Correction: Sensitivity improvement in the detection of V and Mn elements in steel using laser-induced breakdown spectroscopy with ring-magnet con ﬁ nement

Correction for ‘Sensitivity improvement in the detection of V and Mn elements in steel using laser-induced breakdown spectroscopy with ring-magnet confinement’ by Zhongqi Hao et al., J. Anal. At. Spectrom., 2014, 29, 2309–2314, DOI: 10.1039/C4JA00144C.


Introduction
][7][8] Currently, one problem with LIBS is its low sensitivity in determining trace elements.Several methods have been proposed to improve LIBS sensitivity, such as double-pulse (DP), 9 multipulse (MP), 10,11 Nanoparticle-enhanced LIBS (NELIBS), 12 spatial confinement, and magnetic confinement.Among these methods, spatial confinement is simple and cost-effective for enhancing signal strength and improving the detection sensitivity of LIBS.
X.Z.Zeng et al. 13 studied the formation of laser-induced plasmas inside a series of truncated cones.The cavity effects on plasma expansion were investigated.X.K. Shen et al. 14 used a cylindrical pipe to confine aluminum plasmas.A maximal enhancement factor of 9 was obtained at a pipe diameter of 10.8 mm.A.M. Popov et al. 15,16 used a small chamber to confine iron 40 and soil plasmas with enhancement factors of 3 and 10 for arsenic (As) and iron (Fe) signals, and the limits of detection (LoDs) of trace elements in soil were 2-5 times better than those detected using free-expanding plasmas.P. Yeates et al. 17 investigated the variation of spectroscopic, imaging, and probe diagnostics of laser plasma plumes within rectangular cavities of fixed depth and varying height.Z. Wang et al. 18,19 adopted a cylindrical cavity of 1.5 mm in height and 3 mm in diameter to confine coal plasmas and reported enhanced emission intensity with reduced shot-to-shot fluctuation.In addition, several previous works in 50 our research group reported enhancement effects 20,21 with accuracy improvement 22 of quantitative analysis using a series of hemispheric cavities to confine laser-induced plasmas.Furthermore, a few studies reported the magnetic field confinement effects in LIBS.S.S. Harilal et al. 23 reported the 55 confinement and dynamics of laser-produced plasma expansion across a 0.64 T magnetic field.The electron density and temperature in LIBS plasmas were observed to increase because of higher collision frequency in the presence of the magnetic field.X.K. Shen et al. 24 used a magnetic field of ~0.8 T to confine 60 plasmas from aluminum (Al) and copper (Cu) targets, with enhancement factors of 2 and 6-8 for Al and Cu lines.N.R. Virendra et al. 25 investigated the enhancement of LIBS signals of solid and liquid samples under a ~0.5 T magnetic field.The enhancement factors could reach 2 and 1.5 for solid and liquid 65 samples, respectively.From the discussion above, we can understand that both spatial and magnetic confinements are effective in LIBS signal enhancement.To combine the advantages of spatial and magnetic confinements, Guo et al. 26 used a pair of permanent magnets and 70 a hemispherical cavity simultaneously to confine plasmas produced from metal samples using a KrF Excimer laser.Enhancement factors of about 22 and 24 were obtained in emission intensity of the cobalt (Co) and chromium (Cr) lines.Nevertheless, this research focused mainly on the enhancement of 75 optical emissions rather than signal-to-noise ratios (SNRs), which represent the sensitivity of LIBS.Additionally, the quantitative analysis and the LoD were not investigated in that work.
The aim of this study was to investigate the improvement of detection sensitivity of vanadium (V) and manganese (Mn) elements in steels using LIBS with ring-magnet confinement, which combines spatial and magnetic confinements of laser-5 induced plasmas.The experimental results confirmed that the optical emission intensity, SNR, and LoD of LIBS all improved with ring-magnet confinement.

