1. Introduction

Transparent conductive oxides (TCOs) are appealing materials for a vast suite of optoelectronic applications [1–3]. In recent years, TCOs have also emerged as increasingly favorable materials for plasmonic and active photonic devices because of their unique epsilon-near-zero (ENZ) effect in the near-infrared (NIR) [4]. Specifically, TCOs exhibit strong optical nonlinearities, such as extreme nonlinear dispersion and enhanced light-matter interactions [5]. From these properties arise potential applications of TCOs as epsilon-near-zero materials in optical communications and optical data processing [6]. An important additional asset of TCOs is that their optoelectronic properties can be modulated by engineering their defects, often through a post-growth annealing step [7–10]. This can allow for modulation of the carrier concentration, plasma energy, and thus the ENZ region, further into the mid-infrared (MIR) to target the chemical and biological sensing windows [11–14]. Control and optimization of this adjustment is, however, a complex design problem. We seek to address this challenge by establishing a novel high-throughput method of ‘reactive’ laser annealing (ReLA) to selectively tune the carrier concentration of tin-doped indium oxide (ITO) thin films, and thus the plasma energy and ENZ domain. In this way, one can ensure the lowest electronic losses and, subsequently, more desirable “figures of merit” (FoMs) for applications within the desired frequency window of operation [15–20].

The most common techniques for annealing of TCOs are indirect annealing techniques such as conventional furnace heating and rapid thermal annealing [21–23]. However, in recent decades the remarkable capabilities of more “direct” annealing techniques, such as laser annealing (LA), to improve the optoelectronic properties of TCOs have become increasingly apparent [24]. Vitally, the use of high energy lasers, such as excimer lasers, allows for an ultra-fast (few $\textrm{ns}$) annealing process that can match the high thermal energy demand of metal-oxides, while the thermal dose is spatially localized, allowing for micro-patterning of the annealed areas as well as the use of heat sensitive substrates [24]. The use of reactive environments during LA has recently been revealed to induce a purely compositionally driven modulation of the optoelectronic properties of ITO thin films, where laser processing within reducing and oxidizing environments of $5{\%}$H₂ in N₂ and $100{\%}$ O₂ promoted an ambient-dependent enhancement of the carrier concentration and mobility, tuning the ENZ region closer to the telecom wavelength (from $1.91\;\mathrm{\mu}\textrm{m}$ to $1.73\;\mathrm{\mu}\textrm{m}$) [25,26]. To expand the operational window of ITO as a “low-loss” plasmonic and/or photonic material component across the mid-infrared (MIR) spectral window [2,10,13], a strategy to reduce the carrier concentration of ITO via ReLA was noted to be of upmost importance [25,26].

In this work, we follow up the foundational demonstration of ReLA [25,26] by exploring the alternative strategy of starting from an “un-optimized” seed, room temperature, sputtered ITO film, purposely grown to possess a high resistivity $({\sim} 7 \times {10^{ - 3}}\;\mathrm{\Omega}\textrm{cm}$) and low carrier concentration (${\sim} 6 \times {10^{19}} \;\textrm{c}{\textrm{m}^{ - 3}}$). We give special focus onto how the starting film conditions affect the ReLA process by comparing the results to those of our previous work [25]. To elucidate the precise mechanisms behind ReLA, the modifications of the optical and electronic properties were related to changes in the structural and compositional properties. By carefully considering the interplay between the compositional alterations and crystal structure changes, we revealed the influence of both the intra-grain and inter-grain carrier transport properties on the optoelectronic properties of ITO. Depth-dependent, combined structural and compositional changes were uncovered via a combination of advanced ellipsometric modelling, X-ray diffractometry, X-ray photoelectron spectroscopy, and cross-sectional transmission electron microscopy. The upper portion of the processed film was subject to a significant enhancement (${\sim}{\times} 10$) of the carrier concentration that could be tuned by controlling both the laser fluence and the oxygen partial pressure of the reactive environment.

2. Experimental details

2.1 Thin film deposition

${\sim} 130 \;\textrm{nm}$ ITO thin films were deposited onto double side polished n-type ($1 - 10\;\mathrm{\Omega}\textrm{cm}$), $4\mathrm{^{\prime\prime}}$, ($100$) oriented Si wafers with a $2\;\textrm{nm}$ native oxide using radio frequency magnetron sputtering. The seed ITO thin film was deposited with a set of sputtering conditions (power and pressure of $20\;\textrm{W}$ and $5\;\textrm{mTorr}$., respectively) that were found to produce a lower quality material with resistivity being a magnitude greater than the film produced with optimized sputtering conditions [26] (investigation of the sputtering parameters can be found in Supplement 1, section 1) [25]. The base pressure was ${\sim} {10^{ - 5}}\;\textrm{Pa}$ $({\sim} {10^{ - 7}}\textrm{mbar}$). Intentional substrate heating or bias was not applied to the substrate. We used a $3\mathrm{^{\prime\prime}}$ target of $10\;\textrm{wt}.{\%}$ Sn:In₂O₃ ($99.99{\%}$ purity). Ar and O₂ we introduced to the chamber with an O₂ concentration of $0.25 \;\pm\; 0.02{\%}$. The substrate to target distance was $10.8 \;\pm \;0.2\;\textrm{cm}$. Further details on the deposition process can be found elsewhere [27]. Following deposition, the seed material was diced into ${\sim} ({8 \times 8} )\;\textrm{m}{\textrm{m}^2}$ “die”.

2.2 Reactive laser annealing

Laser processing was performed at room temperature with a KrF ($248\;\textrm{nm}$) excimer laser (LAMBDA PHYSIK LPX 305i), which delivers unpolarized light at a 1–50 Hz pulse repetition rate and ${\sim} 25\;\textrm{ns}$ pulse length to the sample surface. Further details of the LA system are published elsewhere [27]. A pressure cell with a UV-transparent window was used to perform the laser processing within pressurized ($6.89 \times {10^5}\;\textrm{Pa}$ or $100\;\textrm{psig}$) environments of $5{\%}$ H₂ in N₂ and $100{\%}$ O₂. For each ambient composition, a single laser pulse was delivered to the sample with a fluence, ${\textrm{J}_\textrm{L}},$ of 25–125 mJcm^-2 in steps of $25\;\textrm{mJc}{\textrm{m}^{ - 2}}$. The beam delivery path can produce a minimum spot size of $\textrm{}\sim 3 \times 3\;\mathrm{\mu }{\textrm{m}^2}$, although in this work a spot size of $({1.20 \;\times\; 1.20} )\;\textrm{c}{\textrm{m}^2}$ was used. We note that to achieve a spot size of $> 1\;\textrm{c}{\textrm{m}^2}$ (large enough to enable reliable characterization), $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ was the largest achievable fluence. The fact that the laser spot size was larger than the die also ensured that the entire area of the die was uniformly laser processed.

