JK ze / ARCHIVES. Analysis of 3D Silicon Pixel Vertex Detector Damage Effects due to Radiation Levels Present in the LHC at CERN

Transcription

1 ARCHIVES MASSACHUSETTS INSTflUTE OF TECHNOLOGY JK Analysis of 3D Silicon Pixel Vertex Detector Damage Effects due to Radiation Levels Present in the LHC at CERN By Matthew R Chapa SUBMITTED TO THE DEPARTMENT OF NUCLEAR SCIENCE AN ENGINEERING IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF BACHELOR OF SCIENCE IN NUCLEAR SCIENCE AND ENGINEERING AT THE MASSACHUSETTS INSTITUTE OF TECHNOLOGY JUNE 2012 Matthew R Chapa. All Rights Reserved. The author hereby grants to MIT permission to reproduce and to distribute publicly Paper and electronic copies of this thesis document in whole or in part. ze / Signature of Author: 7 Matthew R Chapa Department of Nuclear Science and Engineering May 11, 2012 Certified by:_ Accepted by: Dennis Whyte Professor of Nuclear Science and Engineering Thesis Supervisor Dennis Whyte Professor of Nuclear Science and Engineering Chairman, NSE Committee for Undergraduate Students 1

2 Analysis of 3D Silicon Pixel Vertex Detector Damage Effects due to Radiation Levels Present in the LHC at CERN By Matthew R Chapa Submitted to the Department of Nuclear Science and Engineering on May 11, 2012 In Partial Fulfillment of the Requirements for the Degree of Bachelor of Science in Nuclear Science and Engineering ABSTRACT In high energy physics experiments, very high precision tracking of charged particles is needed. Solid state detectors achieve the high precision necessary to provide track and vertex reconstruction of the particles that traverse them, but tracking performance begins to deteriorate at fluxes of radiation around hadrons/cm 2. These radiation levels are congruent with those experienced by the ATLAS pixel detector, the inner most part of the ATLAS tracking system, which is vital to track and vertex reconstruction. During the planned shut-down of the Large Hadron Collider (LHC) in , the energy and luminosity of the LHC will both be increased. The current pixel detector has begun to suffer deterioration of performance, so the ATLAS Collaboration has initiated an upgrade to take place during the scheduled shut-down beginning in 2013, the Insertable B-Layer (IBL). The IBL will be assembled and placed in between a reduced diameter beam pipe and the current pixel detector, acting as the fourth layer of the ATLAS inner detector. The pixel sensors of the IBL will have to sustain a radiation dose of 5 * 10" n,/cm 2. Two sensor technologies are being considered for the IBL upgrade: planar n-in-p silicon pixel sensors and 3D double sided n-in-p pixel sensors. Research of both these technologies is being done by the IFAE in collaboration with CNM-Barcelona. To cope with the increased data rate after the LHC upgrade, a new front-end chip has also been produced, the FE-14 front-end chip. Test results and data analysis from five different 3D pixel sensor devices, all fabricated at CNM-Barcelona were done. Evaluation of these technologies and the test results of irradiated 3D pixel sensor devices are carried out in this thesis. Thesis Supervisor: Title: Dennis Whyte Professor of Nuclear Science and Engineering 2

3 Table of Contents Page T itle P age... 1 A bstract Table of C ontents... 3 L ist o f F igures...4 L ist of T ables Introduction The LHC and ATLAS Detector ATLAS Inner Detector TRT & SCT Pixel Detector D and Planar Pixel Sensors Semi-Conductor Detectors Semi-Conductor Properties Doped Semi-Conductors N oise Two Pixel Detector Technologies P lanar D ATLAS Upgrade Insertable B-Layer (IBL) D Sensor Technology Front-End Chip FE M ethods D Pixel Sensor Tests Current vs. Voltage (I-V) Threshold and Noise Noise vs. Voltage Time over Threshold (ToT) Charge Collection with Strontium-90 (Sr-90) Beta Source Test B eam R esults D Pixel Sensors Un-irradiated vs. Irradiated Su m m ary C onclusion Acknowledgements R eferences

4 List of Figures Figure 1. Schematic of LHC at CERN...9 Page Figure 2. Schematic of ATLAS Detector...10 Figure 3. Cross-section of ATLAS inner detector Figure 4. Propagation of various particles in ATLAS detector Figure 5. Spacing of radial structure of inner detector...13 Figure 6. Schematic of current ATLAS pixel detector...14 Figure 7. Silicon material at 0 K and 300 K...17 Figure 8. A pn-junction and resulting depletion zone Figure 9. Schematic of n-in-p planar pixel sensor...23 Figure 10. Layout of 3D double-sided wafer sensor...23 Figure 11. Schematic of IBL insertion into ATLAS...25 Figure 12. Photographs of an FE-14 USBpix system Figure 13. Hits vs. Charge scan graphed with the ideal and real curves...30 Figure 14. Set-up of post-irradiation scintillator experiment with Sr-90 source...32 Figure 15. DUTs and telescopes test beam set-up at CERN...33 Figure 16. DOBOX sections which house the DUTs and dry ice Figure 17. I-V scan results for 4 irradiated devices from Table Figure 18. Noise vs. Threshold graph for device Figure 19. Noise vs. Veff, # noisy pixels vs. Ve, and pixel map for Veff Figure 20. ToT MPV vs. Veff plot for 3 p-irradiated devices at different doses...39 Figure 21. Sr-90 charge collection of device 36 for 3 different Veff...40 Figure 22. Sr-90 and occupancy scan results for device 34 and