Experimental methods
The schematic diagram of the experimental setup used in this The magnetic field intensity (B) distribution of the NdFeB ring magnet that was placed on the steel sample is given in Fig. 2, and the inset shows the structure of the ring magnet used in the experiment.Its outer diameter (D), inner diameter (d), and height 35 (H) were 12.7, 3.2, and 6.3 mm, respectively.The axial line of the magnet and its intersection with the sample surface were defined as the X-axis and origin O (x = 0).As shown in Fig. 2, the magnetic field of the ring magnet (circle dots) increased gradually along the X-axis.The initial magnetic field was 3080 40 G at the origin O; and then the value reached a maximum of about 4090 G at 3.2 mm, corresponding to the half-height position of the ring magnet.The magnetic field then decreased rapidly as the distance increased and reduced to 0 near the exit of the cavity.To confirm the effect of magnetic confinement, the 45 same magnet was degaussed by heat treatment at 500 ℃ for 10 min.The magnetic field of the degaussed magnet (square dots) decreased to ~0 G, so the degaussed magnet only had the spatial confinement effects for plasma plumes.
Seven steel samples (GSB03-2582-2010, Pangang Group 50 Research Institute Co. Ltd, China,) were used in this work.The concentrations of V and Mn are listed in Table 1.

Optical emission enhancement and physical mechanisms
The time-integrated LIBS spectra from sample No. 7 were 60 measured with a time delay of 2 µs after the laser pulse and a gate width of 0.5 µs.The focal point of the lens was 4 mm under the sample surface to avoid air breakdown.The laser energy was 60 mJ/pulse.The spectra presented here were averaged over 20 laser pulses to reduce the intensity deviation.Figure 3 shows the 65 representative spectra of steel sample No. 7 in the range between 402.5 nm and 440 nm with the ring magnet (solid curve), with the degaussed magnet (dashed curve), and without confinement (dotted curve).Obviously, the emission intensities of Mn and V were enhanced with both the magnet and degaussed magnet 70 confinements.Furthermore, the enhancement effect with the ring magnet was greater than that with the degaussed magnet.It can be seen that the intensity enhancement was different with different delay times, from about 1 to 6 µs.During this time period, there were two times the intensity enhancement effects at time delays of 2 µs and 4.5 µs for confinement both with a ring magnet and with a degaussed magnet.Maximum enhancement factors of 3.1 and 2.8 with the ring magnet for V I 437.92 nm and Mn I 403.08 nm were obtained at a time delay of 2 µs, while the enhancement factors were only about 2.7 and 2.3 with the degaussed magnet at the same time delay.
It is clear that stronger enhancements were observed before the ring magnet was degaussed, which indicates that laserinduced plasmas were under both spatial and magnetic confinements.A shock wave is produced by the initial explosive pressure along with the plasma generation and supersonic 20 expansion. 27When the shock wave encounters the cavity wall, it is reflected and travels back towards the plasma.If the cavity is small and the shockwave energy is sufficiently large, multiple reflections can occur, corresponding to multiple enhanced peaks in the emissive intensity.Shen and Popov et al. 14,15 have observed the phenomenon of twice enhanced peaks by spatially confining the laser-induced plasma, and the latter peaks appear at a longer delay times owing to the velocity decaying of the shock wave. 28Similarly, as shown in Fig. 4, the two enhanced peaks appeared at 2 and 4.5 µs.When an external magnetic field is 30 applied on the plasma, the movement of electrons and ions in the plasmas are affected by the Lorentz force.The plasma expansion is decelerated, and the size of the plume is reduced.This may lead to the increase in effective electron density in the plume and 40 result in the enhancement of optical emission.The enhancement of the plasma emission under the influence of a magnetic field can be given as: 25 3 Eq. ( 1) indicates that the intensity enhancement (I 2 /I 1 ) of plasma 45 emission in the presence of a magnetic field depends on the plasma parameter β as well as the ratio of the duration of the emission (t 1/ t 2 ).At the same delay time and gate width, t 1 = t 2 .The plasma parameter β is the ratio of kinetic energy of plasma to the magnetic energy, and it can be expressed as: where n e is the electron density (cm -3 ), k is the Boltzmann constant, T e is the electron temperature (eV), and B is the magnetic field (G).4][25][26] For this reason, no enhancement occurs for initial hot plasma when the delay time is smaller.When the magnet is degaussed, B = 0 and β tends to be infinity, thus I 2 /I 1 = 1, namely the effect of magnetic confinement 60 to plasma disappears, which explains why the intensity enhancement factors became lower after the magnet is degaussed.
To further understand the intensity enhancement effects due to spatial and magnetic confinement, the variations of T e and n e at different delay times were studied.

