Table 1 Algorithm used for peak extraction from the cathodoluminescence image line scans

so n, the size of the signal to study

so pi, the element at the position i in the signal

so pi−1, the element at the position i−1 (preceding the element)

so the boolean up (true if the curve is ascending, otherwise false)

so the boolean down (true if the curve is descending, otherwise false)

Buckle: from i = 2 until i = n

{In the run i: if pi−1 < pi [up = true check in the run i−1 if down = true yes = presence of a trough]

otherwise: if pi−1 > pi [down = true check in the run i−1 if up = true yes = presence of a peak]}

with the closest significant peak or trough

Nb: signal2 is the result of step 1

so s, the standard deviation of the signal to study

so k, an integer number

so r = s/k, r is the threshold value (all lower variation than r is considered as being noise)

[if the variation (up or down) between two points is lower than r, the peak is considered as

non-significant, r is determined by the operator after observation of the data.]

so m, the number of points extracted during step 1 (peaks and troughs)

so pi, the element at the position i in signal2

so pi−1, the element at the position i-1 (preceding the element i)

Buckle: from i = 1 until i = m

{If pi−1 < pi (trough case): if pi-1 <=pi + r non-significant trough, stacking with the previous trough

otherwise: if pi < pi−1 + r non-significant peak, stacking with the previous peak}