5 List of Tables Table 1. Classification of materials based on their energy band gap...16 Table 2: List of 3D devices studied

6 1. Introduction Currently, there is no unifying theory of physics combining the four known forces of physics acting on fundamental particles: strong force, weak force, electromagnetic force, and gravity.[ 1,2] The Standard Model of Particle Physics presents the leading theory unifying all fundamental forces except gravity, but even the Standard Model leaves questions such as the origin of mass and why the weak force is still 1032 times stronger than gravity unexplained. The recently confirmed mass of the neutrino does not fit into the Standard Model, which levels out the playing field with the other two leading unifying theories of physics: supersymmetry (SUSY) and string theory.[ 1,2] In order to investigate these questions, conditions comparable to the origin of the universe need to be recreated. To study these conditions, particles need to be collided together at very high energies. As a result of the need to achieve high energy particles colliding, the LHC at CERN in Geneva, Switzerland was built. The LHC is the largest and highest energy accelerator in the world. It is designed to accelerate bunches of protons to 7 TeV. The accelerator accelerates bunches of protons in opposite directions and are steered to collide at the four major detectors at CERN, amounting to a combined energy of 14 TeV.[1] Currently, the combined energy of the proton bunches is 8 TeV until the next upgrade is done during the planned shut-down of the LHC is done in One of the two general purpose detectors at CERN is the ATLAS detector. The ATLAS detector provides vertexing information about the particles that result from the high energy protonproton collisions as well as information about the particle momentum. There are many components to the ATLAS detector, each designed specifically to detect different types of particles. Of particular importance to this thesis is the pixel detector of the inner detector. The pixel detector is the first line of detection after the protons collide and produce other particles. Accordingly, it undergoes massive amounts of radiation as it is just centimeters away from the collision point. Following the harsh radiation environment of the pixel detector at the LHC, the material is damaged and degraded, leading to loss of performance. The ATLAS Collaboration has taken the opportunity of the 2013 planned shut-down to insert a fourth layer in between the pixel detector and beam pipe known as the Insertable B-Layer (IBL). [3] The IBL will be even closer to the interaction point than the previous pixel detector. As such, stringent requirements have been placed on the technologies being considered for the IBL: 6

7 3D silicon pixel detectors and silicon planar pixel detectors. Both of these technologies are solid state detectors which provide high precision tracking, and silicon sensors are well suited for the IBL due to their radiation damage resistance and high position resolution. The current pixel detector is already experiencing performance degradation after just a few years of use. The upgraded proton collision energy and hadron fluence of the LHC will require a new pixel detector technology to be used in the IBL.[4] This thesis focuses on the 3D silicon pixel detector technology in consideration for the IBL and how its performance is affected by the radiation levels it will experience in the ATLAS detector. An overview of the LHC at CERN and its mission is given in section 2. The ATLAS detector, of particular interest to this thesis, is described in detail along with its purpose and future. The layout of the inner detector in ATLAS, the pixel vertex detector, is given in order to understand the purpose and necessity of the IBL upgrade described in section 4. Section 3 gives an overview of semiconductor detectors and their general properties. A technical description of the origin of pixel noise is done. With this information, a better insight may be gained into the two competing technologies in consideration for the IBL upgrade described in section 4 and which will provide better performance when subject to the harsh radiation environments of the LHC. Section 4 presents the need for an upgrade to the inner detector of ATLAS and describes the 3D sensor technology studied in this thesis. The 3D pixel sensor chips are bump-bonded onto a front-end chip, the FE-14, which is also an upgraded version of the previous FE-13 front-end chip. The properties of the new FE-14 chip are also explored. The various methods for the different tests involving the 3D sensors are described in section 5. The test are ran on un-irradiated 3D sensors in order to characterize them before they are sent to the test beam sight at CERN which will simulate many years' worth of radiation damage as experienced in the LHC. The test beam setup is also described in section 5. The results of the various tests described in section 5 are given in section 6. The tests that were described in section 5 were done on the 3D sensors pre- and post-irradiation. The results and discussion are presented in section 6. The conclusion of the results and this thesis are given in section 7. 7

8 2. The LHC and ATLAS Detector The European Organization for Nuclear Research (CERN) was founded in 1954 along the Franco-Swiss border near Geneva, Switzerland. CERN was one of Europe's first multi-national collaborations and now contains twenty member states. The mission of CERN from the outset in 1954 was simple, "...provide for collaboration among European States in nuclear research of a pure scientific and fundamental character..." CERN also has no concern with work for militaries and therefore publicly publishes the results of experimental and theoretical work done at CERN. [1] The Large Hadron Collider (LHC) is the world's largest and most powerful accelerator and the latest addition to CERN's accelerator complex. The LHC was built from 1998 to 2008 in the hollowed out tunnel that previously housed the Large Electron-Positron collider (LEP) that operated from 1989 to The LCH was the multi-national collaboration of over 10,000 scientists and engineering from over 100 countries along with hundreds of universities and laboratories. The result is the current LHC ring, which has a circumference of 26.7 km and lies between 50 and 175 m underground.[1] A schematic of the LHC can be seen in Figure 1.[5] The goal of the LHC is to test predictions of various theories in fields of particle and high-energy physics. In particular, discovery of the existence of the Higgs boson, in accordance with the Standard Model, has been the root of many experiments and still remains one of the biggest missing pieces of any complete physics theory. In order to achieve this, the synchrotron is designed to collide beams of either protons or lead nuclei at energies up to 7 teraelectronvolts (TeV) per nucleon or 574 TeV per nucleus, respectively. [3] The beams of particles are accelerated and steered using superconducting magnets in opposite directions to collide at 4 points along the loop. Each of these 4 collision points is the place of a detector, each designed to gain insight into specific areas of physics. The ATLAS (A Toroidal LHC AparatuS) detector and CMS (Compact Muon Solenoid) detector are two complimentary detectors that serve as the LHC's two main general purpose detectors. Having ATLAS and CMS are essential for cross-confirmation of any new discoveries in high-energy physics. Both are used in the search and study for new physics such as the origin of the electro-weak symmetry breaking, Supersymmetry, quark compositeness, leptoquarks, W' and Z', and extra dimensions. They will also look for insight into the production processes of the Standard Model such as W, Z, and jet production.[6] The aim is to precisely measure the W and top masses, which are essential for consistency checks with Higgs studies. 8