2.3 Thin film characterization

Optical characterization via the measurement of the ellipsometric angles $\mathrm{\Psi }(\textrm{E} )$ and $\mathrm{\Delta }(\textrm{E} )$ was performed using a J. A. Woollam Mark II IR-SE in the spectral range of 0.034–0.8 eV ($1.455 - 36.45\;\mathrm{\mu}\textrm{m}$) and a J. A. Woollam M2000 SE in the spectral range of 0.74–3.34 eV (0.37–1.67$\mathrm{\mu}\textrm{m}$) at incident angles of 65°, 70°, and 75°. The J. A. Woollam Mark II IR-SE and a normal-incidence optical reflectance probe were also utilized to measure the IR transmittance, ${\textrm{T}_{\textrm{IR}}}(\textrm{E} )\textrm{}$, and visible reflectance, ${\textrm{R}_{\textrm{Vis}}}(\textrm{E} )$, respectively. Using the known sample geometry (ITO / $2.63\;\textrm{nm}$-SiO₂ / Si), an optical model comprising the complex permittivity, ${\tilde{\mathrm{\varepsilon }}}(\textrm{E} )$, of each layer was fit to the full suite of optical measurements ($\mathrm{\Psi }(\textrm{E} )$, $\mathrm{\Delta }(\textrm{E} )$, ${\textrm{T}_{\textrm{IR}}}(\textrm{E} )$, and ${\textrm{R}_{\textrm{Vis}}}(\textrm{E} )$). By fitting all measurements simultaneously, we reduce the correlation between the fitting parameters to improve the confidence in the uniqueness and physical reality of the extracted values (see Supplement 1, section 3). From the fitting process, we extract the geometric features (film thickness, surface roughness and uniformity) and optoelectronic properties (interband transitions, phonon modes and/or defect states absorption, free carriers etc.) of the ITO films [28]. Electrical characterization was performed via four-point probe (4pp) in both collinear and Van-der-Pauw configuration (${\mathrm{\rho }_{4\textrm{pp}}}$ and ${\mathrm{\rho }_{\textrm{Hall}}}$). Characterization of the “Hall” carrier concentration, ${\textrm{N}_{\textrm{Hall}}}$, was performed with an Ecopia HMS-3000 Hall Measurement System ($\textrm{B} = 0.553\textrm{T}$) at room temperature.

Structural characterization was performed by XRD employing a PanAnalytical Pro Diffractometer. The spot size was set to $5 \times 5\;\textrm{mm}$ using a $10\;\textrm{mm}$ height-limiting slit, ${2^{\textrm{rad}}}$ soller slits, and a programmable divergence slit. The scan step size was $0.0083556^\circ $ with a time per step of $1998.345\textrm\;{s}$ in a range of $25^\circ \textrm{} - \textrm{}40^\circ $. The long step time was required due to the small sample footprint and the amorphous nature of the as-grown ITO films. The range was chosen to cover the most clearly identifiable ($222$) and ($400$) Bragg peaks for bixbyite In₂O₃, at $30.6^\circ $ and $35.5^\circ $, respectively [29]. Cross-sectional images of an indicative set of ITO thin films were obtained via high resolution TEM (FEI Talos S/TEM) using a Schottky-type field emission gun operated with an electrostatic potential of $200\;\textrm{kV}$. Prior to TEM, the films were subjected to focused ion beam milling process using a JEOL 4500 FIB/SEM. Surface elemental characterization was performed with a Kratos Analytical Ltd. AXIS Ultra XPS system using a monochromated Al-K_α1 X-ray beam excitation source ($1486.6\; \mathrm {eV}$) in an ultra-high vacuum chamber (base pressure ${\sim} {10^{ - 9}}\;\textrm{bar}$). The relative elemental abundance across the entire depth of the ITO thin films was determined from EDX measurements with the FEI Talos S/TEM.

3. Spectroscopic Investigation of the seed and laser processed films

3.1. Optical modelling

To reveal the optoelectronic properties of the seed ITO films and those subjected to ReLA we firstly look at the optoelectronic properties, which define the key characteristics of TCOs that are associated with most applications [1,30]. Figure 1(a) presents $\mathrm{\Psi }(\textrm{E} )$ (blue squares) and $\mathrm{\Delta }(\textrm{E} )$. (red squares) of an indicative seed ITO thin film (see Supplement 1, section 2 for optical measurements of all samples). The solid green lines show the best fit across the entire spectral range. We note the more complicated line-shape of $\mathrm{\Psi }(\textrm{E} )$. and $\mathrm{\Delta }(\textrm{E} )$. as compared to the higher-quality (lower resistivity) ITO film deposited with optimized deposition conditions (Supplement 1, Fig. S3) [25]. To fit $\mathrm{\Psi }(\textrm{E} )$ and $\mathrm{\Delta }(\textrm{E} )$. of the seed and laser processed ITO films we need to consider a more-complex inhogeneity than has been previously reported for high-quality ITO films [25,31–33]. It is vital to note that by appending $\textrm{}{\textrm{T}_{\textrm{IR}}}(\textrm{E} )$ to $\mathrm{\Psi }(\textrm{E} )$ and $\mathrm{\Delta }(\textrm{E} )$ during the fit, we were ato exclude the presence of additional IR oscillators to explain the complex line shape of $\mathrm{\Psi }(\textrm{E} )$ and $\mathrm{\Delta }(\textrm{E} )$. A detailed description of this process and a justification of the utilized model is presented in Supplement 1, section 3. Briefly, the ITO film is separated into two distinct layers: a “top” and “bottomlayer, where the top layer comprises most of the film. Each layer is described by a complex permittivity, ${\tilde{\mathrm{\varepsilon}}}(\textrm{E} )$. The imaginary part of the complex permittivity, ${\mathrm{\varepsilon }_2}(\textrm{E} )$, is described by a summation of multiple oscillators (Eq. (1)) and the real part of the complex permittivity, ${\mathrm{\varepsilon }_1}(\textrm{E} )$, is derived from a Kramers-Kronig transformation of ${\mathrm{\varepsilon }_2}(\textrm{E} )$ with a background contribution to the permittivity, ${\mathrm{\varepsilon }_\infty }$ [31,34].

