Characterization of irradiated test structures for the CMS tracker upgrade

Abstract The CMS collaboration is currently conducting a campaign to identify radiation-hard materials for an upgrade of the CMS tracker. This upgrade is needed to be able to cope with the higher radiation background of the future HL-LHC; additionally the performance of the current tracker will be significantly degraded at the time of the upgrade, requiring a replacement. Several different test structures (TSs) and sensors have been designed for a 6 in. wafer layout. These wafers were produced by an industrial supplier (Hamamatsu Photonics K.K.) and differ by their bulk material (Float Zone, Magnetic Czochralski and CVD-Epi), thickness (from 50 μ m to 320 μ m ) and N–P type doping. These TSs consist of different microelectronic devices including diodes, resistors or MOS structures. They enable the extraction of parameters which are not accessible in a silicon detector and allow the assessment of the quality of the sensors produced on the same wafer. The TSs have been irradiated with protons and neutrons to emulate the radiation damage caused by the particle fluence inside the future CMS tracker after 10 years of operation. This contribution will present measurements of non-irradiated and irradiated test structures at different fluences. The changes of the properties of the microelectronic devices will be discussed as well as the design of the TSs.


Introduction and motivation
The current Compact Muon Solenoid (CMS) tracker at the Large Hadron Collider (LHC) has to withstand high levels of radiation due to the high luminosity of the LHC of up to 10 34 cm -2 s -1 . This harsh radiation environment leads to a degradation of the silicon tracking sensor performance. Effects on the sensor are e.g. a change of the bulk resistivity following by an increase of the full depletion voltage, or a change of the coupling capacitance. It has been estimated that the detector performance will be insufficient after 500 fb -1 of delivered luminosity 1 [1]. Thus, a replacement of the current tracker will be required.
Additionally, an upgrade of the LHC is scheduled for 2022. At this upgrade, called HL-LHC, it is planned to achieve an integrated luminosity of up to 3000 fb -1 during its lifetime. Consequently, even higher radiation hardness requirements are imposed on the upgraded tracker. Furthermore, a higher granularity of the sensors is needed to account for the higher detector occupancy in the future.
In order to meet these requirements demanded in the future, the HPK 2 -campaign aims to define the future tracker sensor baseline.
Email addresses: bernhard@lutzer.eu (Bernhard Lutzer on behalf of the CMS collaboration), http://www.hephy.at (Bernhard Lutzer on behalf of the CMS collaboration) 1 This is about the equivalent of 10 years of operation. 2 Hamamatsu Photonics K.K.

The HPK-campaign
At first, a common wafer layout has been designed. The layout, shown in figure 1, features different structures such as teststructures, baby sensors or MSSDs. Each of these structures consists of sub-structures. They are used to investigate different aspects of the future tracker baseline [2], e.g. the choice of wafer material or sensor layouts. This article focuses on the two test-structures (TSs) which are highlighted with red rectangles in figure 1.

Wafer types
A set of possible wafer types to choose the new baseline from has been ordered from HPK. Only one manufacturer was chosen to ensure comparability of the measurement results.
The ordered material types are Float Zone (FZ), Magnetic-Czochralski (MCZ) and CVD-Epitaxy (Epi) silicon. For each type, P-strip implant in N-bulk material (N) and N-strip in Pbulk were produced. N-in-P-bulk types need strip isolation structures to avoid low interstrip resistances due to inversion layers. These are p-stop (P) and p-spray (Y). Moreover, a set of thicknesses (e.g. 320, 200 and 120 µm for FZ material) has been chosen to account for different positions in the upgraded tracker, see section 2.2.
Additionally, wafers with a second metal layer have been designed to investigate the possibility of integrated pitch adapters. In total, the campaign is evaluating 21 different combinations of material type and thickness with a total sum of 156 wafers. Table 1 gives an overview of the ordered wafer types. In the  current CMS tracker Float Zone P-in-N silicon with 320 µm thickness is used. Some of the delivered wafers differed from the order. A study of the delivered material showed for example that deep diffusion instead of wafer bonding wafers have been delivered [3].
All wafers were diced into the respective structures (TSs, MSSDs, etc.) and distributed among the responsible institutes. Thus, each institute has the same set of available wafers types which allows a holistic comparison and triggers an intensive institutional cooperation.

Irradiation plan
In order to investigate the radiation hardness of the different materials irradiation studies are performed. Most structures are irradiated with protons and neutrons at different fluences. The irradiation sites are TRIGA Mark II scientific reactor in Ljubljana, Slovenia and cyclotron in Karlsruhe, Germany. Each fluence corresponds to a radius inside the future CMS tracker. These radii (or distances from the interaction point) were derived from FLUKA simulations, see figure 2. Table 2 gives an overview of the fluences used for the irradiation campaign.
Previous research suggested an optimal thickness of the sensor depending on the location inside the tracker. An active thickness larger than 200 µm is sensible for 40 cm and 20 cm radius. The respective active thickness for 10 cm and 5 cm should be less than 200 µm [4].
Complementary, our institute conducted irradiation studies with the TRIGA Mark II reactor of the Institute of Atomic and Subatomic Physics of the Technical University of Vienna with lower neutron fluences than those of the HPK-campaign.