9 LHC SPS Pb Figure 1: Schematic overview of the LHC and its four main detectors The LHCb (Large Hadron Collider beauty) is another one of the detectors. It specializes in B-physics experiments, measuring the parameters of oscillation period and lifetime differences between mass eigenstates of the B, mesons. It also studies the CP violation in the interactions of b- hadrons, heavy particles containing a bottom quark, and rare decays, which could lead to new physics.[1] The fourth main detector is ALICE (A Large Ion Collider Experiment), which is a heavy-ion program that studies the collision of Pb-Pb (lead) nuclei at energies of 2.76 TeV per nucleon. Post-collision temperatures and densities are expected to be enough to generate a quarkgluon plasma, a state where quarks and gluons are deconfined.[5] For the purpose of this paper, the ATLAS detector will be explored in detail, particularly pertaining to its inner detector, where the 3D pixel sensors explored in section 3 will be mounted. The ATLAS detector has a diameter of 25 m and a width of 46 m. It weighs approximately 7000 tons and is located in an underground cavern approximately 100 m under the surface. There are stringent requirements of the ATLAS detector. A schematic overview of the ATLAS detector is seen in Figure 2.[5,6] The Insertable B-Layer (IBL) explained in section 4 will be inserted between the current pixel detector, the innermost part of the ATLAS detector, and the beam pipe. The LHC is designed to collide protons up to 14 TeV, which allows for the search of new particles up to 5 TeV at a design luminosity of cm 2 s'. At full intensity, each proton beam will consist of 2808 bunches, and each bunch will hold about 1.15 * 10" protons.[6] If the bunches of protons are considered at energies of 7 TeV each, the LHC stores beams of energy of about 360 MJ each. 9

10 Muon Detectors Tile Calorimeter Liquid Argon Calorimeter h.j/ _ Toroid Magnets Solenoid Magnet SCT Tracker Pixel Detector TRT Tracker Figure 2: Schematic overview of the ATLAS detector and its various parts After a proton-proton collision at the center of the ATLAS detector occurs, a large, and therefore short-lived, particle of interest could result and decay into new particles that travel through the detector. The particles that result from the decay carry with them information about the original collision event and parent particle which produced them. In order for the ATLAS detector to precisely measure the particle momentum and energy, the particle must be stopped within the detector, but there are multiple types of particles that could result from a collision event. Accordingly, there are four main sections of the ATLAS detector, each designed to detect and obtain information from the various types of particles that could result from a collision event.[1,3,6] The outer most detector is the muon detector, which identifies and measures the momenta of muons. It employs a muon detection system consisting of precision drift tubes, which measure the curves of tracks, resistive plate chambers, for triggering and 2" coordinate measurement in the central region, cathode strip chambers, which measure precision coordinates at ends of the detector, and a thin gap chamber, for triggering and 2d coordinate measurement at the ends of the detector, all inside a toroidal magnetic field of up to 4 T. [5] The next two layers of the ATLAS detector, moving inward, are the calorimeter detectors. There is the hadron calorimeter, a liquid argon (Ar) calorimeter, which measures the total energy of hadrons and operates at -183 degrees C. The second of the two layers is an electromagnetic 10

11 calorimeter, a device that measures the total energy of "electromagnetic showers" produced by electrons, positrons, and photons.[5] The inner most detector is known as the pixel detector, designed to measure the momentum of each charged particle. It is made from highly segmented silicon strips and contains three main parts: the Transition Radiation Tracker (TRT), the Semi-Conductor Tracker (SCT), and the pixel detector. For the purpose of this thesis, the inner detector will be explored in detail in section 2.1. The pixel detector houses pixel sensors explored further in section 3. A figure of the inner detector is seen in Figure 3.[5,6] Forward SCT Barrel S TR T Figure 3: A cross-section of the ATLAS inner detector Each particle resulting from a collision event contains its own signature that it can be identified by in the ATLAS detector. Neutral particles such as photons and neutrons are not detectable in the tracking chamber at the innermost part of the detector. Instead, photons are detected by the electromagnetic calorimeter, and evidence of neutrons is presented by the energy they deposit in the hadron calorimeter. Charged particles such as electrons and protons are detected in both the tracking chamber and the electromagnetic calorimeter. Muons can be detected in all components of ATLAS while neutrinos are not detected by any components due to their rare interaction with matter, but missing energies that match up with neutrino energies gives evidence of them.[3] The propagation of the various types of particles resulting from a collision event and their tracks through the ATLAS detector is seen in Figure 4. 11

