A lithium laser-ablation based time-of-flight (LILA-TOF) diagnostic for measuring plasma edge ion temperature and toroidal plasma rotation

The measurement of ion temperatures with good spatial resolution at the periphery of fusion plasmas is much more challenging than for their electronic counterpart. While a wealth of techniques exists for evaluating electron temperature, T_e, at the edge and scrape-off layer (SOL) of magnetic confinement plasma, ion temperature, T_i, diagnostics have been limited to the retarding field analyzer (RFA) [1] and the ion sensitive probe [2], both of which are susceptible to degradation and are highly sensitive to alignment errors as well as to model assumptions. On the other hand, passive spectroscopy with sub-Doppler resolution [3] can provide line-integrated T_i values if a good signal-to-noise ratio is achieved and no anomalous velocity components exist, which is not the case in many instances [3, 4]. In contrast, using a different approach, Pitcher and Stangeby measured the toroidal dispersion of a thermal He beam source to validate the recorded ion temperatures in a He plasma [5]. Since ion–ion collisions are the dominant term in the heating of impurities (typically born with kinetic energies of a few eV) along field lines at the relatively highly collisional plasma boundary, the direct rate of thermalization of these impurities with the background species, rather than the local ion temperature value, represents a figure of merit for many simulations.

In some previous works made on the stellarator TJ-II [6, 7], the He/He⁺ line ratio method was successfully tested as a diagnostic for T_i measurement in the plasma boundary region of this device. When combined with the traditional He line-ratio technique and a suitable Collisional-Radiative model [8], it allows microscopic edge plasma parameter profiles to be obtained, i.e. T_i, T_e, and n_e. However, the presence of plasma rotation as a drift term for the locally generated He ions represented an important setback and complicated the interpretation of the diagnostic [7]. Although several methods for the measurement of T_i at the edge of TJ-II have been tried (RFA, He beam, Doppler spectroscopy...) no systematic recording of this important parameter has been achieved so far. In contrast, the injection of a well-localized source of lithium, both in position and in time, opens up the possibility for transport studies in the periphery of such magnetic confinement plasmas. Such a source can be generated using laser-ablation as ablated Li is ionized rapidly on entering the plasma and thereafter transported along magnetic field lines. It should be noted that although the experimental set-up is very similar to the set-up of the laser blow-off technique used on the TJ-II [9], detection is focused principally on the first ionized state of the injected impurity. For this, Li light monitors are located at different toroidal positions around the device. It is considered that, during the toroidal displacement of the injected lithium, collisions with plasma particles lead to rapid ionization of the Li atoms and thereafter to rapid thermalization of the resultant cold Li⁺ ions. This equilibration process can be followed by time-of-flight (TOF) measurements using monitors located at different distances from the source. For this, light sensitive detectors with narrow bandpass filters are used. In this way, the resultant signals can be analysed to extract the parameters of interest.

The goal of this work is to demonstrate the viability of following the toroidal propagation of Li ablated by laser ablation off the inner vacuum vessel wall of the TJ-II device. It should be noted that lithium coatings are regularly applied to the inner surfaces of its vacuum chamber. When compared with other candidate atomic elements, lithium offers several positive benefits as a probing species. In particular, in TJ-II a lithium coating is regularly applied to the hardcore wall in this device to achieve improved density control. Moreover, re-deposition of lithium by the plasma ensures a constant supply of this element at the same location when performing repeated measurements. This simplifies the experimental set-up here. In addition, its low Z and mass result in negligible contamination of the plasma as well as good mass matching with main plasma species (D, T, He) used in TJ-II, the latter enabling fast thermalization with the background plasma. According to neoclassical theory, maximum energy and momentum transfer via inelastic collisions are obtained for particles with equal mass. A good mass matching implies fast thermalization of the injected lithium. Thus, the T_i value of the impurity ions transported along the plasma is constant, having reaches the plasma particle background T_i value almost immediately Moreover, being a low recycling species, only directly injected atoms will be probed by the detector. Since the amount of background Li in TJ-II plasmas is significant, a well localized injection (in time and space) is important to discriminate laser injected Li from sputtered Li and thus perform these measurements. Sputtered lithium continuously enters the plasma. Such a source does not allow synchronizing Li and Li⁺ signals to extract the TOF values of interest here, (this will be discussed in more detail later).