 

Fig. 1. (a) Measured $\mathrm{\Psi }(\textrm{E} )$ (blue squares) and $\mathrm{\Delta }(\textrm{E} )$ (red squares) the seed ITO thin film alongside the corresponding fit (green lines). (b-c) The fitted imaginary permittivity for the (b) top and (c) bottom layers of the seed ITO thin film (solid red line) with the individual components of the imaginary permittivity for the 1^st (dashed orange line) and 2^nd (dashed green line) free carrier species, phonon (dashed cyan line), potential defect state (dashed pink line), and interband transition (dashed purple line). Logarithmic scales in the x- and y-axes are used in (a-c) to exaggerate the IR spectral region. (d) Schematic of the multi-layer geometric and parametrized optoelectronic model.

Download Full Size | PDF

Figure 1(b) presents ${\mathrm{\varepsilon }_2}$ (solid red line), and the individual components which make up ${\mathrm{\varepsilon }_2}(\textrm{E} )$ (dashed lines), for the top layer of the seed ITO film. The optical constants for all seed and laser processed films (alongside tabulated lists of all the corresponding fitting parameters) can be found in Supplement 1, section 2. The two IR Gaussian peaks are shown as the cyan and pink dashed lines. The multiple phonon peaks of ITO between $0.03\;\textrm{eV}$ and $0.07\;\textrm{eV}$ [36] cannot be individually resolved and so are represented here by a single Gaussian oscillator (dashed cyan line). The 2^nd Gaussian oscillator, at ${\sim} 0.15\;\textrm{eV}$ (dashed pink line), is either a phonon peak due to hydroxides [37] or a potential defect state [25]. The purple dashed line shows the tail of the UV absorption, described by a Tauc-Lorentz oscillator (dashed purple line), centered at $3.93\;\textrm{eV}$ [37]. The most significant contribution to the optical absorption in the IR comes from the free carriers. For detailed electrodynamic analysis, the free carrier absorption is described here with two Drude terms (2^nd term in Eq. (1); dashed orange and green lines in Fig. 1(b). The requirement of two Drude terms to achieve a good fit suggests that, in the top layer, there exist two “carrier species” with independent ${\textrm{N}_{\textrm{opt},\textrm{s}}}$, ${\mathrm{\mu }_{\textrm{opt},\textrm{s}}}$ and $\textrm{m}_{\textrm{e},\textrm{s}}^{\ast }$ (i.e., ${\textrm{N}_{\textrm{opt},\textrm{t}1}}$, ${\textrm{N}_{\textrm{opt},\textrm{t}2}}$, ${\mathrm{\mu }_{\textrm{opt},\textrm{t}1}}$, ${\mathrm{\mu }_{\textrm{opt},\textrm{t}2}}$, $\textrm{m}_{\textrm{e},\textrm{t}1}^{\ast }$, and $\textrm{m}_{\textrm{e},\textrm{t}2}^{\ast }$; where the subscript $\textrm{t}$ denotes that the species resides within the top layer). The two carrier species reflect different routes for carrier transport within the ITO layer, the mechanisms for which will be discussed below.

Figure 1(c) presents ${\mathrm{\varepsilon }_2}(\textrm{E} )$ for the bottom layer of the seed ITO film (solid red line). The bottom ITO layer is described using a gradient between the carrier concentration at the ITO/SiO₂ interface, ${\textrm{N}_{\textrm{opt},\textrm{b}}}$, and within the top layer, ${\textrm{N}_{\textrm{opt},\textrm{t}}}$. The overall steepness of the gradient is defined by the layer height, ${\textrm{d}_\textrm{b}}$. The shape approximates an exponential gradient [38]. This layer comprises $20$ individual sub-layers of discrete ${\textrm{N}_{\textrm{opt},\textrm{z}}}$, ${\mathrm{\mu }_{\textrm{opt},\textrm{z}}}$ $\textrm{m}_{\textrm{e},\textrm{z}}^{\ast }$ and ${\textrm{T}_{\textrm{G},\textrm{z}}}$, where the subscript $\textrm{z}$ denotes that the parameter is defined at a distance ${\textrm{d}_\textrm{B}}\textrm{z}/20$ from the ITO/SiO₂ interface. We formulate the Drude terms within Eq. (1) using ${\textrm{N}_{\textrm{opt},\textrm{s}}}$ and ${\mathrm{\mu }_{\textrm{opt},\textrm{s}}}$ (here forth referred to as “the transport properties”) as the gradient is more physically represented by variations in the transport properties than in the optical resistivity, ${\mathrm{\rho }_{\textrm{opt}}}$, and mean time between collisions, ${\mathrm{\tau }_{\textrm{opt}}}$. This also enables us to consider the non-parabolicity of the conduction band due to free carrier population [39] to describe $\textrm{m}_{\textrm{e},\textrm{n}}^{\ast }$ as a function of ${\textrm{N}_{\textrm{opt},\textrm{n}}}$ throughout the sample depth. We also allow the Tauc-Gap, ${\textrm{T}_\textrm{G}}$, [34] to vary with depth, reflecting the potential Burstein-Moss shift with changing carrier concentration [40]. This improves the fit in the visible range. All other parameters are left free during the fit but do not vary with depth (i.e., they are identical for the top and bottom layer). Finally, we follow the suggestions of previous reports [31,33] and include a layer of surface roughness described by an effective medium approximation of $50{\%}$ of the top layer and $50{\%}$ air. Altogether, these steps ensure that the optical model accurately reflects reality.

