Proton irradiation of analogue opto-hybrid circuits for the CMS tracker

Abstract Four analogue opto-hybrid circuits to be used in the front-end electronics of the Compact Muon Solenoid (CMS) tracker were irradiated with 26 MeV protons to fluences of 0.5×10 14 and 0.8×10 14  p/cm 2 in the secondary beam of the cyclotron at the Forschungszentrum (FZK) in Karlsruhe (Germany). The purpose of the test was to measure the tolerance of the circuit to a radiation exposure as that foreseen in the CMS Tracker in 10 years of life. To this end, a set of electrical measurements of the static and dynamic behaviour of the irradiated circuits were made before and after the test. In order to monitor the variations due to the increasing irradiation, the output light power of the opto-hybrids was constantly measured during the test. The measurements taken during the irradiation test are reported and discussed in this paper and the circuit performance, tested before and after the irradiation, is compared.


Introduction
The CMS silicon tracker [1], with its capability of collecting a huge amount of data from high energy proton-proton interactions in the LHC tunnel, is an outstanding example of the frontiers achieved in the fields of physics and technology.
About 10 million channels will be involved in the trajectory reconstruction and this has led to the development of front-end electronics which are fast, compact and very resistant to the harsh radiation environment. The calculated energy integrated fluence for fast hadrons in 10 years is 1.6 14 cm -2 in the inner tracker layer and 0.35 14 cm -2 in the most outer layer [1].
Optical links [2,3], typically 60 m long, will allow the transfer of data at a frequency of 40 MHz from the detector to the control room, where analogue pulses are digitised and processed at the Front-End Driver (FED) [4]. Two types of optical links will be used in the CMS tracker, a bi-directional digital control link and an analogue readout link.
The conversion of analogue electrical signals into the corresponding optical ones is done by means of an analogue opto-hybrid (AOH) circuit. Several versions of it were developed and tested by the CMS collaboration in Perugia, Vienna and at CERN [5,6]. The specific geometric requirements for the various parts of the tracker, i.e. inner barrel and inner disks (TIB/TID), outer barrel (TOB) and end caps (TEC) led to three layouts of the AOH circuit. The total number of AOHs to be built and tested for the CMS tracker amounts to about 17000. Figure 1 shows the scheme of the analogue part of the optical link. The pulses induced by charged particles passing through the silicon detector are sampled and multiplexed by APV chips [7] and MUX. The resulting differential input voltage is converted by the AOH into an analogue optical signal and is transmitted via optical fibres to the back-end electronics. Each fibre will carry the multiplexed signals of 256 microstrips.
Silicon detector APV MUX (2:1) Patch panels group together the fibres into optical cables that arrive at the back end electronics in the control room. Here optical signals are re-converted into electrical and are processed.

The analogue opto-hybrid circuit
The analogue opto-hybrid circuit for the CMS silicon tracker consists of a FR4 (vetronite) substrate with dimensions of 30x22x0.5 mm 2 for the TIB/TID and slightly larger, 30x23x0.5 mm 2 , for the TOB and TEC.
All three AOH types are equipped with one linear laser driver (LLD) chip [8] and 2 or 3 laser diodes. The LLD is programmable via an I 2 C interface and can be set to bias the laser diodes in their linear operational region. Four increasing gains (5, 7.5, 10 and 12.5 mS) of the differential input signals are pre-settable through the LLD in order to compensate for threshold variations due to radiation damage.
The laser diodes are commercially available InGaAsP edge-emitting coupled to single mode optical fibers operating at a wavelength of 1310 nm. They are glued to the substrate with a thermally conductive resin which is resistant to radiation and are ultrasonically bonded. A version of the TOB AOH with laser diodes soldered to the substrate has also been developed, but will not be used in the final system. Figure 2 shows the pictures of the AOHs irradiated in Karlsruhe, where (a) shows the TIB/TID AOH front and back side, (b) the TOB AOH with laser diodes bonded on the substrate and (c) the TOB AOH with soldered laser diodes. The link to the detector module is done through a 30-pin NAIS connector, socket type for the TIB/TID and TEC and header type for the TOB. Due to mechanical constraints, in the TIB/TID AOH the LLD and the connector are on the opposite side of the substrate with respect to the laser diodes, while they are on the same side for the TOB and TEC versions. In order to protect the bonds during assembly and handling, wires are hard covered with a plastic (ABS) stamp (not present in fig. 2).