12 Tracking Electromagnetic Had ron Muon photons muons n Innermost Layer......Outermost Layer Figure 4: The propagation of various particles due to collision events in the ATLAS detector 2.1 ATLAS Inner Detector The ATLAS inner detector was designed to give precise tracking of particles resulting from colliding proton bunches every 25 ns (nanoseconds). A solenoid is used to measure the momentum of charged particles, which makes the inner detector operate in a homogenous magnetic field of 2 T. The inner detector is 7 m long with an outer radius of 1.15 m. The outer tracking systems of the TRT, the microstrip detector of the SCT, and the pixel detector all contribute to the reconstruction of the particle paths through the ATLAS detector. Two types of tracking systems are combined for this task: a high-resolution tracking system at the closest point to the interaction and a continuous tracking system located at the outer radius of 115 cm.[5] As seen in Figure 3, the high-precision end-cap detectors are mounted in concentric cylinders around the beam axis. Figure 5 shows a more detailed layout structure of the inner detector and its spacing. 12

13 R =1082 mm TRT TRT R =514 mmn SCT R =443 mm R =371 mm R =299 mi {R = 122.5mm Pixels Pixels R = 88.5 mm ir = 50.5 mm R = 0mm Figure 5: Detailed spacing of radial structure as a part of the inner detector to ATLAS Transition Radiation Tracker (TRT) and Semi-Conducting Tracker (SCT) Beginning at the outer radius of the inner detector, the TRT takes up about half the volumetric area of the inner detector with 12 M 3. It is made up of 4mm diameter straw tubes with 30 pm gold-plated tungsten wire. There are 50,000 straw tubes in the barrel portion, each 144 cm long, and there are 250,000 straw tubes in the end caps, each 39 cm long.[5] The purpose of the TRT is to provide additional information about the particle that could not be seen in the pixel detector of SCT such as whether the particle was an electron or pion. This is done by filling the straw tubes up with Xenon gas that ionizes as charged particles pass through them and gives information about them. [6] The SCT constitutes the middle section of the inner detector. It contains 4 concentric barrels with radii of 30.0 cm, 37.3 cm, 44.7 cm, and 52.0 cm, make up an area of 60 m 2 over the 4 cylindrical barrel layers and 1 8 planar end caps. The layers are formed by p-in-n silicon microstrip detectors unlike the smaller silicon pixel sensors used in the pixel detector, making the coverage of a large area more practical.[5] The SCT is the one of the most important parts of the inner detector due to its basic tracking in the plane perpendicular to the beam since it measures particles over a much larger area than the pixel detector and contains more sampled points. [3,6] 13

14 2.1.2 Pixel Detector The pixel detector is the innermost part of the inner detector that detects charged particles. It encompasses 1.7 m 2 and contains 1744 modules in total, each with area 10 cm 2, and 46,080 readout channels per module. Each module contains 16 pixel chips, each of area 50 x 400 sm 2 and resolution 14 x 115 sm 2.[5] These chip specs are required to provide high granularity since high precision measurements of the momentum of charged particles close to the interaction point is essential to understanding the physics of the collision event.[3] It is subdivided into three barrel layers of radius 50.5 mm, 88.5 mm, mm and three disks on either side of the forward direction at lengths 49.5 cm, 58 cm, and 65 cm from the center of the detector. The total length of the detector is 1.4 m and contains 1456 modules on the barrels and 244 modules on the discs. These pixel specs can be seen in Figure 6.[5] 1442mmfl 430MMv~ \Barrel Layer 2 // 1 arre Layer I Barrel Layer 0 (b-layer) End-cap disk layers Figure 6: Depiction of the current pixel detector in ATLAS The pixel detector faces the highest amount of radiation of any component of the ATLAS detector since it is the closest part of the detector to the interaction point. Because of this, there are stringent requirements that the detector be radiation hard so that it can continue operating with little efficiency loss under years of significant exposure. The pixel detector needs to retain the speed of its read-out chips, so that all particle events can be recorded. Accordingly, silicon was chosen as the material for the pixel detector due to its radiation hardness and fine granularity, which gives excellent impact parameter resolution. The pixel detector specifics and its planned upgrade for 2013 are further explored in sections 3 and 4. 14

15 3. Planar and 3D Pixel Sensors Two different pixel sensors are being investigated for inclusion into the 2013 ATLAS IBL upgrade described in section 4. The two types of pixel sensors being tested are the planar and 3D pixel sensors. Due to the extreme radiation environments of the pixel detector in the ATLAS inner detector, stringent requirements have been placed on the pixel sensors that they be radiation hard. They are required to sustain radiation environments up to 250 Mrad without seriously compromising their efficiency or active area.[7,8] Before understanding how radiation levels present at the LHC will damage the silicon pixel sensors and how that damage will affect their performance, it is first important to understand how silicon semi-conducting detectors operate. 3.1 Semi-Conductor Detectors In order to reconstruct the particles of the collision event in high energy physics and identify the collision vertexes, track resolution is vital. The ATLAS tracking detector has a resolution of about 10 pm.[3,7,8] This precision is achieved by a detector close to the interaction point by having a high single hit detection resolution. There are some detectors that can meet these characteristics, but it must also be radiation hard, which is achieved by semiconductor detectors Semi-Conductor Properties The properties of semiconductor material are determined their energy band structure. The energy band structures determine the energy states that electrons are allowed to occupy and largely determine the electrical properties of the material, which is essential in electronics and detectors. The valence band is the highest band in which electrons are present at absolute zero temperature (0 K). The conduction band is band of energies above the valence band, which are sufficient to free an electron from binding with its individual atom and allow it to move freely within the lattice of the material.[9] Energy band gaps constitute the regions within the electronic band structure of the material in which there are no energy states available for electrons to occupy. Depending on the band structure of the material and the magnitude of its band gap, it can fall into three different categories: metals, semiconductors, insulators.[6,9] Metals are materials in which the valence band overlaps with the conduction bands, and electrons are allowed to move freely throughout the material. This is useful for sending electrical current through a material with little resistance. Insulators are materials in which the energy band gap is about 6 ev or above. This could be useful if no electrical conduction is desired. Of particular 15

