showers with LOFAR Laura Rossetto 35 th ICRC July 13 th 2017 - PowerPoint PPT Presentation

Characterisation of radio frequency spectrum emitted by high energy air showers with LOFAR Laura Rossetto 35 th ICRC – July 13 th 2017 – Bexco, Busan, South Korea

The LO LOw w F Frequency requency AR ARray ray The M. van Haarlem et al., A&A 556, A2, 2013 → 50 stations in Northern Europe: The Netherlands, France, Germany, Ireland, Poland, Sweden, and UK → 24 stations ( ~ 2 km radius) located in Northern Netherlands form the LOFAR core → 6 stations ( ~ 320 m diameter) form the LOFAR “Superterp” 2 2 Laura Rossetto – 35 th Laura Rossetto – 35 th ICRC – July 13 ICRC – July 13 th th 2017, Busan 2017, Busan

The LO LOw w F Frequency requency AR ARray ray The Each Dutch station has: → 96 Low Band Antennas (LBAs) frequency = 10 – 90 MHz → two orthogonal dipole arms with orientation NE – SW and NW – SE → 48 High Band Antennas (HBAs) frequency = 110 – 240 MHz 3 3 3 3 Laura Rossetto – 35 th Laura Rossetto – 35 th ICRC – July 13 ICRC – July 13 th th 2017, Busan 2017, Busan

The LO LOw w F Frequency requency AR ARray ray The Posters id 402 id 413 → the six central stations are instrumented with 20 scintillators which give the main trigger for Cosmic Ray event detection → cosmic ray E = 10 16 – 10 18 eV S. Thoudam et al., Nucl. Inst. Meth. A 767, 339, 2014 4 3 4 3 Laura Rossetto – 35 th Laura Rossetto – 35 th ICRC – July 13 ICRC – July 13 th th 2017, Busan 2017, Busan

Radio emission processes of EAS Radio emission processes of EAS Geomagnetic synchrotron emission Charge excess emission → separation of charged particles due to → negative charge excess produced at the geomagnetic field the shower front A. Nelles et al., JCAP 05, 18, 2015 O. Scholten et al., PRD 94, 103010, 2016 3 5 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 5 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Frequency spectrum study Frequency spectrum study GOALS: → to characterise the pattern of radio signals in the frequency–domain → to improve the reconstruction of the showers, i.e. position of the shower axis at ground, energy and mass composition of primary particle ANALYSIS: → Analysis applied to data collected by LOFAR since 2011 and CORSIKA/CoREAS simulated showers D max = 3.6 km D max = 5.9 km D max = 6.1 km 6 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 6 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Frequency spectrum study: the method Frequency spectrum study: the method 1. Signal in the time–domain has been converted to the frequency–domain by applying a Fast Fourier Transform on Δ t = 128 samples = 640 ns (1 sample = 5 ns) 3 7 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 3 7 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Frequency spectrum study: the method Frequency spectrum study: the method 1. Signal in the time–domain has been converted to the frequency–domain by applying a Fast Fourier Transform on Δ t = 128 samples = 640 ns (1 sample = 5 ns) 2. |FFT| 2 Signal → evaluated on Δ t = [ t 0 – 240 ns , t 0 + 400 ns ] where t 0 = time of the pulse–peak 3 7 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 3 7 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Frequency spectrum study: the method Frequency spectrum study: the method 1. Signal in the time–domain has been converted to the frequency–domain by applying a Fast Fourier Transform on Δ t = 128 samples = 640 ns (1 sample = 5 ns) 2. |FFT| 2 Signal → evaluated on Δ t = [ t 0 – 240 ns , t 0 + 400 ns ] where t 0 = time of the pulse–peak 3. |FFT| 2 Background → evaluated on 400 sub-windows outside the pulse region 7 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 7 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Frequency spectrum study: the method Frequency spectrum study: the method 1. Signal in the time–domain has been converted to the frequency–domain by applying a Fast Fourier Transform on Δ t = 128 samples = 640 ns (1 sample = 5 ns) 2. |FFT| 2 Signal → evaluated on Δ t = [ t 0 – 240 ns , t 0 + 400 ns ] where t 0 = time of the pulse–peak 3. |FFT| 2 Background → evaluated on 400 sub-windows outside the pulse region 4. |FFT| 2 = |FFT| 2 Signal – |FFT| 2 Background 3 7 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 3 7 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Frequency spectrum study: the method Frequency spectrum study: the method 1. Signal in the time–domain has been converted to the frequency–domain by applying a Fast Fourier Transform on Δ t = 128 samples = 640 ns (1 sample = 5 ns) 2. |FFT| 2 Signal → evaluated on Δ t = [ t 0 – 240 ns , t 0 + 400 ns ] where t 0 = time of the pulse–peak 3. |FFT| 2 Background → evaluated on 400 sub-windows outside the pulse region 4. |FFT| 2 = |FFT| 2 Signal – |FFT| 2 Background 5. linear fit applied to log 10 |FFT| 2 in the range ν = 33 – 70 MHz log 10 |FFT| 2 = a + slope · ν 3 7 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 3 7 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Real data results Real data results → Event selection criteria: 1) at least 1 station with half antennas having signal > 10 σ → 142 events 2) events with at least 10 antennas with | FFT( ν i ) | 2 > RMS ( | FFT( ν i ) | 2 B ) → 103 events → Linear–fit applied to all antennas of each selected event 8 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 3 8 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Real data results – Shower axis distance Real data results – Shower axis distance → Event selection criteria: 1) at least 1 station with half antennas having signal > 10 σ → 142 events 2) events with at least 10 antennas with | FFT( ν i ) | 2 > RMS ( | FFT( ν i ) | 2 B ) → 103 events → Linear–fit applied to all antennas of each selected event → the slope–parameter shows a E = 1.7 · 10 17 eV parabolic distribution as function of X max = 757 g/cm 2 distance to the shower axis 3 9 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 3 9 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Real data results – Shower axis distance Real data results – Shower axis distance → Event selection criteria: 1) at least 1 station with half antennas having signal > 10 σ → 142 events 2) events with at least 10 antennas with | FFT( ν i ) | 2 > RMS ( | FFT( ν i ) | 2 B ) → 103 events → Linear–fit applied to all antennas of each selected event syst. unc. → comparison to simulations X max (sim) = 755 g/cm 2 → maximum around 100 m in agreement X max = 757 g/cm 2 with the Cherenkov ring region 3 9 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 9 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Real data results – D max correlation Real data results – D max correlation Slope at 180 m ( MHz –2 ) Slope at 220 m ( MHz –2 ) 33/103 events 22/103 events D max (km) D max (km) → the slope parameter depends on the geometrical distance of the observer from X max (i.e. D max ) α → at fix distances from the shower axis: slope ( D max )= D max )−γ 1 + exp (−β⋅ 10 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 10 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan

Conclusions Conclusions • Radio frequency spectrum has been studied in the range 30 – 70 MHz • linear–fit applied to real data detected by LOFAR since 2011, and to corresponding CORSIKA/CoREAS simulated showers • clear dependence of the frequency spectrum as function of distance to the shower axis has been obtained → this can be used as an independent method to reconstruct the position of the shower axis at ground Poster id 402 • At fixed distances from the shower axis, the frequency spectrum depends also on the geometrical distance to X max → independent method for X max reconstruction 11 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan 11 3 Laura Rossetto – 35 th ICRC – July 13 th 2017, Busan