As discussed, there exist three independent carrier species that define the optical carrier transport properties across the sample depth of the ITO film. This inevitably makes it difficult to extract useful information about how the optoelectronic properties of the low-$\textrm{N}$ seed ITO films are affected by ReLA. To aid in this matter, it is helpful to note that $\textrm{N}$ and $\mathrm{\mu }$ are linked through the influence of scattering mechanisms. The precise nature of this interdependence of the transport properties has been described for the case of ITO [41,42]. The formulation of the “scattering equation” that defines $\mathrm{\mu }(\textrm{N} )$ by considering each scattering mechanism in turn, is detailed elsewhere [32]. Therefore, to see the full picture of the transport properties of each seed and laser processed film, we compare the transport properties for the 1^st and 2^nd top-layer species in Fig. 2. The case for the bottom layer and a comparison of optical and electrical results is presented within Supplement 1, section 3 (Fig. S21). Due to the complexity introduced by the existence of multiple species of carriers across the film, the analysis requires careful, step-by-step, examination before summarizing the overall picture. Firstly, we examine the seed films (and those annealed at “low” fluences) to elucidate the identity of the two carrier species in the top layer. Then we examine how ReLA at higher fluences affects each layer and carrier species. Finally, we compare the results with those for the bottom layer and Hall Effect values of the transport properties.

 

Fig. 2. Optical, “top layer”, carrier concentration, ${\textrm{N}_{\textrm{opt},\textrm{ts}}}$, and mobility, ${\mathrm{\mu }_{\textrm{opt},\textrm{ts}}}$, of the seed ITO thin films (grey stars) and those subject to single-pulse ReLA at $25\textrm{} - \textrm{}125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in $5{\%}$ H₂ in N₂ (red squares) and $100{\%}$ O₂ (green triangles). The 1^st and 2^nd carrier species are denoted by a dark and light symbol color, respectively. For easy comparison, we also present the corresponding transport mechanisms for seed and laser processed ITO films that were deposited with optimized deposition conditions, taken from [25]. The dashed lines indicate the theoretical trends of $\mathrm{\mu }(\textrm{N} )$ from various scattering mechanisms. The parameters of $\mathrm{\mu }{(\textrm{N} )_{\textrm{IIS} + \textrm{GBS}}}\textrm{}$ that are related to grain boundary scattering (grain size, $\textrm{L}$, and trap density, ${\textrm{Q}_\textrm{t}}$) are fit to the 1^st carrier species within the seed materials and those subject to ReLA at $\le 100\;\textrm{mJc}{\textrm{m}^{ - 2}}$ (gold dashed line) and at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ (cyan dashed line). The corresponding $\mathrm{\mu }{(\textrm{N} )_{\textrm{IIS}}}$ relation, which ignores the effect of grain boundary scattering, for the films annealed at $100\;\textrm{mJc}{\textrm{m}^{ - 2}}$ and $12\;\textrm{mJc}{\textrm{m}^{ - 2}}$ are indicated by the purple and pink dashed lines, respectively. The green and red arrows indicate ambient dependent transitions in carrier concentration after ReLA in $5{\%}$ H₂ in N₂ and $100{\%}$ O₂, respectively.

Download Full Size | PDF

3.2. Dual carrier species of ITO

Figure 2 presents ${\mathrm{\mu }_{\textrm{opt}}}$ against ${\textrm{N}_{\textrm{opt}}}$ for the 1^st (darker shaded symbols) and 2^nd (lighter shaded symbols) carrier species that exist within the top layer of the seed ITO films (grey stars) and those subject to single-pulse ReLA at 25–125 mJcm^-2 in $5{\%}$ H₂ in N₂ (red squares) and $100{\%}$ O₂ (green triangles). The labels indicate the fluence used during laser processing and are colored red or green for annealing in $5{\%}$ H₂ in N₂ or $100{\%}$ O₂, respectively. For comparison, the case for seed and laser processed ITO films that were deposited with optimized deposition conditions [25] are presented with the open symbols.

We indicate multiple theoretical trends of $\mathrm{\mu }(\textrm{N} )$ with the dashed lines, where the parameters defining the trend are fit to particular “clusters” of data. Subsequently, it is found that each cluster relates to a set of samples where the scattering mechanisms that influence ${\mathrm{\mu }_{\textrm{opt}}}$ in relation to ${\textrm{N}_{\textrm{opt}}}$ are the same. The fits of $\mathrm{\mu }(\textrm{N} )$ to each cluster are indicated with the purple, orange, pink, and blue dashed lines. The first cluster (indicated by the gold shaded area) represents the 1^st carrier species within the top layers of the seed ITO films and those subject to ReLA at fluences of $25 - 100\;\textrm{mJc}{\textrm{m}^{ - 2}}$. We immediately notice the significantly lower ${\textrm{N}_{\textrm{opt}}}$ in comparison to previous reports for “optimized” (lower resistivity) ITO films [25,31–33], which is the primary source of the larger ${\mathrm{\rho }_{\textrm{opt}}}$ for the ITO films of this work. For this cluster of samples, we observe that ${\mathrm{\mu }_{\textrm{opt}}}$ increases with ${\textrm{N}_{\textrm{opt}}}$. This indicates that, in this case, grain boundary scattering (GBS) is the dominant scattering mechanism [42]. We indicate the trend of $\mathrm{\mu }(\textrm{N} )$ fit to the 1^st cluster with the dashed gold line. Within the small range of $\textrm{N}$ covered by this cluster, only the parameters relating to the effect of GBS and ionized impurity scattering (IIS) matter (due to Matthiessen’s rule). Therefore, only the grain size, $\textrm{L}$, and trap density, ${\textrm{Q}_\textrm{t}}$, were left free during the fit of the scattering equation [41,42] to the data, and were determined to be $2.0 \;\pm\; 0.2\;\textrm{nm}$ and $({7.0 \pm 0.3} )\times {10^{12}}\;\textrm{c}{\textrm{m}^{ - 2}}$, respectively. The fine grain size indicates that the film is nanocrystalline/amorphous. The other parameters were fixed to previously reported values [42] or (for the IIS-diminished mobility, ${\mathrm{\mu }_{\textrm{IIS}}}$, and the coefficient of ionized cluster scattering (ICS), ${\mathrm{\alpha }_{\textrm{ICS}}}$) to those fit to previous results for a set of optimized ITO films [32].