16 interest to semiconductors are the semiconducting materials, classified as having a band gap of about I ev to about 6 ev.[6,9] This information can be seen in Table 1. Table 1: Classification of materials according to their band gaps Material Band Gap Metal 0 ev (valence and conduction bands overlap) Semiconductor ~ 1eV ev Insulator > 6 ev Semiconductor detectors provide excellent position resolution, high detection efficiency, while operating in high radiation environments. The tracking detectors in the ATLAS experiment are placed very close to the interaction point, around 5 cm currently. [5] The small distance from the interaction point implies a high radiation dose received by the semiconducting material. It is then important to consider many factors when determining which material to use for semiconducting detectors. One of the most important factors for use in the pixel detector in the ATLAS experiment is radiation hardness. Accordingly, many materials were explored to determine the best possible material to use in the high radiation environments of the LHC. Among them were pure materials and doped semiconductors. Doped semiconductors are semiconducting materials that were doped with other elements to improve the material properties such as material stability and providing energy levels where there previously were none. Some common semiconducting materials used in semiconducting detectors are: germanium, gallium arsenide, and silicon carbide.[9] Silicon is the material most used in semiconductor detectors; it is the most understood and does not have to be cooled to temperatures below 100 K such as germanium. At 0 K, the lowest energy state of a semiconductor, all the electrons in the valence band are covalently bonded between the lattice atoms. Silicon has four electrons in the valence band at absolute zero temperature, but even at room temperatures, there is enough thermal energy to excite an electron from the valence band to the conduction band. When this occurs, a hole will be created in place of the excited electron that is now free in the lattice. A neighboring electron may jump from its energy state to fill the hole, again creating a hole where that electron came from. Through this process the hole effectively moves through the lattice. Since the hole is an absence of an electron in a field of electron, it can be seen as a positively charged "particle." Therefore, the semiconductor current comes from electrons and holes.[6] Figure 7 shows silicon at 0 K and room temperature. 16

17 0 K (No electrons in conduction band.) 300 K Fermi.k" ev Figure 7: Silicon material at 0 K (left) and 300 K (right) and its effective electron/hole pairs It can be seen in Figure 7 that while at room temperatures, silicon can experience electron/hole pairs being created. At the same time, there are a number of them that eventually recombine, establishing an equilibrium between the two. The concentration of electrons (or holes), ni, is then given by ni = NcNy exp(- ) = 2 u' T exp (3.1) where Nc and Nv are the number of states of the conduction and valence bands, respectively, E. is the energy gap at 0 K, kb is the Boltzmann constant, and T is the temperature. The electrons are found to vary as T 3 / 2 by Fermi-Dirac statistics.[6] If equation 3.1 is used in conjunction with the values given in Table 3.1, it can be seen that the exponential factor is on the order of - 10 for insulators while only ~ 10-4 for semiconductors. This can be surprising since the difference in the energy gap is only a few ev. As such, equation 3.1 has large implications for the effect of the energy gap in the electrical properties of a material. Typically, the energy gaps of semiconductors are lower than the energy needed to create an electron-hole pair.[6] Doped Semi-Conductors Pure semiconducting material such as silicon used in a detector cannot detect ionizing particles since the free charge carriers present at room temperature (25 C) is almost four orders of magnitude higher than the amount produced by a minimum ionizing particle (mip). To see this, a 1 cm x 1 cm x 300 pm slap of pure silicon is considered. An electron-hole pair created by a minimum ionizing particle is de.p MeVCM2.[9] According to equation 3.1, the concentration dx f of free electrons, ni, is about 108. The concentration of electron-hole pairs produced by a mip, nh 17

18 mip, is about 3.2 * In order for a mip to produce any readable signal from the pure silicon slab, the free carriers have to be removed from the material. Pure semiconducting materials contain the same number of free electrons in the conduction band and holes in the valence band. The amount of each can be changed by adding impurities that integrate into the lattice which would effectively "steal" a free electron or hole from the conduction band. This forms a doped semiconductor which can contain extra electrons (n-doped) or extra holes (p-doped). It should be noted though that although the doped semiconductor may contain an extra electron or hole, it is still electrically neutral since the seemingly negative or positive charges in the lattice are compensated by the impurity. If a p-doped and an n-doped piece of silicon are joined together, a pn-junction is formed, and a depleted zone with no free charges is formed by diffusion. This can be seen in Figure 8. When this occurs, free electrons are joined with the free holes of the doped silicon lattices, and an electric field is created by the excess charges in both lattices. The charge density can be found from the concentration of donors, n-doped material, and acceptors, p- doped material end, 0 <x <Xn ( P (X) = I-eNA, --x < X < 0 (32 where ND and NA are the concentrations of donors and acceptors, and xn and x, are the extensions of the depleted zone into the n- and p-doped silicon materials.[6] If the material as a whole is neutral, then NAxp = NDxn (3.3) must also be true in accordance with charge conservation. According to Poisson's equation, the voltage, V, is given by d 2 V p(x) (3.4) dx 2 e where E is the dielectric constant. If equation 3.4 is integrated once, the result is E = - end (3.5) dx -x + ce, xp < X < 0 18

