Temperature Dependent Investigations of Dark Current E. Engelmann, - - PowerPoint PPT Presentation

1

Extraction of Activation Energies from Temperature Dependent Investigations

• f Dark Current
• E. Engelmann, S. Vinogradov, F. Wiest, E. Popova, P. Iskra, W. Hansch
• T. Ganka, Ch. Dietzinger, W. Gebauer, S. Löbner, A. Márquez Seco, R. Fojt

PhotoDet15, 9th of July, 2015

Motivation

PhotoDet15, 9th of July, 2015 2

Goal of this work:  reduction of dark count rate of Silicon Photomultipliers

• gain initial information on dark generation and extract contributions

to dark current General approach:

• activation energies determined from T dependencies are expected to

be a good indicator of physical mechanisms

• conventional methods of extraction of Eact at fixed voltages/overvoltages

may not be suitable

• effects dependent on voltage and overvoltage cannot be separated

Proposed method:

• independent measurements of photo- and dark-response
• separation of overvoltage dependent responsivity and

voltage dependent high-field effects

• find expression for field-independent generation component

PhotoDet15, 9th of July, 2015 3

• temperature dependent investigations were performed in a range from 20°C to -30°C
• the measurements were executed on a KETEK 3x3mm2 SiPM which was mounted on a

Peltier element and evacuated in a TO8 module

KETEK 3x3 mm2 in a TO8

Conventional Method

PhotoDet15, 9th of July, 2015 4

• conventional method as proposed by R.Pagano et al.; „Dark Current in Silicon Photomultiplier Pixels:

Data and Model“; IEEE Transactions on Electron Devices; Vol.59 NO. 9; 2012 is not suitable here

• Eact can not be attributed to a certain mechanism

Proposed Method

PhotoDet15, 9th of July, 2015 5

• dark and illuminated data was taken
• assumption of an equal responsivity R, for electrons
• riginating from dark generation and photoelectrons
• in general this approach is applicable for any

Response ρ, e.g. Idark or DCR

1st Approach

PhotoDet15, 9th of July, 2015 6

• analysing Igen consisting of a multiplied and

non-multiplied component

• Igained is assumed to be a small fraction
• f Inot_gained

V0 I0

1st Approach-Reconstruction of Dark Current

PhotoDet15, 9th of July, 2015 7

2nd Approach-Determination of Inot_gained

PhotoDet15, 9th of July, 2015 8

• fitting the not multiplied component yields an expression for V independent non-multiplied generation
• dark currents at all investigated temperatures could be fitted with good agreement in the voltage

range V<15V using the square root dependence of V

PhotoDet15, 9th of July, 2015 9

2nd Approach-Determination of Igained

• in order to determine the multiplied component, the difference between the measured dark current

and Inot_gained is investigated as a function of the responsivity

• Idiff could be described with a parabolic function

in good agreement in the range between R=0 to R=4x106

2nd Approach-Reconstruction of Dark Current

PhotoDet15, 9th of July, 2015 10

Results-Activation Energies

PhotoDet15, 9th of July, 2015 11

• Igained shows two activation energies E1

act≈ Eg and E2 act≈ Eg/2

• Fhigh_field lowers the effective activation energy by ΔEact
• ΔEact is close to expected value for Poole-Frenkel effect

Confirmation of Model

PhotoDet15, 9th of July, 2015 12

• Eact extracted from DCR is a sum of

field-independent Igained and field-dependent Fhigh_field

• DCRmeasured and DCRreconstr show comparable Eact

within the uncertainties

• this result is an indicator of the parameter fit quality

Reconstruction of DCR

PhotoDet15, 9th of July, 2015 13

• DCRreconstr overestimates DCRmeasured
• internal dark generation rate is expected

to be higher than DCR

• the measured DCR has to be addtionally

corrected for avalanche triggering probability

Summary

PhotoDet15, 9th of July, 2015 14

Results so far

• the chosen model for dark current could describe the measured data

to a sufficiently precise level

• field-enhanced effects could be separated from generation components
• the extracted activation energies indicate that dark currents at T>-5°C are diffusion

dominated, whereas currents at T<-5°C are dominated by generation (KETEK devices) Further investigations

• confirmation of model for different type of devices
• identification of micro-cell regions as origin for diffusion currents
• change technological process in order to reduce DCR

PhotoDet15, 9th of July, 2015 15

Thank you for the attention

PhotoDet15, 9th of July, 2015 16

Additional Slides

Results-Eact at fixed Overvoltage

PhotoDet15, 9th of July, 2015 17

• the extracted Eact directly from Idark (T) and DCR (T) at a fixed overvoltage show slightly different

values, but agree within the uncertainties

• Eact from „raw“ data is an indicator for physical mechanisms
• for a precise analysis, a more advanced analysis is necessary

Results- SensL

PhotoDet15, 9th of July, 2015 18

• a 3x3mm2 C-Series device from SensL was investigated for comparison
• only one slope could be observed in the Arrhenius plot
• Eact of (0.57 ± 0.02) eV is attributed to generation current
• the contribution of diffusion current is expected to be suppressed for this device

Reconstruction of Dark Current

PhotoDet15, 9th of July, 2015 19

Photocurrent

PhotoDet15, 9th of July, 2015 20

DCR in extended T range

PhotoDet15, 9th of July, 2015 21

Fhigh_field at fixed Responsivity

PhotoDet15, 9th of July, 2015 22