The second cluster (purple shaded area in Fig. 2) represents the 2^nd carrier species (lighter shaded symbols) within the top layer of the seed ITO films and those subject to ReLA at $25 - 100\;\textrm{mJc}{\textrm{m}^{ - 2}}$. The second cluster has considerably (approximately a factor of $10$) lower ${\textrm{N}_{\textrm{opt}}}$ and larger ${\mathrm{\mu }_{\textrm{opt}}}$ than the first cluster. There is a negative trend of $\mathrm{\mu }(\textrm{N} )$ that is equivalent to the trend of $\mathrm{\mu }(\textrm{N} )$ for the first species, if only the influence of GBS is ignored (i.e., $\mathrm{\mu }(\textrm{N} )$ is not fit to this cluster but arises simply by removing the contribution of GBS to the scattering equation). We show this with the dashed purple line in Fig. 2. Indeed, the values of $\mathrm{\mu }(\textrm{N} )$ for the 2^nd species agree well with those for epitaxial films of In₂O₃ and ITO [43–45]. From this observation, we can formulate an understanding of the physical mechanism behind the existence of two charge species within the top layer. Specifically, the first species experiences the influence of GBS whereas the 2^nd species does not. Therefore, the 1^st species can be explained as the free carriers which oscillate in response to the driving electromagnetic field across grain boundaries. Conversely, the 2^nd species oscillates within the individual grains. To further validate this claim, we consider that due to the nanocrystalline nature of the room-temperature deposited ITO films, the grains can be smaller than the mean free path, ${\textrm{l}_{\textrm{e},\textrm{IIS}}}(\textrm{N} )$, that is calculated by considering the effect of IIS using the degenerate electron gas model [46,47],

This gives values of ${\textrm{l}_{\textrm{e},\textrm{IIS}}}(\textrm{N} )= 3.9 - 5.5\;\textrm{nm}$ for $\textrm{N} = {10^{18}} - 2.7 \times {10^{19}}\;\textrm{c}{\textrm{m}^{ - 3}}$ (the $\textrm{N}$-range of the 2^nd cluster). Vitally, these values for ${\textrm{l}_{\textrm{e},\textrm{IIS}}}(\textrm{N} )$ are larger than $\textrm{L}$, suggesting these carriers should not scatter within the grain. It should, however, be noted that the scattering events which give rise to ${\textrm{l}_{\textrm{e},\textrm{IIS}}}(\textrm{N} )$ are, inherently, statistical phenomena. Even if ${\textrm{l}_{\textrm{e},\textrm{IIS}}} > \textrm{L}$, a fraction of the carriers will be scattered off ionized impurities before the grain boundaries and another fraction will be scattered within the grain boundaries. The “transmitted” fraction of carriers, ${\textrm{T}_{\textrm{N},\textrm{IIS}}}$, (i.e., those that make it to the grain boundaries) can be described by:

3.3. Effect of ReLA

The third and fourth clusters (light blue and pink shaded areas in Fig. 2, respectively) represent transitions to different “branches” of $\mathrm{\mu }(\textrm{N} )$ brought on by ReLA at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$. The population density for the intra-grain carriers is greatly increased, into the order of the inter-grain carriers. This indicates that the number of high mobility conduction pathways have increased, possibly by breaking up the neutral clusters of (SnO₂)₂. We fit $\mathrm{\mu }(\textrm{N} )$ to the 3^rd and 4^th clusters simultaneously with $\mathrm{\mu }(\textrm{N} )$ considering either only IIS (dashed pink line in Fig. 2) or IIS + GBS (dashed cyan line in Fig. 2). To do so, ${\mathrm{\mu }_{\textrm{IIS}}}$ and ${\textrm{Q}_\textrm{t}}$ are required to be free parameters. Due to the limited sample set, we are required to fix $\textrm{L}$ at $6\textrm{nm}$ (determined by XRD; see Fig. 3) to avoid over-correlation of the free parameters. ${\mathrm{\mu }_{\textrm{IIS}}}$ and ${\textrm{Q}_\textrm{t}}$ are subsequently determined to be $10 \pm 4\;\textrm{c}{\textrm{m}^2}/\textrm{Vs}$, and $({18.2 \pm 0.3} )\times {10^{12}}\;\textrm{c}{\textrm{m}^{ - 2}}$, respectively. This represents a decrease in the IIS/lattice limited mobility and an increase in both the grain size and trap density after laser processing at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$. This is indicative of a transition of the structure to one that contains more distinct and crystallized granular phases, with higher grain boundary potentials, dispersed within the amorphous phase [54]. The lower value of ${\mathrm{\mu }_{\textrm{IIS}}}$ also infers an increased concentration of ionized impurities (activated Sn or oxygen vacancies; to be explored further below). For both the 1^st and 2^nd carrier species, the samples annealed in $5{\%}$ H₂ in N₂ tend to show larger values of ${\textrm{N}_{\textrm{opt}}}$ than for those annealed in $100{\%}$ O₂. Indeed, we note a decrease in ${\textrm{N}_{\textrm{opt}}}$ with increasing fluence for ReLA in $100{\%}$ O₂. We indicate this with the red and green arrows in Fig. 2.

 

Fig. 3. X-ray diffractograms for the seed ITO film (grey squares) and those subject to ReLA in $5{\%}$ H₂ in N_­2 (red squares) and $100{\%}$ O₂ (green squares) at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$. The solid black, red, and green lines represent the corresponding fit of two pseudo-Voigt functions to the data. For the seed film, we present the component peaks, representing regions of the film under tensile and compressive stress, as the gold and cyan dashed lines respectively. For the laser processed films, we present the additional “unstressed” ($222$) and ($400$) Bragg peaks as the pink and purple dotted lines, respectively. Also shown are the TEM cross-sectional images of (b) the seed ITO film and those subject to ReLA at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in (c) $100{\%}$ O₂ and (d) $5{\%}$ H­­₂ in N_2.