19 where CN and CP are constants that are determined by the points in x. and -x, where E =0, which gives rise to the full electric field equation dv (X -x), O < x < x" E-- - ena (3.6) dx 9 (x + x,), x, < x < 0 Equation 3.6 can be integrated one more time to obtain the electrical potential, -end x2 _Xx, 0<x<X V = e C 2 (3.7) ) '+ A e xx +C' -x, < x < 0 and the constants C and C' can be found by recognizing the potential is continuous across x = 0.[6] By setting V(x)= 0 and V(-x,) = Vo, o= -(N x+nax) (3.8) is obtained. If equation 3.8 and equation 3.3 are used in conjunction, it is possible to obtain xn and xp, 1 1 x- e= ' and x = (3.9) end 1+YD ena 1+N (NA)! \ ND) When one of the concentrations of either NA or ND is much higher than the other (such as NA >> ND), the depletion depth into the material of lower concentration takes the form of a Debye length and is given by d(xn + x, ~e (3.10) By inspection of equation 3.10, it can be seen that it is possible to increase the depth of the depletion zone by increasing Vo, bias voltage. Therefore, by increasing the bias voltage of a doped semiconductor, the free electrons or holes will be removed and allow for the detection of electrons generated by a charged particle. 19

20 depletion zone S oo 000 n P bias metal contacts Figure 8: Joining of n-doped and p-doped semiconductors to form a pn-junction and a resulting depletion zone Since charge electrons and holes are constantly being created, a volume current forms. The volume current is proportional the concentration of electrons and holds which leads to I= 1 Ad (3.11) 2r where A is the area of the junction, ni is the concentration of electrons or holes from equation 3.1, d is the thickness from equation 3.10, and T is the electron/hole life-time.[6] If equation 3.10 and equation 3.11 are combined, the volume current in a doped semiconductor sensor is then vol = e (3.12) Therefore, the volume current, also known as leakage current, increases proportionally with Vo" 2 if Vo < Vbias. The pn-junction as discussed and seen in Figure 8 acts as a plane capacitor since the depleted zone is an acting insulator and the un-depleted zones are conductors. The capacitance of a planar capacitor is A A C = e - oc e -r (3.13) d g where A is the area of the capacitor plates and d is the distance between them. [6] In the case of a doped semiconductor, A is the area of the pn-junction and d is the depletion depth. Equation 3.13 states that the capacitance of a semiconductor detector decreases as the bias voltage increases. 20

21 3.1.3 Noise The doped semiconducting material used in the pixel detector in ATLAS is obtained from small signals of the charged particles that traverse it. Due to the very small signals, they must be amplified, but this provides a problem, noise. There is a certain amount of noise that arises from three different sources: thermal noise, shot noise, and I/f noise. If the signal to noise ratio is small, then the noise will be amplified along with the signal, and large uncertainties in the data will result. Thermal noise is a function of temperature, but the other two sources of noise arise from the electronics. Various filters must be implemented in the electronics in order to reduce the amount of noise and achieve the highest possible signal to noise ratio.[10,11,12] The thermal noise arises from the temperature of the lattice. The only reasonable way to reduce the thermal noise is to operate at a lower temperature, which is a common strategy when using doped semiconductors. [61 Shot noise arises from the number fluctuation of quanta and is temperature and frequency independent. It is connected to the way photons spatially arrive at the pixel detector. Shot noise is the standard deviation (RMS) for the number of interactions per pixel when charge quanta are emitted over a barrier, such as energy gap of the silicon.[ 11] A way to reduce shot noise is to operating the detector in the dark such that external light does not ionize the semiconducting detector and cause noise. The third type of noise is 1/f noise, where f is the frequency of the electrical signal. The reason for this noise is neither well known nor understood, but it believed to be caused by the preamplifier system. 1/f noise has also been thought to be caused by interface traps at the silicon/silicon diode interface as a results of very high quality deposited oxides.[10] These three sources of noise combine together to form the equivalent noise charge (ENC), which is given by ENC = Qnoise = d1(cd + C,)Q + d 2 (Cd + Cf) 2 + d 3 IleakRC (3.14) where d 1, d 2, and d 3 are constants, R and C are the resistor and capacitor values of the preamplification circuitry, and Cd and Cf are the capacitance of the semiconductor and a controlled capacitor value in the pre-amplification circuitry, respectively. The first term of equation 3.14 is the thermal noise; the second term comes from the 1/f noise, and the third term arises from the shot noise. This is known as the ENC, which is the charge necessary to produce the same signal at the end of the amplification system. Equation 3.14 shows the noise of the semiconductor detector is 21

