume trajectories that takes them to the appropriate area on the specimen 800 000071 Rev A 47 Principles of Operation detector screen annular heated cathode filament filament supply deflectors Figure 14 Schematic Diagram of Flood Gun The Funnel The funnel is a shallow pan shaped metal structure that is suspended below the ion beam column primarily for the purpose of providing an electrostatic shield between the electron flood and whatever static electric charges may be present on the mask An orifice in the center of the pan allows the ion beam and the electron flood to pass through to the mask while the funnel itself provides a relatively large grounded plane A vertical motion mechanism is provided that extends lowers the funnel to a position close to the mask to maximize its effectiveness during charge neutralization and that retracts raises it sligh
42. of spot size 20 to 30 nm and beam current 10 to 20 pA that is typically preferred for mask repair An aperture eventually deteriorates with use the orifice becomes enlarged by the ion beam hence the need for redundancy While the ion column remains powered up the currently selected beam limiting aperture remains in the path of the beam and is eroded by it To extend the useful life of the aperture set a sacrificial aperture should be selected parked in the beam path if the ion column is to remain powered up but idle for any appreciable time The sacrificial aperture would be one of the little used apertures e g 60 um or any aperture that has already deteriorated beyond use 800 000071 Rev A 40 Principles of Operation The Lens 1 focus voltage that produces the highest current density irrespective of spot size for each of the beam limiting apertures is one of several items of data specified in a data structure called a lens table and is asserted by the software when it positions a particular aperture for use The amount of stigmation force that is required also tends to vary with the beam s diameter and current density which means that selecting a beam limiting aperture of a different size usually requires re stigmating the beam Therefore the software also saves stigmation values in association with each of the beam limiting apertures in the lens table See Stigmation below When it is desirable to alter the
43. ple 800 000071 Rev A 39 Principles of Operation Same SS LI lt a Limiting B _ SX SS Aperture a Lens 1 focused for high b Lens 1 focused for lower current density current density Figure 11 The Effect of Lens 1 on Current Density Beam Limiting Apertures A beam limiting aperture is any of 14 apertures that are arranged in two parallel lines on a horizontal sliding bar and that can be brought into position by a mechanical drive mechanism Homing the positioning mechanism withdraws the sliding bar completely so that the beam is unrestricted this is normally done only when servicing the column A beam limiting aperture reduces the beam diameter at the point of interception and has a proportional effect on the spot size the smallest spot to which Lens 2 can focus the beam on the specimen The combination of the size of the beam limiting aperture and the current density established by the focusing of Lens 1 is the major determinant of beam current delivered by the column and to a lesser degree of spot size the spot size increases somewhat with the current density The standard set of beam limiting apertures installed on Accura 850 systems consists of six 25 um apertures two 30 um apertures four 40 um apertures and two 60 um apertures The 25 um aperture size is expected to be used most often because it yields a combination
44. proximately the same time the bit streams from the transmitting stations will interfere collide with each other and both transmissions will be unreadable If that happens each transmitting station must be capable of detecting that a collision has occurred before it has finished sending its frame Each must stop transmitting as soon as it has detected the collision and then must wait a quasi random length of time determined by a back off algorithm before attempting to retransmit the data frame USB The Joystick which can be used for manual stage movement communicates with the system via the USB Universal Serial Bus port on the Windows NT workstation 800 000071 Rev A 36 Principles of Operation The lon Column The VisION ion column generates the ion beam that does the work for which the FIB system is intended All of the other system hardware and software serves either to create a suitable environment for the generation of the beam or to provide a more or less convenient means by which it can be put to use Column performance is generally given in terms of beam current and spot size For the mask repair application the column is operated such that it produces ion beam currents in the range of 5 to 100 picoamperes and spot sizes from 10 to 40 nanometers Figure 10 is a schematic diagram of the ion column Although simplistically depicted in the schematic as two dimensional shapes most of the elements through which the ion
45. r mounted near the bottom of the ion column attracts and captures the secondary particles emitted by the mask in response to excitation by the ion beam and it develops an electrical signal whose magnitude is proportional to the number of particles it is receiving at any given instant Variations in surface topography and composition that are encountered as the ion beam scans the object being imaged produce variations in the rate of secondary particle emission which result in variations in the electron current the signal produced by the detector The detector that is normally used is often referred to as the MCP pronounced like the letters which is the abbreviation for the kind of device that it is a microchannel plate The signal thus detected at successive dwell points is applied to the signal acquisition electronics where it is filtered for noise suppression and digitized to a binary number in the range of 0 to 255 where 0 is the weakest darkest signal and 255 is the strongest brightest signal As the scan proceeds the digital values produced by the signal acquisition electronics are stored in successive locations in a random access memory array called the frame buffer in the graphics board in the FIB system computer The frame buffer is read sequentially and continuously in synchronism with the raster scan of the CRT monitor such that the value of each byte of image data determines the combination of red green and blue RGB signals