Download Full Size | PDF

To the best of our knowledge, this is the first time that ReLA in oxidizing environments can reduce the carrier concentration of ITO. This has profound implications for the field of plasmonics, as ReLA can be used to tune the plasma energy of ITO further into the IR, increasing the operational window for “low-loss” plasmonics [13,55,56]. An indicative set of ‘figures of merit’ (FoMs) (for surface plasmon resonance (SPR) [16], localized SPR (LSPR) [16], near-field enhancement [15], and photothermal conversion [15]) are given in Fig. S16 in Supplement 1, section 2 and demonstrates how ReLA can both enhance the peak values of each FoM and expand the range of negative permittivity where plasmonic effects can be exploited. Finally, examination of the 1^st carrier species within the film annealed at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ reveals, again, an ambient dependence to the increase in ${\textrm{N}_{\textrm{opt}}}$, where there is greater enhancement for the reducing environment. For the bottom layer, ${\textrm{N}_{\textrm{opt}}}$ is significantly higher than in the top layer and is largely unaffected by ReLA due to the depth-dependence of the thermal treatment (see Supplement 1, section 3).

In summary, the optoelectronic properties of the studied seed ITO film grown with “un-optimized” sputtering conditions are far more complex than observed during previous reports of ReLA of high-quality ITO films [25]. The seed film has an inherent inhomogeneity (described by a gradient in the carrier transport properties) and multiple carrier species (due to “inter-grain” and “intra-grain” carrier species). To the best of our knowledge, this is the first observation of a dual carrier species present in ITO. Only by fitting a complex optical model to the ellipsometric measurements could we accurately extract the carrier transport properties. Furthermore, by investigating the resulting properties through the lens of the various scattering mechanisms at play could the electrical behavior throughout the sample, and the influence of ReLA, be fully elucidated. We revealed the nature of the photo-induced modifications of the optoelectronic properties of the “un-optimized” ITO film, with similar effects as found for the optimized film [25]. We find an ambient-dependent modulation of $\textrm{N}$ at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$, increased for more reducing environments, and, vitally, find that ReLA in $100{\%}$ O₂ can reduce the carrier concentration. However, the carrier mobility is reduced. Using previous studies as a guide [25], we have been able to infer some of the underlying physical mechanisms behind the conversion of the optoelectronic properties during ReLA. However, due to the complex nature of the film structure further experimental effort is required to fully elucidate the precise nature of such photo-induced modifications.

4. Structural ReLA-induced modifications

To investigate whether ReLA has induced any modifications to the film crystallinity which may elucidate the changes to the optoelectronic properties observed in Fig. 2. The crystal structure of the seed and processed ITO films was studied via XRD. Figure 3(a) presents the X-ray diffractograms for the seed film (grey squares) and the corresponding films processed with a single laser pulse at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in $5{\%}$ H₂ in N₂ (red squares) and in $100{\%}$ O₂ (green squares). The X-ray diffractograms for all samples can be found in Supplement 1, section 4. The seed film shows a broad, asymmetric peak at centered at ${\sim} 32.5^\circ $. We indicate the ($222$), ($321$), and ($400$) peak positions for crystalline, unstressed In₂O₃ at $30.607^\circ $, $33.127^\circ $, and $35.49^\circ $, respectively [9,29], with the dashed brown and purple lines, respectively.

The diffraction peak is fit with two pseudo-Voigt functions that represent each crystalline orientation, which we indicate these with the gold and cyan dotted lines, respectively. For the films annealed at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$, two new peaks emerge from the broad asymmetric peak. This indicates that the film now has a semi-crystalline nature [57]. The positions of these peaks more accurately reflect those for an unstressed, crystalline ITO film. Therefore, we note that ReLA has induced crystallization of a portion of the ITO film, and we present the pseudo-Voigt peaks that represent the crystallized ($222$) and ($400$) peaks with the pink and purple dotted lines, respectively, in Fig. 3(a). This is confirmed by examining Fig. 3(b)-(d), where we present the TEM cross-sectional images for the seed ITO film and those subject to ReLA at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in (c) $100{\%}$ O₂ and (d) $5{\%}$ H₂ in N₂. The seed film is highly uniform across the sample depth and no crystallization can be observed. However, for the film annealed at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in $100{\%}$ O₂, a clear region of distinct crystallites is observed within the top ${\sim} 67\;\textrm{nm}$ of the film. The “fingerprint” like patterns which represent the grains are approximately $5 - 9\;\textrm{nm}$ in diameter.

This region can also be faintly observed for the film annealed in $5{\%}$ H₂ in N₂, but determination of the crystallite size is inhibited by the lower quality of the image. This observation reveals that the two regions of an amorphous and polycrystalline character observed via XRD corresponds to the top and bottom layers of the film (as described above). The distinct layers arise from the inherent depth-dependence of the ReLA thermal treatment. The existence of a lower, amorphous layer and an upper, polycrystalline layer in the ITO film after ReLA can be ascribed to the heating mechanism of laser annealing [24,25]. The penetration depth, ${\mathrm{\delta }_\textrm{p}}$, (purple line in Fig. 4(a)) for the seed ITO film was calculated to be $43.2 \pm 0.1\;\textrm{nm}$ [1]. Within the region defined by ${\mathrm{\delta }_\textrm{p}}$, the heat is predominantly generated by the absorption of the $\textrm{KrF}$ photons. This region is indicated by the purple shaded area in Fig. 4(a). The lower portions of the film are then heated by heat diffusion from the top ${\sim} 67\;\textrm{nm}$. The diffusion will also result in an exponential decay of the peak temperature as a function of the depth (Fig. 4(b)). This region is indicated by the red shaded area in Fig. 4(a). Below this region, the temperature increase would be insufficient to significantly affect the material. Depending on the thickness of the film, this threshold may occur within the film, or extend into the substrate itself so that the entire film is thermally affected [24,25]. This region is indicated by the red shaded area in Fig. 4(a). The reduced significance of the thermal annealing further down the film explains why the transport properties only vary by ${\sim}{\pm} 10{\%}$ for the bottom layer of the laser processed ITO films (see Supplement 1, section 3). In Fig. 3(c)-(d), however, it was seen that the region of polycrystallinity stops at a depth of ${\sim} 67\;\textrm{nm}$, which is ${\sim} 60{\%}$ larger than the calculated ${\mathrm{\delta }_\textrm{p}}$. The absorption of the photons does not cut-off at the edge of the penetration depth. Instead, there is a smooth exponential gradient of the laser photon intensity (Fig. 4(a)) and thus temperature increase due to photon absorption. Across the film, the difference in the absorption would result in a gradient of the peak temperature reached during the ReLA process.