22 directly dependent on the leakage current of the semiconducting sensor and the bias voltage, which ultimately changes the value of Cd. Therefore, the ENC will raise as the leakage current rises as well as if the bias voltage is lowered.[6] 3.2 Two Pixel Sensor Technologies There are two main pixel sensor technologies: planar and 3D. The pixel sensors are at the inner most part of the ATLAS detector, subject to the harshest radiation environments from the LHC of any other ATLAS detector component. As such, the pixel sensors must be radiation hard, suffer little loss to efficiency and material properties as it is continually irradiated over its years of operation in the ATLAS detector. Other desirable sensor properties include little inactive area, usually most prominent at the edges, minimum silicon chip thickness, below 250 pm, so multiple scatterings do not result and cause uncertainties in the data, and continuing to offer excellent position resolution after years of operation.[3,4] At the current radiation levels experienced at the LHC, even the state-of-the-art silicon sensors fail to operate properly due to radiation damage to the crystalline structure which alters the silicon sensor properties. The planar and 3D technologies are being further developed and tested to try and negate these adverse effects of radiation damage Planar The planar sensors are a proven technology and what are currently being used as the ATLAS pixel detector sensors. They are an n*-on-n pixel sensor diffused on an oxygenated floatzone silicon bulk. The opposite sides of the electrodes are in contact with a p* layer. This requires that the current planar pixel sensors be fabricated on both sides of the substrate. The new planar pixel sensors that are being developed are n-on-p pixel sensors. There are two pixel sensors mounted on each modules of the pixel detector. The new substrate has a thickness of 200 pm, which is a sizeable reduction from the current 256.pm thickness in the pixel detector, and only requires single-sided wafer processing, which could lead to higher yields and reduction in fabrication costs. [13] The planar sensors require a bias voltage between V to achieve a high hit efficiency.[3] A schematic of the n-in-p planar pixel sensor is shown in Figure 9. Since the purpose of this thesis is to explore the characteristics of 3D pixel sensors, this is the extent to which the planar sensors will be discussed. 22

23 UBM 50 um 30 um 6um G 0 0 p pixels guard ring Figure 9: Cross-sectional schematic of the n-in-p planar pixel sensor D The 3D pixel sensors are produced on 230 pm thick silicon wafer with a double sided process. The 3D pixels wafers take advantage of new silicon technology advances that produce column-like electrodes which penetrate the substrate instead of being implanted on the surface like the planar technology. The wafer also exhibits n-in-p technology as columns of roughly 10 pm diameter are etched into opposite sides of the wafer alternating n- and p-type in a process known as deep reactive ion etching (DRIE).[4] A layout of the 3D double-sided pixel sensors can be seen in Figure 10, where the columns are the pixel sensors etched into the silicon substrate. BumP Si-P Poly-id P-stop N-diffusi 250pm holes P-diffusion 285pm Slicon Figure 10: Layout of 3D double-sided wafer sensor, where the columns are the silicon pixels 23

24 The 3D sensor is intrinsically radiation hard because it decouples the electrode distance from the bulk thickness and the small collection distance reduces the trapping probability, governed by the separation of the columns. Due to this fact, 3D sensors can reduce the charge collection path without having to suffer a reduction in sensor material the charged particles traverse. This makes it possible for the 3D sensors to continue to fully deplete even after hadron fluencies of 5 * 10"5 neq/cm 2 planned for the LHC upgrade in The 3D sensors only require about 200 V to obtain the same high hit detection efficiencies that planar sensors achieve at V.[3,4,6] This is important as it alleviates the stringent requirements of the cooling system for the pixel detector. Due to the advanced design of the 3D pixel sensors, there is a low dependence on the external magnetic field and little charge sharing between the pixels of the sensor. The pixel sensors are mounted on the modules using only 1 chip compared to the 2 chip modules of the planar pixel detector. 24

25 4. ATLAS Upgrade During the scheduled 18 month shut-down of the LHC in , the ATLAS detector will undergo some upgrades. During this time, the LHC luminosity will be upgraded as well to more than double the expected increase from the onset. A total of 29 proposals for upgrades to components in the ATLAS detector were reviewed, and 14 have been fully approved considering the expected increase of luminosity to 103" fb- 1. This correlates to an expected hadron fluence increase to 5 * 10" ne/cm 2 or 250 Mrad to the pixel detector, and the current pixel detector only provides acceptable charge collection efficiency up to 2 * 10 ne/cm 2.[4,14] Since the inner detector of ATLAS is not scheduled to be fully replaced until 2020, one of the approved projects for the upgrades to ATLAS is the Insertable B-Layer (IBL) addition to the inner detector. 4.1 Insertable B-Layer (IBL) In order to deal with the current and expected degradation of the current pixel detector, a fourth layer to the inner detector, the IBL, is being developed. It will be inserted in between the current pixel detector and the beam pipe. The space is too small to house the IBL, so the beam pipe will have a reduced radius in the ATLAS detector to appropriate room for the IBL. A schematic of where the IBL will be located in ATLAS is seen in Figure 11.[7,8] The purpose is to achieve the smallest possible radius for IBL in order to have better b-tagging performance, but due to the decrease radius, it will sustain a higher radiation dose due to increased interacting particle density. Figure 11: A schematic of where the IBL will be inserted in the inner detector of ATLAS There are stringent requirements for the IBL; it must remain in the detector and function without serious degradation until the 2020 upgrade when the entire inner detector will be replaced. A high track resolution is essential of the IBL in order to distinguish the tracks of the various 25