In addition, the large difference between the first and second ionization potentials results in a long dwelling time for the Li⁺ ions, hence, only local transport away from the injection point has to be considered as loss term at edge relevant parameters. The impurity has to dwell into the plasma long enough for full thermalization with the background plasma before being lost by imperfect confinement. Here, a long dwelling time signifies that the time for Li⁺ to be ionized to Li⁺² is long compared to thermalization times with background plasma and to expected TOF values.

In this work, which is performed on the stellarator TJ-II, the technique is described from first principles and examples of its use for studies of edge ion temperature and toroidal rotation are given. Basic aspects of the application of this new method to other fusion devices and possible challenges are also addressed.

When low-energy ions are introduced into plasma, they undergo Coulomb collisions with electrons as well as with main plasma ions and impurities. This leads to their full thermalization provided that the residence time within the plasma is long enough to allow for sufficient collisions. The rate, $t_{{\text{eq}}}^{ - 1},$ at which the laser induced breakdown spectroscopy (LIBS) injected impurity thermalizes with background (hydrogenic) plasma ions is given by the neoclassical expression [10]:

$$\begin{equation}t_{{\text{eq}}}^{ - 1} = C\frac{{{n_{{\text{e }}}}Z_{{\text{imp}}}^2{\text{ln}}\Lambda}}{{{M_{{\text{imp}}}}{T_{\text{i}}}}}\end{equation} \tag{ 1 }$$

where C is a constant, n_e is the electronic plasma density, M_imp and Z_imp are the main-ion mass and charge of the injected impurity respectively, m is the hydrogenic mass, ln Λ is the Coulomb logarithm, also assumed constant here, and T_i is the plasma ion temperature. It should be noted that if the injected impurity temperature is assumed to be much smaller than the T_i, (T_{i mp}≪ T_i), then the rate is maximized when M_imp = m. If a Maxwellian distribution function of velocities, v, f(v) is assumed for the thermalized ions:

$$\begin{equation}f\left( v \right) \propto T_{\text{i}}\,^{-{3}/{2}}{v^2}{e^{\Big({\frac {- {m_i}{v^2}}{{k_B}{T_i}}}\Big)}},\end{equation} \tag{ 2 }$$

where T_i is ion temperature, m is ion mass, and k_B is Boltzmann's constant. The distribution function f (v) can be converted into the distribution g(TOF) by changing the variables in the velocity expression using the Jacobian matrix of partial derivatives [11]. Here, TOF refers to time-of-flight. Then the transformation of the f(v) distribution to a g(TOF) distribution, where TOF = v/l, leads to:

$$\begin{equation}g\left( {{\text{TOF}}} \right) = 4\pi {\left( {\frac{{{m_i}}}{{2\pi {k_{\text{B}}}{T_i}}}} \right)^{{3}/{2}}}\frac{{{l^3}}}{{{\text{TO}}{{\text{F}}\,^4}}}{e^{ - \frac{{{m_i}}}{{2{k_{\text{B}}}{T_i}}}{{\left( {\frac{l}{{{\text{TOF}}}}} \right)}^2}}}.\end{equation} \tag{ 3 }$$

By differenciating the function g(TOF), we find that it peaks at

$$\begin{equation}{\textrm{TO}}{{\textrm{F}}_{{\textrm{max}}}} = l/\sqrt {5{k_{\textrm{B}}}{T_i}/m} {\mkern 1mu} {\mkern 1mu} .\end{equation} \tag{ 4 }$$