46. rocess It can also deflect the ion beam thus affecting the placement accuracy of repairs In ion imaging mode the detector is biased to attract positive ions and a flood gun see below which directs a stream of free electrons at the specimen can therefore be employed for positive charge neutralization during rasters The flood gun cannot be used while the detector is biased to capture electrons because the electron flood would saturate the detector reducing its useful life and in any case totally obscuring electron emission from the specimen Multiplex imaging mode is a variation of electron imaging mode in which the detector is switched off and the flood gun is switched on during the time that the ion beam is moving between dwell points While the beam is dwelling at any particular point the flood gun is switched off and the detector is switched on This has the advantages of both electron and ion modes clearer images with concurrent charge neutralization However it requires significantly more time to produce an image than the other modes and to be effective it requires relatively frequent recalibration of the flood gun to verify that it is correctly aimed and energized Imaging Speed The rate at which an image raster proceeds is generally referred to as the imaging speed The dwell time the time during which the ion beam is directed at each dwell point is the main determinant of the imaging speed for ion mode or electron mode For
47. ry during deposition to allow the deposition gas to be replenished on the surface of the mask between frames lon Dose The ion dose is the quantity of ions delivered to the mask per unit area which translates to the depth of milling or the thickness of deposition 800 000071 Rev A 44 Principles of Operation The dose delivered for any particular operation is a function of the beam current the dwell point spacing the dwell time and the number of frames rasters Dose does not take acceleration potential which determines the velocity or energy level of the ions into account so the effect on the mask of a given dose is likely to be less for example with the acceleration potential at 25 kV than at 30 kV The units of dose are either nanocoulombs of electrical charge per square micron nC um2 or ions per square centimeter ions cm2 One nC um is equivalent to about 6 25 x 1017 ions per square centimeter FIB Imaging The FIB system produces gray scale images by applying the same principles used by scanning electron microscopes SEMs That is the mask is probed by the ion beam in a raster scanning pattern and at each dwell point the energy transferred to the mask from the decelerated ions excites it and thereby causes it to emit electrons positive ions and negative ions This phenomenon is referred to as secondary emission and the particles emitted are collectively referred to as secondary particles A device called a detecto
48. s pass are either hollow cylinders or disks with central orifices The element stack is aligned so that the geometric centers of the elements are on a vertical axis called the central axis Ideally the stream of ions passing through the column assumes the shape of a long narrow cylinder whose longitudinal center corresponds to that central axis The electrically active elements that make up the column are operated electrostatically by the application of relatively constant voltage as opposed to electromagnetically by the passage of current through them Note the presence of the gate valve VGD that can be closed by the software to isolate the upper part of the column including the ion source and the limiting aperture assembly from the lower part of the column for servicing see Jon Gun Isolation Valve on page 58 for related information lon Emission An ion is an atom that has acquired a positive or negative electrical charge by virtue of losing or gaining electrons an effect known as ionization The FIB system uses liquid metal gallium as the ion source In the ionization process gallium atoms usually lose one electron thus becoming singly charged positive ions Being charged particles ions can be accelerated and directed along particular paths by electrostatic fields and their relatively high mass compared with that of subatomic particles allows them to be used to produce the milling and deposition effects Located at the top of the
49. selects a plate size 2 5 6 or 7 inches and a plate file then clicks the Load button at the bottom of the dialog box 800 000071 Rev A 50 Principles of Operation x axis amp z axis mask registration surfaces axis mask registration z axis registration clamp surface release pressure points x y motor z motor actuator cam registration pins etc Figure 16 The Mask Holder 800 000071 Rev A 51 Principles of Operation Figure 17 Mask Holder with Mask in Place See Figure 18 Mask Carrier RSP An automated loading system involving two transfer robots and automated alignment mechanisms transfers the mask to the positioning stage When imaging and repair procedures are complete the mask is automatically returned to its mask carrier in the load port module Diagnostic Fixtures Refer to Figure 19 Several devices that are used for alignment and calibration are mounted near the rear edge of the stage These include the silicon grid the ion faraday and the electron faraday The stage position that places each of these devices under the ion column is recorded on disk as a stage label file so that the software can bring them into play whenever necessary A solid state sensor for gauging workchamber temperature is also mounted on the worktable The other devices are explained below The Silicon Grid The silicon grid is a flat silicon chip about 10 mm by 16 mm on which is etched a grid of 5 micron lines on 35 m