65
In this work, T e was estimated base on Boltzmann plot using iron atomic lines. 29As shown in Fig. 5, T e is about 12,000 K at delay time of 0.5 µs without any confinement, and then T e started to decrease quickly.Comparably, the values of T e with ring magnet were highest and those with the degaussed magnet were 70 moderate at the delay times from 1 to 6 µs.The reason may due to the magnetic and spatial confinement.
The electron densities with/without confinement were evaluated from width of H α following the formula:  H α -lines, 2 1 α is the half width of the reduced Stark profiles. 31The temporal variations of the n e were shown in Fig. 6.The values of n e without confinement are about 6 × 10 17 cm -3 at delay time of 0.5 µs, then the n e decreased rapidly with the delay times.The n e with the degaussed magnet were higher than the values without 5 confinement at delay times from 1 to 6 µs owing to the spatial confinement effect, and the n e with the ring magnet were highest due to the spatial and magnetic confinement simultaneously.Furthermore, the values of T e , n e and B with the ring magnet were substituted into Eq.( 2) to calculate the values of β.At a 10 certain delay time, the T e and n e are constant, but the B is varied along with X-axis.Therefore, the variation of β with X can be obtained.After that, the values of I 2 /I 1 with X were acquired by substituting β into Eq.( 1).
The variations of β and I 2 /I 1 along the X-axis at different 15 delay times (0.5, 2, 3.5, and 4.5 µs) were shown in Fig. 7 (a) and (b), respectively.As Fig. 7 (a) shows, along with X-axis, the values of β begin to decrease gradually, and then increased sharply.The minimum appeared at X=3.2 mm, corresponding to the position of the maximum magnetic field.Comparing the curves at different delay times, the values of β decreased with the 30 increasing of the delay times due to the decreasing of T e and n e .In Fig. 7 (b), the I 2 /I 1 increase continuously along with Xaxis until reached a peak at X of 3.2 mm, and then the I 2 /I 1 rapidly decreased.The values of I 2 /I 1 >1, that means the magnetic field can cause the emission enhancement.For the curves with 35 decreasing delay times, the I 2 /I 1 became smaller.Therefore, at the delay time of 0.5 µs, the enhancement effect was very weak.

Enhancement of signal-to-noise ratios (SNR) ) ) )
A substantial amount of effort was devoted to increasing the spectral SNR, as larger SNRs result in lower LoDs.The SNR is 40 defined as the ratio of the spectral intensity to the standard deviation of the background.The spectral range from 436.09 to 436.40 nm was selected as the blank background in this work.nonmagnetic ring cavity in improving LIBS sensitivity (SNR).

Quantitative analysis and limit of detection
The discussion above reveals that the enhancement of emission intensity and SNR appeared during the delay times from 1 to 6 µs.Therefore, in the quantitative analyses and LoDs estimation for V 10 and Mn, the delay time and the gate width were fixed to 1 and 5 µs, respectively.The calibration curves of seven reference steel samples for V I 437.92 nm and Mn I 403.08 nm in three cases are shown in Fig. 9.To estimate LoDs of V and Mn in the steels, the well-known 3σ-criterion was used in the following formulae: 15 where B σ is the standard deviation of the background, and S is the slope of the calibration curve.As seen in Fig. 9 (a) and (b), the slopes of the calibration curves with the ring magnet, the degaussed magnet, and no confinment decreased gradually.The 20 LoDs of V were 11, 18, and 41 ppm when a ring magnet, a degaussed ring magnet, and no confinement were used, respectively.The corresponding LoDs of Mn were 30, 41, and 56 ppm, respectively.Obviously, the sensitivity can be further improved using ring-magnet confinement rather than a nonmagnetic ring.In addition, the intercepts of the calibration curves were almost unchanged before and after the magnet was degaussed for both elements, presumably because the magnetic field enhanced the optical emission but had no effect on the bremsstrahlung radiation. 25Hence, the continuous background 35 was almost unchanged in the presence of the magnetic field.
It should be noted that when the LoDs were reduced, the accuracy of the quantitative analyses with the magnetic and nonmagnetic ring confinements showed improvement rather than deterioration.As shown in Fig. 9, the determination coefficients 40 (R 2 ) of the calibration curves for V I 437.92 nm and Mn I 403.08 nm with the ring-magnet confinement were larger than those without confinement.The reason might be that the compressed plasma plumes became more uniform under the spatial and magnetic confinements than those in open air.