For data analysis, it is convenient to define a function h(u), where $u = {1}/{{\text{TOF}}^2}$ as:

$$\begin{align} h\left( u \right) & = \ln g\left( {{\text{TOF}}} \right) + 4\ln {\text{TOF}} \nonumber \\ & = {\text{ln}}\left( {4\pi {l^3}{\left( {\dfrac{{{m_i}}}{{\pi {k_{\text{B}}}{T_i}}}}\right)^{ {3}/{2}}}} \right) - \dfrac{{{m_i}{l^2}}}{{{k_{\text{B}}}{T_i}}}u .\end{align} \tag{ 5 }$$

So far, it has been assumed that ion velocity is determined solely by the thermalization process. However, if the contribution of the toroidal plasma rotation is considered, the observed velocity, v_x, would be formed by 2 components, one due to ion rotation (v_rot) and the other to the thermalization velocity (v_{t herm}) contribution, hence v_x is given by:

$$\begin{equation}{v_x} = {v_{{\text{rot}}}} + {v_{{\text{therm}}}}.\end{equation} \tag{ 6 }$$

Consequently, the corresponding g(TOF) function, in terms of the TOF, can be written as:

$$\begin{equation}g\left( {{\text{TOF}}} \right) = 4\pi {\left( {\frac{{{m_i}}}{{2\pi {k_{\text{B}}}{T_i}}}} \right)^{{3}/{2}}}\frac{{{l^3}}}{{{t^4}}}{e^{ - \frac{{{m_i}}}{{{k_{\text{B}}}{T_i}}}{{\left( {\frac{l}{{{\text{TOF}}}} - {v_{{\text{rot}}}}} \right)}^2}}},\end{equation} \tag{ 7 }$$

Equation (7) shows a maximum at:

$$\begin{equation}{\textrm{TO}}{{\textrm{F}}_{{\textrm{max}}}} = {v_{{\textrm{rot}}}} \pm l/\sqrt {2{k_{\textrm{B}}}{T_{\textrm{i}}}/m} {\textrm{ }}.\end{equation} \tag{ 8 }$$

These equations are applied to the TJ-II, a four-period heliac-type device [12] in which the experiments described in this work were performed. As an example, figure 1(a) shows the evolution of TOF simulated traces for Li⁺ ions (equation (3)) for a toroidal travel distance of l = 2.33 m (from sector A1, where laser ablation is performed, to sector A8, where a Li⁺ monitor is located, see figure 2). Note: ion propagation along TJ-II magnetic surfaces is considered when determining this l [13]. In the simulations in figure 1(a), no toroidal rotation is considered for Li⁺ ion temperatures between 10 eV and 45 eV. In contrast, figure 1(b) shows the effect of toroidal rotation on the TOF trace for T_i = 30 eV (see equation (7)). Here, a negative v_rot corresponds to anti-clockwise rotation from the injection point in figure 2 to the monitoring point, while a positive v_rot corresponds to a clockwise rotation in the same figure 3.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The experiments reported here were performed in the TJ-II, a four-period, low-magnetic-shear stellarator with major and average minor radii of 1.5 m and ⩽0.22 m, respectively [12]. For these experiments, plasmas were generated using hydrogen as the working gas in the standard TJ-II configurations and heated using electron cyclotron resonance heating (ECRH) operated at the 2nd harmonic of the electron cyclotron resonance frequency (ν = 53.2 GHz, P_ECRH < 600 kW). This gyrotron is coupled to the plasma by means of a quasi-optical transmission line and up to 250–300 kW of power can be injected in X-mode at densities up to 25 W cm⁻³. No auxiliary heating was provided. As a result, central electron densities and temperatures up to n_e (ρ = 0) $\, \cong \,$0.7 × 10¹⁹ m⁻³ (measured by a microwave interferometry) and T_e (ρ = 0) $\cong$ 0.6 keV (by electron cyclotron emission (ECE) monitors), respectively, were obtained, (ρ is the normalized effective radius). In the TJ-II plasma discharges are sequentially numbered.