50. sent the data This concept makes it easier to service or upgrade the network since data transmission is not based on the availability of specific types of stations Ethernet The remaining electrical subsystems of the XT100 communicate via an Ethernet based LAN local area network The XT100 uses a 10Base T Ethernet which provides transmission speeds up to 10 Mbps Megabytes per second The Ethernet LAN bus is configured in a star topology with each station or node connected to a central multi port repeater also known as a hub All connections in the network are point to point links implemented with unshielded twisted pair UTP cables Each Ethernet station competes for bus access using the Carrier Sense Multiple Access with Collision Detection CSMA CD protocol The CSMA CD protocol was originally developed as a means by which two or more stations could share a common media cabling type Each Ethernet station determines for itself when it will be allowed to send a data frame The CSMA CD access rules are summarized by the protocol s acronym e Carrier Sense Each station continuously listens for traffic on the medium to determine when gaps between data frame transmissions occur e Multiple Access Stations can begin transmitting any time they detect that the network is quiet there is no traffic e Collision Detect If two or more stations in the same CSMA CD network collision domain begin transmitting at ap
51. series of short periods during the dwell averaging the signal levels thus obtained and digitizing that average 800 000071 Rev A 46 Principles of Operation as the information for the dwell point The principle is that when the signal produced by the particle detector as a result of the impingement of the ion beam at a given dwell point is sampled repeatedly and is averaged over the number of samples the randomly occurring information noise will eventually constitute a relatively small part of the final result while the valid information will tend to prevail The duty cycle for the sampling electronics is chosen by the software based upon the dwell time The Flood Gun The flood gun is an electron emitting device that is suspended horizontally near the lower end of the ion column to direct a beam of low energy electrons at the mask in an area roughly corresponding to the field of view This electron flood neutralizes the positive electrostatic charge that tends to build up on the mask during rastering Electrostatic charge buildup interferes with imaging and may result in electrostatic discharge that damages the specimen See Imaging Modes on page 45 for related information The application program s Power Up function powers up the flood gun as well as the ion column While it is powered up the flood gun operates continuously as long as the selected imaging mode is ion It is operated intermittently during imaging rasters whi
52. ted in the opposite direction to the anode Thus an apparent DC flow through the pump occurs as the free electrons arrive at the anode and the ions arriving at the cathode are neutralized by cathode electrons This current is 5 The surface area is large enough so that at operating pressures of 1 3 x 10 4 Pa 10 torr or less the life expectancy of an ion pump may be well in excess of 50 000 hours 800 000071 Rev A 60 Principles of Operation to site exhaust extraction i foreline roughing toreine isolator pump approx 610 mm separation Du AA plastic with std length s s vacuum hose vacuum ose T connection chemical cabinets FIB system turbopump enclosure O roughing manifold loadlock Figure 22 Auxiliary Vacuum Pump Configuration Top View proportional to the number of gas molecules being ionized is therefore proportional to the gas pressure in the pump FIB Chemistry Any of several process chemicals may be used in conjunction with the ion beam in the course of repairing a mask The source for each of the chemicals is stored in one of four chemical cabinets mounted in a vertical stack inside the rear of the process module enclosure Process chemicals are delivered under control of the application software by a system of delivery lines and control valves to nozzles mounted on the Funnel see page 48 from which they are directed at the mask in the vicinity of the ion beam scan field 80
53. terposed between the forepump and the foreline This consists of ballast mass for suppressing the transmission of forepump vibration to the foreline an isolation valve which the application software can close to protect the turbopump against a sudden gas load if the gas pressure in the foreline should start rising for any reason and a pressure sensor which allows the application software to monitor the pressure in the foreline The effective pumping capacity pumping speed of the auxiliary pumps is limited by the conductance of the vacuum hoses which increases with internal hose diameter and decreases with length Therefore to obtain adequate pumping speed the vacuum hoses are made as short as is feasible lon Pumps To help maintain high vacuum in the ion column without introducing mechanical vibrations an ion pump is attached to the lower column housing and a second ion pump is attached to the upper column housing to maintain very high vacuum in the ion source housing Each has a separate roughing duct and roughing valve but the same controlling signal operates both valves simultaneously they are collectively known as VI new column pumping information tk An ion pump works by ionizing gas molecules with an intense electrical field By virtue of their charge the ions are accelerated to and become implanted in a cathode structure having a very large surface area gt The electrons freed by the ionization process are accelera