 

Fig. 4. (a-c) Schematic illustrating the thermal depth-dependence of the ReLA process. Extracted (d) grain size and (e) internal film stress of the seed un-optimized ITO films (grey stars) and those subject to ReLA at 25–125 mJcm^-2 in $5{\%}$ H₂ in N₂ (red circles) and O₂ (green triangles). The light shaded symbols denote top layer of the film after ReLA at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$.

Download Full Size | PDF

An illustration of the variation of the peak temperature, ${\textrm{T}_{\textrm{Peak}}}$, across the depth of the sample is shown as the solid red line in Fig. 4(b). The peak temperature closely follows the laser photon intensity. For a critical intensity of photons, ${\textrm{T}_{\textrm{Peak}}}$ would be sufficiently high to induce a crystallization of the ITO film. An illustrated example ${\textrm{T}_{\textrm{Peak}}}$ is indicated by the vertical dashed line in Fig. 4(b). Figure 3(c)–(d) demonstrates that this critical temperature must be reached within the upper, polycrystalline region that extends ${\sim} 67\;\textrm{nm}$ into the film. This region is indicated with the purple slab in Fig. 4(c). Below this region the peak temperature is insufficient to induce crystallinity and the layer remains amorphous. This is indicated with the pink slab in Fig. 4(c). It should be noted that the photon absorption relative to the incident photon intensity (i.e., the laser fluence) as a function of the depth is dependent only on the optical constants and the sample geometry [24,25]. However, the absolute absorption and thus the peak temperature as a function of the depth would depend on the fluence and other properties of the material such as the thermal conductivity and entropy. Furthermore, the critical temperature for the promotion of crystallinity must not only depend on the absorption coefficient at $5\;\textrm{eV}$.

This can be inferred by how laser processing of an ITO film with optimized sputtering conditions [25] did not show an upper crystallized layer even though the optical constants in the UV (above the band gap) were highly similar [26] (see Supplement 1, section 2). The heating mechanism described above results in a depth-dependent modification of the microstructural and optoelectronic properties. Figure 3(a) showed that the peak position of the ($222$) ITO peak for the seed ITO film was lower than the bulk ($222$) peak position of $30.6^\circ $. For films annealed at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$, the ($222$) peak is split into two parts for both the amorphous and crystalline layers. The peak for the amorphous (bottom) layer exists below the bulk ($222$) position and thus remains under compressive stress. Conversely, the peak for the crystalline (top) layer exists above the bulk ($222$) peak position and indicates the top portion of the film is under a tensile stress. The change in the crystalline state between the top and bottom layers cautions a direct comparison of the XRD peaks. However, the compressive stress of the seed ITO film and lower region of the laser processed films likely arises from the interfacial mismatch between the atomic spacings of the $\textrm{Si}$ and In₂O₃ lattices [58]. Furthermore, the high oxygen content at the bottom of the seed film (Fig. 5(d)) leads to a stretching of the lattice and a compressive stress, increasing the $\textrm{d}$-spacing and reducing the Bragg peak position. For the upper layer, the reduction in oxygen and creation of oxygen vacancies significantly reduces the $\textrm{d}$-spacing, resulting in a tensile stress. We consider it of great interest that the structural modifications occur at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ for the “un-optimized” ITO film studied in this work but not for an ITO film deposited with optimized sputtering conditions [25]. This is compounded by the observation in our previous work that ${\mathrm{\varepsilon }_2}(\textrm{E} )$ at $5\;\textrm{eV}$ (the KrF excimer laser energy) is lower for the un-optimized (low-$\textrm{N}$) film than for the optimized (high-$\textrm{N}$) film [32].

 

Fig. 5. Relative areal intensities of the deconvoluted component peaks of the core level X-ray photoelectron spectra for (a) O1s peaks: O_I (red bar), O_II (green bar) and O_III (blue bar) and (b) Sn: Sn²⁺ (pink bar) and Sn⁴⁺ (yellow bar) for the seed ITO film and that subject to a single ReLA pulse at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in $5{\%}$ H₂ in N₂. (c) The O/(In + Sn) ratio for the seed ITO film (grey bar) and that subject to a single ReLA pulse at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in $5{\%}$ H₂ in N₂ (red bar), calculated from surface-XPS. In (a-c), the solid horizontal lines indicate the values reported for an ITO film, with larger carrier concentration, deposited with “optimized” sputtering conditions. (d) Depth profile of O/(In + Sn) for the seed ITO film (grey stars) and those subject to single ReLA pulse at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in $5{\%}$ H₂ in N₂ (red circles) and $100{\%}$ O₂ (green stars), calculated from EDX imaging. O/(In + Sn) for stoichiometric In₂O₃ is indicated with the dashed black line. We map the XPS results onto the ‘surface’ (purple dashed line) of the EDX profile with the grey and red dashed lines.

Download Full Size | PDF

We posit two possible causes for the increased “sensitivity” of the un-optimized film to the energy dose:

Indeed, both factors may be at play and thus a deeper theoretical investigation into the heat diffusion mechanisms during ReLA is highly warranted.

In summary, at higher fluences ($125\;\textrm{mJc}{\textrm{m}^{ - 2}}$) in both reducing and oxidizing ambient environments, a polycrystalline nature was introduced to the upper layer of the film (with a grain size of 5–9 nm) while the lower layer remained amorphous. The depth of the polycrystalline layer was shown to result on the change in the photon absorption and peak temperature as a function of depth across the film. The depth-dependence of the laser processing explains the minor changes to the bottom of the ITO film during ReLA. However, the modulation of the carrier transport properties observed for the upper layer in Fig. 2 cannot be explained by a change in the structure alone. Furthermore, optical modelling revealed a high depth inhomogeneity in the seed ITO film itself. To explain these observations, an extensive investigation into the compositional properties of the seed and laser processed films was required.