A 6–8 ns pulse from a Nd:YAG laser (750 mJ) is used to ablate lithium off the inner wall of the vacuum chamber. Table 1 summarizes the main experimental parameters. Thin layers of lithium are deposited regularly on the inner walls of this chamber in order to improve plasma control [14]. The poloidal cross-section of the plasma is kidney shaped with strongest plasma/wall interaction in the region called the hardcore. The laser is focused through an optical vacuum viewport in sector A1 to a ∼1 mm spot size on the surface of the hardcore, at a position just above the central coil vertical position, using a quartz lens with f' = 900 mm which is located outside the vacuum chamber at 3 m from the laser (see figure 3).

Table 1. Experimental parameters.

Working gas	Hydrogen
Injected heating power	ECR: ν = 53.2 GHz
Power heating	<600 kW
Line averaged electron density (interferometry)	ne (ρ = 0) $\, \cong \,$ 0.7 × 1019 m−3
Central electron temperature (ECE)	Te (ρ = 0) $\cong$ 0.6 keV
Magnetic field on-axis	1 T
Nd:YAG laser energy	750 mJ
Nd:YAG laser pulse length	6–8 ns (single pulse: one pulse per discharge)
Nd:YAG laser wavelength	1064 nm

To understand this LIBS process, several consecutive phases can be differentiated: Firstly, the laser beam is focused on the wall of the vacuum chamber. Later, a laser pulse creates a micro-plasma, by ablating the outer layer of the wall, which propagates in the opposite direction to the propagation direction of the laser beam. This creates a micro-crater in the material (this phase is known as rupture). Consequently, the laser energy is captured in part by the chemical species that make up the wall, whose elements are excited, creating a plasma. During the expansion phase of the micro-plasma, a continuous emission spectrum is produced. Finally, the emission of this radiation causes a cooling since the chemical species lose energy through the emission of discrete spectral lines where the desorbed elements can be identified. Neutral lithium is first present in the local ablation micro-plasma. Since the plasma produced by LIBS expands, if there is hot plasma in the vicinity of this ablation plasma, the neutral LIBS plasma can penetrate through the closed magnetic field surfaces and come into contact with the hot plasma. The particles at the edge of the plasma, in particular the electrons, impact on the LIBS plasma inducing ionizations, and once partially ionized this plasma (or plasmoid) expands toroidally in both directions along the magnetic field lines as occurs with solid hydrogen pellets injected into TJ-II [15]. As this lithium plasmoid expands toroidally along the field lines collisions occur with plasma particles (particularly electrons) that thermalize the initially cold Li⁺ ions. Although the thermalization time competes with the TOF, for a first approximation, we assume that Li is ionized (2.5 μs), thermalized, and then propagated along the field lines. When this process reaches equilibrium, and the ions are thermalized, measurements of the TOF can be carried out. This consists of recording the temporal evolution of the intensities of a neutral Li emission line (Li I) in sector A1 and of a Li⁺ emission line (Li II) in other sectors, that is, at different distances from the source, as indicated in figure 2. Then, by analysing the evolution of the Li I and Li II emissions until ion thermalization occurs, TOF measurements can be obtained which will allow T_i to be deduced for the plasma periphery region, thereby resulting in a novel technique. The ablation of lithium and its subsequent displacement can be seen by following the evolution of neutral and singly ionized Li line emissions [16], the latter at different toroidal distances from the ablation location. During their toroidal transport, collisions with plasma particles thermalize the initially cold Li⁺ ions. Moreover, the use of monitors located in different machine sectors of TJ-II allows TOF measurements to be determined for a range of distances, l, as shown in figure 2, for instance, in sectors A3, D4, A8, C8, C3, where l = 0.67, 1.67, 2.33, 2.90, and 4.64 m, respectively [12].