54. the Windows NT registry The servers accept requests from the application program on behalf of the system user and operate the hardware by issuing the appropriate commands to the device drivers Each server is responsible for a particular subsystem For example the vacuum server communicates with the vacuum controller the hardware unit that operates the loading unloading mechanisms and all of the pumps valves and gauges that are involved in maintaining high vacuum in the workchamber When the FIB system user commands the application program to transport a mask holder from the loadport module through the air lock chamber known as the oadlock and onto the moveable stage in the workchamber an event referred to as loading the program sends a load request to the vacuum server which sends the appropriate sequence of commands to the vacuum controller to evacuate the loadlock to operate the transport mechanism and to otherwise carry out the loading event Other servers include the positioning server the HV high voltage server which handles requests for setting the outputs of the high voltage power supplies and otherwise governs the operation of the ion column and the MUI server which monitors the positions of the knobs and switches on the manual user interface also referred to as the knob panel Refer to page 78 for more information about the servers Control and Communications The 100nm XT tool uses a distributed controls architecture that
55. the ions increases the sputtering effect as suggested by the illustration or the average thickness is less the practical implication is that the ion dose that will remove the interior of a chrome defect without damaging the glass substrate will tend to over etch and produce a riverbed at the perimeter Figure 24 Edges Etch Faster than Flat Surfaces 800 000071 Rev A 63 Principles of Operation Various scanning arrangements have been devised to minimize the riverbed effect but the best results are obtained through the dose reduction allowed by the use of the chrome GAE process described earlier Gallium Removal The implantation of gallium ions in the quartz substrate in the course of chrome removal is problematic because the presence of the gallium reduces the transmission of short wave length light by the quartz and it therefore compromises the effectiveness of the repair process However because xenon difluoride XeF gt vapor is highly reactive with quartz it can be used in conjunction with a relatively light ion dose to remove the implanted layer of quartz without resulting in further implantation The xenon difluoride used for gallium removal is stored in its chemical cabinet at room temperature 800 000071 Rev A 64
56. tion counters is available from linear encoders that are incorporated in the stage piezoelectric drive motors As a drive motor moves its encoder produces digital pulses that indicate displacement from which stage motion can be extrapolated Vacuum The ion beam can be generated only in an extremely low pressure environment which is to say an enclosed space as nearly devoid of gas molecules as is feasible This is because the intense electrostatic fields in the column will ionize any gas molecules that might be present Ionized particles conduct electricity if the particle density is such that electrons can jump from particle to particle Thus if the column is powered up with enough gas present electrical arcs will occur between electrodes electrode vaporization and ion implantation will occur and the column will thereby be damaged or destroyed Ionized particles are also accelerated to electrodes having an opposing charge where they accumulate eventually affecting the operation of the column For these reasons as well as for the safety of service personnel the FIB system is interlocked such that an abnormal rise in internal pressure switches off the high voltage to the column The need to keep the interior of the ion column evacuated when power is applied to it dictates that the workchamber must also be evacuated This is because the opening through which the beam passes at the bottom of the column will admit gas molecules and in any case the
57. tly to provide vertical clearance for maneuvering the stage in the x y plane The funnel also serves as a mount for the carbon deposition gas delivery nozzle as well as for the delivery nozzles for any other chemicals that may be installed in the system The Stage The ion column is stationary and its field of view scan field is relatively small Therefore a moveable stage is incorporated to allow the mask to travel about in the x y plane to bring points of interest under the column The stage is implemented with piezoelectric linear drive mechanisms as depicted by Figure 15 Working travel is 4 inches 100 mm from center which accommodates masks up to 7 inches 200 mm square and allows diagnostic fixtures see below on the stage to be positioned under the column 800 000071 Rev A 48 Principles of Operation ion column stage zero 7 position y travel lt x travel gt load unload path loadlock Figure 15 X Y Stage Motion Positioning Server A software program known as the positioning server operates the stage at the hardware server level It maintains an awareness of the position of the stage and processes requests for stage motion by application programs Stage Motion Controls The FIB system user can move the stage with a joystick or with a variety of controls provided by the application software The joystick operates the stage directly through the computer s USB port using DirectX