Measurement types
A test structure, shown in figure 3, contains several substructures like diodes, MOS capacitors or resistors [5]. Their first use was to monitor the process quality during the production of several thousand silicon sensors of the current CMS tracker. This undertaking proved very successful [6]. Now each test structure is characterized before and after irradiation to determine the change of detector properties, e.g. the full depletion voltage of diodes 3 . A comparison of those measurements allows to assess the radiation hardness of different wafer types. Table 3 shows the characterized parameters with the measurement type and the structure which is used therefor.
The motivation for using test-structures is that they allow the extraction of additional information about the process quality. For instance, the measurement of the dielectric breakdown voltage on a sensor is not possible without destruction.

Parameter
Meas. Structure

Results
This section presents an overview of test structures measurements. All samples have been annealed for 10 minutes at T=60 • C. The measurements have been performed at a temperature of T=-20 • C and in a low humidity atmosphere. With the exception of those in section 4.1, all results are of the HPK campaign.

C(V)-curves of irradiated MOS capacitors
Due to the high fluences used in HPK irradiation campaign the MOS characteristics changed crucially. In order to understand this behaviour, smaller fluence steps were necessary. Therefore, our institute performed complementary irradiations at the reactor in Vienna. Figure 4 shows the same C(V)-plot in different zoom steps. Each curve represents a different irradiation fluence between 10 11 and 10 14 cm -2 as noted in the plot legend.
A fluence depending transition behaviour of the C(V)-curve with different effects can be observed. The capacitance drop decreases with increasing fluences while the flat band voltage shows only minor deviations. This can be explained with an increasing bulk resistivity and minor changes of the fixed oxide charges. Figure 5 shows the flat band voltages of MOS capacitors before irradiation. This value is an important parameter regarding the oxide process quality. The behaviour of the C(V)-curves of MOS capacitors, discussed in 4.1, suggests that only unirradiated measurements make sense. The flat band voltage is homogeneous between N and P materials , but not dependent on the bulk material or the thickness (for N/P). V f b of N and P are generally low which is desired for a high process quality.

Sheet resistances
Different structures have been designed to evaluate the sheet resistance quality. The fluence dependent sheet resistivities can be found in figure 6. The n + -implant resistance shows a minor decrease with irradiation while the polysilicon resistance plot seems not to be dependent on the irradiation fluence. The only distinctive feature are higher values for the MCZ200N TSs. The aluminum resistance is hardly influenced by the irradiation. This measurement has proven to be quite difficult and will be replaced by a four-wire method in the future.

Interstrip resistance
The method used obtaining the interstrip resistance is only possible on test-structures. In contrast to sensors, the R Intstructure consists of isolated strips without bias resistors. Hence, an easy characterization of the pure interstrip resistance is possible. Figures 7 and 8 show the interstrip resistance for the different materials. A sharp decrease of the resistance can be found after irradiation. This can be explained by modifications of the bulk and implant resistivity.

Interstrip capacitance
The interstrip capacitance is shown in figure 9. One can see a major increase after the first irradiation step. Afterwards, a saturation can be observed for higher fluences.

Coupling capacitance and dielectric breakdown voltage
The coupling capacitance (figure 10) and the dielectric breakdown voltage (figure 11) are a parameters of the oxide quality. Those are measured on the C AC -structure.
The generally high breakdown voltages point to a high quality oxide. The coupling capacitances do not show a trend depending on the fluence.

Summary and Outlook
These first irradiation results show that as expected some parameters change with irradiation while others remain mostly stable. The hugest difference before and after irradiation can be found for the interstrip resistance, where the value changes over several orders of magnitude. While the values at the 20 cm fluence step of R Int are still fine, it is not yet clear if this will be the same for the highest fluences. The interstrip capacitance increases slightly, but shows a saturation for higher fluences while the n + -implant resistivity decreases slightly.
The poly silicon and aluminium sheet resistances, the coupling capacitance and the dielectric breakdown voltage do not seem to be dependent on the fluence.
The C(V)-curves of irradiated MOS capacitors showed interesting behaviours which could lead to further investigations in this area.
About the half of the irradiation steps of table 2 have been already made. The next irradiation steps with protons and neu- trons are currently undergoing and will be analysed soon. The new irradiation steps with higher fluences will show which materials are able to withstand the highest fluencies and are therefore suitable for the production of the new CMS tracker which should be in operation in 2022. Complementary, our institute is currently scheduling new irradiations at the scientific reactor in Vienna to enhance the experience at lower fluences.

Acknowledgements
The research leading to these results has received funding from the European Commission under the FP7 Research Infrastructure project AIDA, grant agreement no. 262025. I would like to thank all collaborators within this campaign. In the scope of the irradiations in Vienna I would like to thank all our collaborators at the Institute of Subatomic and Atomic Physics of TU Vienna. Additionally I would like to thank my colleague E. Fruehwirth for assistance in taking the measurements and W. Treberspurg for valuable help in performing the irradiation studies in Vienna.