For TOF measurements, a photomultiplier (PMT) detector (model R3896 by Hamamatsu, Japan) fitted with a suitable narrowband transmission filter (full width at half maximum = 1 nm, center wavelength = 548.5 nm, model 010FC10-50/5484, by Andover Corporation, USA) is used to follow Li II emission at 548.5 nm [17]. In addition, to follow neutral Li emissions in sector A1, a ½ m focal length spectrometer (ACTON Spectra Pro 500, Princeton Instruments, USA) with a PMT detector (H9305-04 Hamamatsu, Japan) is used (see figure 3). It is set to 670.7 nm, the wavelength of the resonant Li I emission line [17]. In addition, a broad wavelength range spectrometer (PMA-12 by Hamamatsu, Japan) collects radiation between 200 and 900 nm to provide a global view of plasma emissions, and x-ray monitors (AXUV detectors & Al/C filter) and bolometers (naked or unfiltered AXUV detectors), that collect global plasma radiation, are used to support the experiments. Finally, a light-sensitive photodiode located in sector A1, which collects scattered laser emission, permits precise timing of the laser pulse to be obtained.

A representative TOF trace is displayed in figure 4(a). Here, the experimental TOF is the time difference between ablation of the Li I and the Li II (Li⁺) arrival time signals for l = 2.33 m, i.e. the distance travelled along field lines from sectors A1 to A8. Several features are apparent from this figure. First, a sharp peak with a 13 μs full-width at half-maximum, is observed in the Li I emission. This corresponds to the Li ablation at 1100.26 ms along the discharge. Next, a peak is seen in the Li II emission at ∼0.9 ms later. This second peak is significantly broader than the Li I peak. Next, figure 4(b) shows some representative plasma signal traces together with the chord-averaged line density. As can be seen, there are no detectable perturbations in plasma density, H_α signal and outer electron temperature, as shown by ECE traces for the normalized minor effective radius, ρ = 0.64 and ρ = 0.75, during a 40 ms time window after injection, i.e. non-perturbative conditions are fulfilled.

Zoom In Zoom Out Reset image size

A fit to the Li II signal for T_i = 25 eV is made in figure 4(a). This fit is made using equation (8) for TOF_max. While excellent agreement is observed in this plot between 50 and 160 μs, fit values are systematically lower than experimental Li II values for longer TOFs. This becomes more evident if the Li II signal is plotted using the function h(u) defined in equation (5). See figure 5. In this case, a straight-line provides a good fit to TOFs values between 70 and 160 μs, with a slope corresponding to T_i∼ 26 eV, in good agreement with the previous value. However, for values outside of this range, a deviation from expected behaviour predicted by equation (3) is apparent. At TOF values shorter than 70 μs (right side of the figure) a straight-line fit yields a lower T_i value. Since a rapid rise of the Li II signal is observed at the onset of the trace in figure 4(a), under-sampling by the detector analogic-digital converter partly contribute to this, although other physical processes may also contribute significantly (see below). In contrast, for TOF values above 160 μs (left side in the figure), a fast rise of h(u) is observed and values of T_i ⩽ 5 eV would be deduced from a similar fit. In this time range, the Li II signal merges with the background signal and therefore this type of analysis becomes less reliable. Indeed, a possible contribution from a slower and colder tail in the velocity distribution, coming from the SOL region, cannot be ruled out as radial discrimination is not possible with the current detector geometry, and a neutral lithium cloud encompassing a fraction of the minor radius can be generated by the laser ablation process. Only by properly selecting the flux surface to be monitored with a suitable detection geometry could this issue be tackled. Efforts in this direction are presently underway.

Zoom In Zoom Out Reset image size

Possible modifications to lithium traces induced by toroidal plasma rotation are analytically described by equation (6). The largest expected change occurs in the location of the Li II signal maximum. The effect of such rotation is to displace the light peak in time, i.e. to shorter TOF values when rotation is negative and to longer TOF times when positive (see figure 1(b)). Thus, the contribution of negative or positive toroidal plasma rotation is to yield apparent T_i values that are higher or lower, respectively, than the true experimental ones. Moreover, the width of the distribution is also affected. For instance, a negative rotation leads to a narrower TOF distribution while a positive rotation broadens the same distribution. Simulations of distorted distributions are plotted in figure 6(a) and compared to a Li II signal recorded for an ECRH plasma, for n_e∼ 6 × 10¹⁸ m⁻³, made with He as the working gas. As shown, several combinations of T_i and v_rot are used to fit the distribution. In contrast, a simple thermal distribution with a higher ion temperature, T_i= 40 eV, and no plasma rotation (as deduced from TOF_max = 1100.35 ms as displayed above) yields a significantly broader TOF trace.