26 charged particles in the high track density environment. As such, radiation hard sensors are required to have a hit efficiency above 97% after irradiation to continue to collect charge in the pixel sensors. The sensors need to have very little inactive edges as well to achieve the highest possible active area ratio possible; this is vital since there is no long enough room to tilt the pixel sensors to an optimum angle, but even so, the IBL will increase the parameter impact resolution from the current pixel detectors. Along with these requirements comes a consciousness of the material budget. It is desired to have the smallest amount of material on two different fronts. The first reason is an obvious implied reduction in cost due to using less material. The second reason is to avoid multiple scatterings off of the detector material which causes uncertainties in the determination of the track and to avoid interfering with the other detector layers.[3,6] 4.2 3D Sensor Technology The 3D sensors studied are produced from the Centro Nacional de Microelectronica - Instituto de Microelectronica de Barcelona (CNM-IMB) at the Universitat Autonoma de Barcelona (UAB) in Barcelona, Spain. At current, the plan for the IBL is to house 75% planar sensors and 25% 3D sensors. The main reason for this was that planar sensors are a well understood technology that is currently being used in the pixel detector of ATLAS, and the 3D sensors have never before been used on a large international experiment such as the ATLAS in the LHC. The specs of the 3D sensors were explained in section 3, but it would do well to reiterate the advantages of 3D sensors over the current pixel detector. The current pixel detector sensors are degrading due to the radiation environment. Accordingly, the bias voltage required for full depletion has increased along with the leakage current. This requires an extensive cooling system. 3D sensors, even with equal radiation damage, will give off less heat due to a lower required bias voltage for full depletion of the 3D sensor. With the planned increase in luminosity, the LHC detectors could take 1000 fb 1 data which would significantly extend the LHC's physics reach.[ 14] The 3D sensors can cope with this increased data intake due to their faster charge collection over the current pixel detector. Because of the geometry, there is a smaller charge trapping probability, and there is no charge shift from Lorentz angle. Since the 3D sensors are more radiation hard than the current pixel detector, its leakage current will also be less, again, requiring less cooling. 26

27 4.3 Front-End Chip The signal from the pixel sensors is read out by a front-end chip attached to the sensor. When a charged particle traverses the pixel sensor's depleted zone, it ionizes the sensor which carries information about the particle with it. The front-end chip amplifies the collected charge and sends a digital output signal from the pixel sensor to be evaluated. The current front-end chip used in the current pixel detector is the FE-13, but an upgraded FE-14 is required by the IBL. The FE-13 does not have a high enough hit rate to keep up with the planned increase in luminosity, and its active footprint is too small. The high luminosity would cause the FE-13 column drain architecture to saturate.[4] FE-14 The upgraded FE-14 front-end chip employs 130 nm technology and features an array of 80 x 336 pixels with a pixel size of 50 x 250 sm 2. Its larger chip size of 20.2 x 19.0 mm 2 results in a larger active fraction than before, increasing it from 74% to 89%. Since the IBL radius is too small to tilt the modules as done in the current pixel detector, it is vital to increase the active fraction area.[4] The larger chip also reduces periphery and, therefore, the cost of manufacturing it. The FE- 14 front-end chip also has a primary output rate that is four times faster than that of the FE-13. A picture of the FE-I4 front-end chip alongside a planar pixel sensor chip is shown in Figure 12. The FE-14 was developed to improve on the FE-I3's resolution, performance, and radiation hardness. The radiation tolerance of the FE-I4 exceeds the specified 300 Mrad. The FE-14 is bump-bonded to the 3D pixel sensor chip using a tin-silver (Sn-Ag) solder. The process involves heating the substrate and the chip to melt the bumps in order to establish an electric connection. [4] Figure 12: Photographs of an FE-14 USBpix system: (left picture) A multi-io board on the left side and an FE-I4 adapter card on the right, (right picture) a FE-14 single chip card 27

28 5. Methods The 3D pixel sensors will be mounted to an FE-14 chip as explained in section 4. Before the pixel sensors can be used in the IBL, they must be tested and tuned; each pixel can differ slightly and must be configured similarly with all the other pixels on the chip. This is achieved through various tests first characterizing the sensors then tuning them based on their behavior D Pixel Sensor Tests Before any test is scan is done or any test administered to the chip, digital and analog scans are done. These scans consist of injecting a certain charge several times through the front-end pixels and receive a reading back. This makes sure all the pixels are operating and can receive charge as well as are capable of sending signal. If equal numbers of events for the digital and analog tests are recorded, the chip is operating properly, and further scans and tests described in this section may be done Current vs. Voltage (I-V) The current vs. voltage scan done is classified as a bare sensor property. The scan is done directly to the sensor, and the how the leakage current depends on the bias voltage is a property of the sensors. At a certain voltage, known as the breakdown voltage, the charge carriers resulting from the bias voltage, or charged particles when in use in the IBL, have enough energy to ionize the semiconductor which produces more electron-hole pairs, which in turn leads to more ionizations. This kind of avalanching cascade saturates the pixel and detector. The goal is to operate below the voltage breakdown to avoid damaging the pixel and electronics while maximizing the bias voltage and minimizing the leakage current. Both of the goals work to lower the noise as shown in equation The leakage current has three major contributing components: volume current generation, surface generated current, and avalanche breakdown current. The volume current generation is produced by the bias voltage as it generates charge flow. From equation 3.12, it can be seen that the volume current increases as the square root of the bias voltage, JV7ias The surface generated current arises due to large radiation doses that are to be experienced when in use in the IBL. Specifically, the large radiation doses sustained by the IBL induce SiO 2 -Si interface states that produce a surface current. The last contribution to the leakage current is the avalanche breakdown current which occurs when the charge carriers produced by charged particles from the LHC then 28