Proton irradiation
The irradiation of the analogue opto-hybrid is the final step in the qualification of this circuit for resistance to radiation damage.
Previous qualification was dedicated only to components of the AOH, i.e. laser diodes, fibres and connectors [9,10], LLD [8], substrate, passive electrical components and glues [11]. The results obtained already gave a good level of confidence on the AOH radiation hardness. Nevertheless, the response to radiation had also to be proven once assembled with all its parts.
Four AOHs, two TIB/TID and two TOB, were irradiated in a secondary beam of the KIZ isochrone cyclotron at the Forschungzentrum in Karlsruhe. The irradiation source is a proton beam of 26 MeV with a current adjustable from 400 nA to 2 µA. Figure 3 shows the beam area.  Figure 3. Picture of the cyclotron beam area with the scanning system facility.
A scanning system, placed on a rail, allows uniform irradiation of the target. The circuits are contained inside the sealed box placed in front of the beam (represented by the white X in fig. 3). The box to the right contains the nitrogen used to cool the target.
The electrical setup used in the AOH's irradiation was built to drive the AOH circuit without the detector module, like in the final readout chain shown in fig  The AOHs are plugged onto two PCBs inside the sealed box, represented by the dotted rectangle in the right of the figure. They are powered and biased through the five line flat cables (V DD , V SS , V 125 , clock and data) arriving from the front-end electronics (gray test box on the left). An interface connector board (centre of the figure) splits the signals from the test box to feed both AOHs under test. The optical fibres from the AOHs run outside the beam area to an optical power-meter (Agilent 8163A lightwave multimeter). The twisted flat cable and the power supply cable go, respectively, to a VME module and power supply generator located in the control room.
Two irradiation runs were done, with 2 AOHs at each time. Several beam stops were requested during the test, in order to monitor the rate of degradation after each irradiation step. In the first irradiation run the beam current was initially set to 400 nA and then rose up to 1 µA after 10 steps, reaching a total fluence of 0.54 14 p/cm 2 . The second irradiation was performed in 10 steps, each about 1 min long, with a beam current of 1 µA and a final fluence of 0.76 14 p/cm 2 . Both fluences are calculated with the beam current and are affected by a systematic error of 15%. The temperature inside the sealed box was measured by a probe placed near the AOHs and varied between -8 and -2 °C during the first irradiation run and between -18 and -14 °C during the second. All the results reported in this paper refer to the two AOHs (one TIB/TID and the soldered TOB) irradiated at higher fluence. x

Sealed box
Nitrogen box

Monitoring during the irradiation
The signals carried from two optical fibres (one for each AOH) were measured during the irradiation test by the Agilent 8163A lightwave multimeter.
The measurement of the output light power as a function of time is shown in figures 5a and 5b for some increasing steps of the irradiation. The temperature inside the cooled box is reported together with the increasing fluence value. Before the beam was switched on, the laser diode was biased with a current of 45 mA, well above its threshold current that is usually around 5 mA. The initial decrease of the output light power in figs. 5a and 5b is related to the self-heating of the laser diode. Once the output power is settled, after about 60 s, the AOH is ready to be irradiated. When the beam crosses the laser diode under test, the emitted power suddenly decreases until the beam stop. A partial recovery of the initial light power was observed after some tens of second. The maximum variation of the initial output light power during the irradiation amounts to about 100 µW for the TIB/TID AOH and about 200 µW for the TOB AOH. The relationship between this variation and the irradiation is not so evident, since there is no clear correlation in these results between the expected decrease in laser diode efficiency with increasing fluence. This is instead more evident in the second type of monitoring described below. The measurement of the output light power as a function of the bias current of the same two laser diodes (channel 0 of each AOH) was done soon after each irradiation step and is shown in figure 6. The diodes were biased with an increasing current from 0 to 54 mA in steps of 0.45 mA, that corresponds to one unit of the LLD register. In particular, the data in fig. 6 were plotted in the range 0-10 mA in order to show the laser threshold shift of about 3 mA caused by the irradiation. The temperature inside the cooled box was measured by a local probe, placed near the AOHs, and is reported together with the increasing fluence value. The dependence of the laser threshold on the temperature for non irradiated AOHs was measured in the range -15 °C to 25 °C. According to these measurements, the temperature variation reported in fig. 6 contributes to less than 10% of the total current shift.