Zoom In Zoom Out Reset image size

When choosing suitable combinations of T_i and v_rot for such a fitting, consistency with the location of the maximum of g(TOF) is required. Thus, the relationship given above can be reformulated as 1/TOF_max (see equation (8)), where positive or negative values of the rotation velocity can appear. This boundary condition has been used for the selection of the fitting curves shown in figure 6(a): for instance, Rapisarda et al [18] measured toroidal v_rot of −8 km s⁻¹, using C III emissions at the plasma edge of TJ-II. In contrast, in a separate work [19], the same authors measured, using He II emissions, positive v_rot of 5 km s⁻¹ in regions closer to the plasma core, (for TJ-II plasmas heated with ECRH and with similar averaged plasma densities). Here, as seen in figure 4(b), the combination of T_i = 18 eV and v_rot = −13 km s⁻¹ provides a good fitting to the data within the experimental uncertainties. Alternatively, a combination of T_i = 25 eV and v_rot = −10 km s⁻¹ could also be a valid solution. Another way to justify the incorporation of toroidal plasma rotation in the TOF distribution is equation (5). In the absence of rotation or other components in the distribution, a linear fitting in the h(u) plot should be obtained, as seen in figure 5. However, deviations from linearity are expected in the presence of rotation. This is clearly seen in figure 6(b) where different fittings are made with suitable T_i and v_rot combinations. As shown, all possible solutions provide a linear fit with the right slope within a given interval of the abscissa at lower times. However, bending of the h(u) function towards lower values at longer TOF times, as experimentally observed, provides evidence of negative toroidal rotation. Conversely, bending of the curve upward indicates positive rotation, which, unfortunately, is hard to distinguish from the presence of a low T_i tail in the TOF distribution. Similar results are observed for these detectors, with maxima at the times predicted by equation (8). For these measurements, we had only one detector/filter set available to us. Nevertheless, these measurements were not obtained simultaneously, they were obtained on a shot to shot basis. This was moved from one sector to another when performing tests. An upgrade of our technique would consist of employing several detectors, located in the different sectors, to measure the T_i with different flight lengths for a single injection. This would strengthen this technique.

The method proposed here relies on several simplifications which may not be fulfilled in some instances. Perhaps the most obvious of these is the assumption that Li ions, generated by laser injection and subsequent electron impact ionization of the corresponding neutrals, are instantaneously thermalized with background plasma. In a hot magnetic confined plasma, ablated Li atoms are ionized quite rapidly compared to the TOF times of interest here. With a 1st ionization potential of 5.41 eV and an ionization-rate of 9 × 10⁻¹⁴ m³ s⁻¹ for T_e (edge) < 30 eV [20], the ionization time is expected to be of the order t_ion ≈ 6 μs for n_e (edge) < 2 × 10¹⁸ m⁻³ [21]. The resulting Li⁺ ions are subsequently transported along field lines. During this process, ionization to Li²⁺ by electron impact (the ionization potential is 75.64 eV) together with thermalization with the main plasma species takes place. However, the ionization time is significantly longer for Li⁺ ions, i.e. ≈2.5 ms for an ionization-rate of 2 × 10⁻¹⁶ m³ s⁻¹ [20], for the same plasma conditions, this being much longer than the transit times recorded here. Thus, although the technique is based on TOF measurements and the amplitude of the Li⁺ light pulse is secondary (as long as it is above the noise level), the lifetime of the Li⁺ ions sets the range for TOF values (and hence the distance from the laser injection port) of practical use in an experimental device. Thus, the thermalization of cold ions with initial temperature, T_i0, with the plasma background can be described by the expression: T_i (t) = T_i0 [1 −exp(−t/τ_{t herm})] [22] where the characteristic thermalization time, τ_therm is given by the inverse of the thermalization rate, $t_{{\text{eq}}}^{ - 1},{ }$as defined in equation (1). Even when good mass matching between the plasma particles (H, He) and lithium exists, thus fostering a quick thermalization between the injected and plasma species, the thermalization process is not instantaneous and part of the time from the injection to the detection locations is transited at a continuously increasing ion temperature of the, initially cold, Li ions. Although this effect can become minor for large source to measurement separations, it is necessary to consider this effect for locations where the Li II emission is still above detector noise levels, i.e. at short distances. Only estimates of the thermalization time are possible as the impurity content and their ionization state at the plasma edge will strongly accelerate the thermalization process. Thus, for example, for an ion plasma temperature of T_i= 25 eV and an edge electron density n_e = 2–4 × 10¹⁸m⁻³, the thermalization time for Li⁺ in a pure H plasma is ∼100 μs. This time is drastically reduced in the presence of impurities, such as Li²⁺ ions (according to equation (1)) (the ionization potential for Li⁺ is 75.64 eV according to NIST [17] and therefore the contribution of these Li²⁺ ions could be a negligible fraction for the considered radii). So, in any case, the deduced ion temperatures may be considered as a lower limit, and testing the technique at different distances from the source could provide a full picture of the thermalization process, including the background plasma composition.

