**3. Results**

Plasma emission spectra in the range of 652.0–659.0 nm at di fferent times after the laser pulse (2–5 μs) and with several gate times (100, 200, 300, 500, and 1000 ns) were obtained. The spectra were measured in LSP conditions (flow water over the sample). The main feature of the spectrum was the appearance of the H α-line emission. In contrast to the spectrum obtained in air, in this case and as was observed in our previous work (Moreno-Diaz et al., [24]), the second order line of 327.39 nm Cu I does not appear.

The results obtained about Stark width and red shift after the fitting processes of the experimental data are presented in Table 1. In the first column the delays are presented, and in the successive columns, for each gate time, the corresponding Stark width and shift are shown. In these data, <sup>Δ</sup>λ1/2 (Å) and *d* (Å) represent the Lorentzian component (in Å) of the Voigt profile (once the instrumental contribution had been discounted) and the Stark red shift, respectively.

**Table 1.** Experimental Stark broadening (Δλ1/2 (Å)) and red Stark shift (*d* (Å)) of the Hα-line obtained in the LSP experiments of this work.


The experimental red shifts obtained in this work were a ffected by a general uncertainty of around 15%, except for those that appear in parentheses (with time gates of 100 ns and 200 ns and for which the statistics of the data are poor, and the uncertainty may be higher). However, they are included in the table since they seem to indicate the experimental trend indicated by Griem [27]. The 100 ns and 200 ns time gates appear to have light-integration times that are too short. These low statistics are the reason why the use of short gate times for measurements in industrial conditions are discouraged.

An example of these results can be observed in Figure 6, where the displacement of the maximum emission of the H α-line can be clearly appreciated.
