Improved radiation tolerance of MAPS using a depleted epitaxial layer A. Dorokhov ∗ ,a , G. Bertolone a , J. Baudot a , A.S. Brogna a , C. Colledani a , G. Claus a , R. De Masi a , M. Deveaux b , G. Dozi` ere a , W. Dulinski a , J.-C. Fontaine c , M. Goffe a , A. Himmi a , Ch. Hu-Guo a , K. Jaaskelainen a , M. Koziel a , F. Morel a , C. Santos a , M. Specht a , I. Valin a , G. Voutsinas a , F.M. Wagner d , M. Winter a a Institut Pluridisciplinaire Hubert Curien (IPHC), 23 rue du loess, BP 28, Strasbourg, France 67037 b Goethe-Universit¨ at Frankfurt am Main, Senckenberganlage 31, Frankfurt am Main, Germany 60325 c Groupe de Recherche en Physique des Hautes Energies (GRPHE), Universit´ e de Haute Alsace, 61, rue Albert Camus, Mulhouse, France 68093 d Forschungsneutronenquelle Heinz Maier-Leibnitz (FRM II), Garching, Germany 85748 Abstract Tracking performances of Monolithic Active Pixel Sensors (MAPS) developed at IPHC [1] have been extensively studied [2], [3]. Numerous sensor prototypes, called MIMOSA 1 , were fabricated and tested since 1999 in order to optimize the charge collection efficiency and power dissipation, to minimize the noise and to increase the readout speed. The radiation tolerance was also investigated. The highest fluence tolerable for a 10 μm pitch device was found to be ∼ 10 13 n eq /cm 2 , while it was only 2·10 12 n eq /cm 2 for a 20 μm pitch device. The purpose of this paper is to show that the tol- erance to non-ionising radiation may be extended up to O(10 14 )n eq /cm 2 . This goal relies on a fabrication process featuring a 15 μm thin, high resistivity (∼1kΩ · cm) epitaxial layer. A sensor prototype (MIMOSA-25) was fabricated in this process to ex- plore its detection performances. The depletion depth of the epitaxial layer at standard CMOS voltages (< 5V) is similar to the layer thickness. Measurements with m.i.p.s 2 show that the charge collected in the seed pixel is at least twice larger for the de- pleted epitaxial layer than for the undepleted one, translating into a signal-to-noise ratio (SNR) of ∼ 50. Tests after irradiation have shown that this excellent performance is maintained up to the highest fluence considered (3 · 10 13 n eq /cm 2 ), making evidence of a significant extension of the radiation tolerance limits of MAPS. Key words: Radiation Hardness, Monolithic Active Pixel Sensor, Tracking. PACS: 29.40.Gx, 29.40.Wk, 61.80.-x 1. Introduction 1 A MAPS [1] is a semiconductor detector, which uses a matrix 2 of p-n diodes (formed by N-well/P-epitaxial layer junctions in 3 CMOS technology) as a sensing element. The charge liberated 4 by the particle traversing the semiconductor volume is collected 5 and transformed into signal by the readout electronics. The pe- 6 culiarity of MAPS is that the readout electronics and sensing 7 elements (N-well diode) are implemented in the same CMOS 8 technology substrate (Fig. 1, left). This allows for a small pitch 9 (∼10 μm) and for a reduction of the equivalent noise charge 10 (10 to 20 e), because the sensing element is located very close 11 to the preamplifier. The standard CMOS technology substrate 12 has very low resistivity (< 1 Ω · cm), hence the epitaxial layer 13 (typically 5-20 μm thick), featuring a relatively high resistivity 14 (∼10 Ω · cm), is used as a sensing volume. 15 At standard CMOS voltages (< 5V), the depletion depth 16 of the epitaxial layer is a fraction of micrometer (Fig. 1, 17 right). The signal electrons get therefore predominantly col- 18 lected through thermal diffusion. Non-ionising irradiation in- 19 creases the trapping time for charge carriers in the epitaxial 20 ∗ Corresponding author Email address: Andrei.Dorokhov@IReS.in2p3.fr (A. Dorokhov) Figure 1: Left: N-well/P-epi diode connected to a 3 transistor pixel cell, in a low resistivity epitaxial layer. The color scale describes the doping concentration [cm −3 ]. Right: Potential [V] calculated using Synopsys TCAD [4]. The white zone shows the depleted region. Preprint submitted to Nuclear Physics A July 23, 2009