Finally, it is necessary to consider the initial thermal energy of the injected lithium atoms. It is known that laser ablation generates suprathermal atoms, with kinetic energies that depend on the mass of the ejected atom, as well as other species, (ions, excited neutrals and clusters [23]. For Li, a velocity ∼8 km s⁻¹ was measured for laser blow-off ablation with the same Nd:YAG laser off a LiF layer [24], which will result in a penetration of a few centimeters in the TJ-II plasma edge, although the kinetic velocity may not necessary be the same for LIBS. Aside from limiting the use of the LiLA-TOF diagnostic to ion temperatures above a few eV, a radial spreading of the Li⁺ source can result in a blending of different flux tubes sampled by the detector, and hence, multiple TOF components. This could explain the presence of a low T_i tail in the present experiments, indicating the presence of a T_i gradient near the plasma boundary [7].

The detector set-up here used also imposes some limitations on the accuracy of the measurements. The sampling rate, at 100 kHz, limits reconstruction of the fast signal rise at the onset of the Li⁺ trace. Also, the fact that the injected Li trace has a non-zero-time width implies some convolution of the signal at short times. Both limitations will especially affect the location of the maximum of the TOF trace and fitting of the data to the function given in equation (5) should always be chosen as it provides a more accurate T_i value.

As described above, plasma toroidal rotation has a direct impact on the shape of the TOF trace. The values of v_rot between −10 and −13 km s⁻¹. (However, the edge toroidal rotation values typically found in TJ-II are ⩽−10 km s⁻¹, thus the T_i= 25 eV would appear to be more credible) and associated ion temperatures deduced from figure 6 are in good agreement with previous measurements in TJ-II by RFA (17–21 eV in the SOL) [25]. However, an independent, more accurate estimate of v_rot and T_i would be only possible by simultaneously recording the lithium traces at two toroidal locations, on opposite sides of the injection point whenever available. Work on the required improvements of the diagnostic addressed so far is in progress in TJ-II.

An original technique to measure T_i and v_rot at the periphery of fusion plasmas has been developed. The viability of injecting Li using LIBS and recording the lithium traces of the plasma-generated Li⁺ ions to study ion toroidal, and, consequently, deducing the TOF by means of its propagation has been demonstrated. The preliminary values of T_i and v_rot deduced are consistent with those measured with other diagnostics in TJ-II. The present limitations have been identified and steps oriented at the improvement of the new diagnostic are in progress.

This work was funded by the projects from the Spanish Ministerio de Ciencia e Innovación RTI2018-100835-B-I00 (MCIU/AEI/FEDER, UE) and PID2020-116599RB-I00.

The data that support the findings of this study are available upon reasonable request from the authors.