58. umping of the upper column area See Figure 10 on page 38 for the location of GIVi in the ion column Gas Pressure Units The relative vacuum in the FIB system is expressed in terms of gas pressure The international SI unit of measurement for pressure is the pascal Pa The customary unit of measurement for gas pressure in the United States in the context of vacuum technology is the torr One pascal is equivalent to 7 5 x 10 3 torr 1 torr is equivalent to 1 3 x 102 pascal Gas pressure in the workchamber is typically maintained in the range of 10 5 Pa 10 7 torr which is roughly one billionth of atmospheric pressure Pressure Gauges Mechanical pressure gauges do not work at the low pressures normally sought in the FIB system The number of gas molecules that would find their way into such a gauge would be so small that the mechanism would not respond Therefore operating pressure in the workchamber is measured by a combination of thermocouple gauges and either a hot cathode ionization gauge HCIG or a cold cathode ionization gauge CCIG These work as follows Thermocouple TC Gauges are used at several locations in the FIB system to indicate gas pressure from atmospheric pressure down to the rough vacuum range which is about 1 3 Pa 10 2 torr They 800 000071 Rev A 58 Principles of Operation apply two principles that a junction of dissimilar metals develops a voltage proportional to the temperature of the junction and that
59. when the junction is heated and placed in a gaseous environment the heat escapes from the junction by convection at a rate proportional to the ambient gas pressure Thus the lower the gas pressure the hotter the junction remains and the higher the voltage it develops The voltage is measured by an electrical metering circuit and a predetermined correlation between voltage and gas pressure is used to derive the reading At pressures below 1 3 Pa 10 2 torr the gauge bottoms out produces its lowest reading because the convection of heat from the device becomes negligible lonization Gauges are used to obtain pressure readings in the workchamber when the pressure therein is below the range of the TC gauge Two types of ionization gauges have been installed in FEI 8000 Series systems hot cathode ionization gauges HCIGs or cold cathode ionization gauges CCIGs The HCIG consists of a glass envelope attached to a port at the rear of the workchamber in which several electrodes are mounted and associated electronics located in the FIB system s electronics module The gauge works by heating a cathode element to produce thermionic emission of electrons The electrons are accelerated to a wire grid having a positive potential and as they pass through the gas present in the glass envelope to reach it they ionize the gas molecules by colliding with them and displacing their outer orbit electrons The displaced electrons ionize more gas molecules
60. xample open valve V1 open valve V4 etc The vacuum server sends the low level commands through a communications link to a continuously running firmware program that is stored in read only memory in a hardware device called the vacuum controller The program in the vacuum controller includes firmware interlocks whereby it is coded to reject commands that in context would be unsafe or damaging Thus for example if a malfunction resulted in the vacuum server issuing a command to close valve CLV the loadlock workchamber door while the transport mechanism was still in the process of loading a mask plate onto the stage the vacuum controller program would reject the command The vacuum controller s firmware program is also capable of automatic actions under certain conditions For example the program will open the high voltage interlock circuit to power down the ion column if the pressure in the workchamber rises to a level above which arcing can be expected lon Gun Isolation Valve Designated G Vi this valve can be closed by the vacuum server to maintain vacuum in the lower part of the ion column when the upper part of the column must be vented to atmospheric pressure to allow service access to the ion source and or the beam limiting apertures The service apparatus that vents and pumps the upper column area is enabled only when GIVi is closed The service apparatus includes a controller that allows one button venting and one button p
61. y high ion dose otherwise required for chrome removal tends to result in riverbeds see below and gallium ion implantation both of which reduce the transmission of short wave length light through the quartz and thereby compromise the quality of repairs The reservoir for the bromine used with the chrome GAE process is refrigerated to 12 C 10 4 F by a thermostatic control system that includes a current regulated power supply mounted in the Electronics Module and a water cooled heat sink with a remotely located recirculating refrigeration unit The water vapor used with the chrome GAE process is stored in its chemical cabinet at room temperature Riverbeds A riverbed so called because in a SEM image it seems to resemble a dry riverbed is the outline of an opaque defect that has become etched into the glass substrate in the course of removing the chrome see Figure 23 800 000071 Rev A 62 Principles of Operation chrome chrome a the defect unwanted chrome b the riverbed over etch Figure 23 A Riverbed Riverbedding also called the riverbed effect tends to occur because the sputtering rate is usually greater at the perimeter of a chrome defect than it is in the interior a phenomenon that seems to be a consequence of the edges or sides of chrome features defects included typically sloping outward from top to bottom see Figure 24 Whether the low relative angle of incidence of

User Manual - David Hoyle

Contents

Download Pdf Manuals

Related Search

Related Contents