Conclusions
In summary, the sensitivity improvement of V and Mn in steels using LIBS combined with ring-magnet confinement was investigated by optical emission intensity, SNR, and LoD.The intensity enhancement effect with a ring magnet is attributed to 50 the simultaneous spatial and magnetic confinement, which can increase the temperature and electron density of the plasma.The SNR enhancement factors of about 2.9 and 2.3 were obtained at a time delay of 2.5 µs for the V I 437.92 nm and Mn I 403.08 nm lines with the ring magnet and 2.4 and 1.7 with the degaussed 55 magnet.The calibration curves of V I 437.92 nm and Mn I 403.08 nm were given with a ring magnet, with a degaussed magnet, and without confinement.Finally, the quantitative analyses curves of V I 437.92 nm and Mn I 403.08 nm elements with/without confinement were established.The LoDs of V and 60 Mn were estimated using 3σ-criterion.The results showed that with ring-magnet confinement, the LoDs of V and Mn in steels decreased to 11 and 30 ppm, lower than the 18 and 41 ppm with a degaussed magnet, and the 41 and 56 ppm with no confinement.

Fig. 2
Fig. 2 Magnetic field intensity before (circle dots) and after (square dots) degaussing along the X-axis of the ring magnet (inset ： magnet 55

Figure 4
Figure4depicts the temporal evolution of the emission intensities from (a) V I (437.92nm) and (b) Mn I (403.08 nm) peaks as functions of the gate delay with the ring magnet, with the degaussed magnet, and without confinement.The error bars may come from the pulse-to-pulse energy fluctuation and the deviation of the magnet position among different measurements.It can be seen that the intensity enhancement was different with different delay times, from about 1 to 6 µs.During this time period, there were two times the intensity enhancement effects at time delays of 2 µs and 4.5 µs for confinement both with a ring

Fig. 4
Fig. 4 Emission intensities of V I 437.92 nm (a) and Mn I 403.08 nm (b) at different delay times with a ring magnet (circle dots), with a degaussed magnet (triangle dots), and without confinement (square dots).Gate width: 0.5 µs.

Fig. 5
Fig. 5 Variation of the electron temperature at different delay times with a ring magnet (circle dots), with a degaussed magnet (triangle dots), and without confinement (square dots).Gate width: 0.5 µs.

Fig. 6 Fig. 7
Fig. 6 Variation of the electron density at different delay times with a ring magnet (circle dots), with a degaussed magnet (triangle dots), and20

Figure 8
illustrates the SNRs of V I 437.92 nm (a) and Mn I 403.08 nm (b) with a ring magnet, with a degaussed magnet, and 45 without confinement, respectively.As shown in Fig.8, the SNR reached maximum at a time delay of 2.5 µs but not 2 µs, corresponding to the maximum intensity enhancement factors in Fig.4(a) and (b).This is because the decay of the noise is faster than the signal under confinement.At the beginning, n e and T e 50 increase sharply because of the compression of the reflected shock wave.The noise is enhanced during this fast variation of plasma conditions.After the energy of the shock wave is consumed, the noise decreases quickly due to the compressed plasma expanding relatively smoothly, but the optical emission of 55 excited atomic decays slowly with the decrease in n e and T e in expanding plasma plumes.

Fig. 8
Fig. 8 Signal-to-noise ratios of V I 437.92 nm (a) and Mn I 403.08 nm (b) 60

Fig. 9
Fig. 9 Calibration curves of (a) V and (b) Mn in the referenced steel samples with a ring magnet (circle dots), with a degaussed magnet (triangle dots), and without confinement (square dots). 45

Table 1 .
Compositions of V and Mn in Steel Samples