Characterization before and after the irradiation
The monitoring of the AOH's output light during the irradiation is useful to check the immediate effect of the irradiation in the circuit behaviour. More specific tests were performed after 3 months, when the activity of the irradiated AOHs allowed a safe handling. These are the measurement, at room temperature (between 25°C and 28 °C) of some static and dynamic electrical parameters, i.e. gain, integral linearity deviation, RMS noise, bandwidth and crosstalk. The setup scheme for the AOH characterization before and after the irradiation is shown in figure 7. The AOHs are biased slightly above the laser diode threshold, in their linear operational range. The gain is the slope of the linear fit of the output voltage as a function of the input voltage in the range -300÷300 mV. The input pulse is generated by the DAC and is converted into a differential signal by the Analog Devices AD8192 amplifier. The optical output signal is transferred through the optical fibres to a 4-channel receiver module where it is converted back to an electrical output and sent to the ADC. Table I states the gain values measured on channels 0 and 1 of the TIB/TID AOH (a), and channels 0 and 2 of the TOB AOH (b), before (upper value) and after (lower value) the irradiation. The measurements show a general decrease of the gain, of 14% on average (with a spread between 0 and 30%) due to the radiation damage, which causes a lack of efficiency in laser diodes. The four columns correspond to the four LLD pre-settable gains (5, 7.5, 10 and 12.5 mS). The integral linearity deviation (or integral non linearity, INL) is also calculated by the linear fit of the output voltage of the optical link as a function of the differential input voltage in the interval -300÷300 mV (operational voltage range). The measurements before (a) and after (b) the irradiation are reported in figure 8 for the TIB/TID and TOB AOHs for a LLD gain setting of 5 mS. The horizontal solid lines in fig. 8(a) and 8(b) correspond to the maximum value of 1.5% for INL in the operational voltage range, as requested by the CMS tracker optical link specifications [12]. Both channels of the two AOHs satisfy this constraint, before and after the irradiation.
The output voltage RMS noise is measured by a high bandwidth ADC for differential input voltage levels varying from -300 and 300 mV. The results are reported in figure 9 before (a) and after (b) the irradiation. Some increase in the noise after the irradiation was observed in all channels except for channel 2 of the TOB AOH, where even some decrease is present. The horizontal solid lines are again the limits for the average noise allowed by the specifications for the CMS tracker optical link [12].   and that of the input gives a measurement of the bandwidth at -3 dB. A comparison of the results obtained before and after the irradiation is reported in figure 10. The bandwidth at -3 dB (this value is indicated by the horizontal solid line) measured on the two channels of each AOH before and after the irradiation is about 100 MHz, above the minimum of 90 MHz requested by the CMS tracker optical link specifications [12]. The measured values are in fact limited by the bandwidth of the AD8192 amplifier rather than the optical link itself. The crosstalk between the AOH channels was measured by injecting a fast input signal from the pulse generator (Agilent 81110A/81112A, rise time 0.8 ns) (see fig. 7) on one channel and measuring the corresponding output on the other after 20 ns with the scope. Figure 11 shows the crosstalk on channel 0 of the TIB/TID AOH before and after the irradiation. The limits, represented in the figure with the dashed lines, correspond to the maximum value of the crosstalk allowed by the CMS tracker specifications for the optical link of -54 dB [12].
(a) (b) Figure 10. Bandwidth of two channels of the TIB/TID and TOB AOHs before (a) and after (b) the irradiation.
(a) (b) Figure 11. Right axis: crosstalk measured on channel 0 of the TIB/TID AOH before (a) and after (b) the irradiation. Left axis: normalized output voltage on channel 1.

Conclusions
The irradiation of analogue opto-hybrid circuits with a proton beam of 26 MeV up to a fluence of 0.8 14 p/cm 2 has shown an increase of the laser diode threshold current of about 3 mA. Some decrease of the gain of about 14% on average was measured. The small effect on electrical parameters, like linearity, noise, crosstalk and bandwidth demonstrates the radiation tolerance of such electronic devices for proper functionality during the CMS lifetime. These low levels of damage are also consistent with earlier test results and can be easily compensated by the LLD.