5. Compositional ReLA-induced modifications

To reveal the precise nature of compositional modifications and how they relate to the alterations of optical properties of ITO during laser processing, we examined the surface oxidation states of the individual elements, via XPS, for the seed ITO film and for the ITO film subjected to single-pulse ReLA at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in the reducing environment ($5{\%}$ H₂ in N₂). Figure 5(a)-(b) presents the resultant areal intensities of (a) the O_I (red bar), O_II (green bar) and O_III (blue bar) components of the O1s core-level X-ray photoelectron spectra and (b) the Sn⁴⁺ (pink bar) and Sn²⁺ (yellow bar) components of the Sn3d_5⁄2 core-level X-ray photoelectron spectra . For comparison, the solid horizontal lines of each color presented across each bar indicate the values for an “optimized” ITO film with a larger carrier concentration [25]. Inspection of the deconvoluted O1s peaks reveals that the surface oxidations states for the seed ITO film of this work are different to those of the optimized film. Specifically, in relation to the optimized ITO film: the O_I peak that is related to the metal oxide [59] is significantly decreased while the O_III peak, related to surface contaminants [60], is increased. The O_II peak, related to oxygen atoms in the amorphous phase of ITO and/or oxygen atoms bound to Sn [60–62], is largely the same. This suggests that the ITO film of this work is highly entropic or amorphous [63], that there is a higher concentration of surface defects related to hydroxides [64], and/or that there is a smaller presence of metallic In [8]. Inspection of the deconvoluted Sn3d peaks for the seed ITO film revealed that the Sn⁴⁺ states are reduced in comparison to the high-quality ITO films [25], partly explaining the lower values of ${\textrm{N}_{\textrm{opt}}}$ and ${\textrm{N}_{\textrm{Hall}}}$ observed in Fig. 2.

The slight decrease in the O_II peak observed in (a) corresponds to the reduction in the Sn⁴⁺ peak observed in (b) and supports the case for Sn-O bonds as the source of the O_II peak [60–62]. For the film subject to a single ReLA pulse at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in the reducing environment of $5{\%}$ H₂ in N₂, the O_I peak is enhanced at the expense of the O_II peak, indicating that there is an enhancement of the crystalline order and a reduction of both Sn-O bonds. It is also found that the surface defects (O_III peak) are slightly reduced. There is also a corresponding activation of Sn⁴⁺ donors, partially explaining the increase in ${\textrm{N}_{\textrm{opt}}}$ and ${\textrm{N}_{\textrm{Hall}}}$ observed after laser processing. The concentration of oxygen vacancies cannot be reliably determined by de-convoluting the O1s XPS peak [25]. Therefore, we instead examine the abundance of oxygen relative to the abundance of the O1s, In3d_5/2 and Sn3d_5/2 peaks, where the individual peak areas were corrected by their respective relative sensitivity factors [65, 66].

The surface region O/(In + Sn) ratio is presented for the seed ITO film (grey bar) and that subject to single-pulse ReLA at $125\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in $5{\%}$ H₂ in N₂ (red bar) in Fig. 5(c). The O/(In + Sn) ratio is decreased by laser annealing, where the solid horizontal red line indicates the post-ReLA O/(In + Sn) ratio for an optimized ITO film [25]. The higher, more stoichiometric (dashed black line), concentration of oxygen likely corresponds to a smaller density of oxygen vacancies, explaining the other cause of the lower $\textrm{N}$ for the ITO film of this work. The increase in the oxygen deficiency after ReLA correspondingly explains the increase in the carrier concentration (alongside Sn⁴⁺ activation). Significantly, the oxygen vacancy modulation is directly related to the ambient dependence modulation of the carrier concentration during ReLA. It is important to note that XPS is sensitive only to the surface of the thin film, such that we lose information on the observed depth-dependence of the ReLA process. Figure 5(c) is used to calibrate the EDX-calculated depth profile of the normalized O/(In + Sn) ratio across the sample depth for the seed ITO film (grey stars) and those subject to single-pulse ReLA at $12\;\textrm{mJc}{\textrm{m}^{ - 2}}$ in $5{\%}$ H₂ in N₂ (red circles) and in $100{\%}$ O₂ (green triangles). This is presentation as a function of the distance from the bottom of the film in Fig. 5(d). Note that the EDX results are offset so that the values of O/(In + Sn) for the seed and annealed film, at roughly $10\;\textrm{nm}$ below the imaged surface (to account for the sampling depth [67]), match the values determined from XPS. Further details on the analysis of the EDX images is presented in Supplement 1, section 6. These steps allow us to examine how the oxygen content varies across the film. A gradient of the carrier concentration is observed throughout the depth of the film. The sharp increase of O/(In + Sn) at the bottom of the seed film may explain the existence of the bottom layer required to model the optical properties of the film. We note that the gradient in O/(In + Sn) is sharper for the film processed in $5{\%}$ H₂ in N₂ than for the film processed in $100{\%}$ O₂, and that near the surface there is a sharp dip in O/(In + Sn). From this, it is apparent that the reducing environment creates a steeper reduction of the oxygen across the depth of the film. This effect has been explained to arise from the promotion of oxygen mobility and Sn activation in response to the photo-induced localized heating [25].

6. Conclusions

Laser processing an “un-optimized”, low-free carrier, seed ITO material within pressurized environments containing reducing and oxidizing reactive gasses was investigated in order to further reveal the physical mechanisms behind the conversion of the optical properties of ITO during ReLA. A complex interplay between the film structure, composition and optoelectronic behavior was revealed, building upon previous investigations of ITO and ReLA. Specifically, multiple charge species and a complex inhomogeneity were required to model the optical properties of ITO. By examining the carrier transport mechanisms of the two carrier species for the larger set of seed and laser processed films, considering the effect of various scattering mechanisms, the nature of the multiple carrier species was revealed. The two carrier species were found to arise from carriers that were scattered by ionized impurities and those that were scattered by grain boundaries or by clusters of defects that act like grain boundaries. This has implications for ITO films with lower carrier concentration used to target ENZ regions further into the IR. Furthermore, the modelling used to elucidate the dual carrier species could be investigated for a wider range of nano- and poly-crystalline materials.

Clear structural modifications occurred when laser processing at a high fluence ($125\;\textrm{mJc}{\textrm{m}^{ - 2}}$). Specifically, the upper layer of the film became polycrystalline after ReLA. The depth dependence of the structural modifications was shown to arise from the change of the peak temperature across the film, due to the natural depth-dependence of the ReLA process. Despite the complexity of the structural modifications, it was found that the compositional modifications induced by ReLA were near-identical to previous reports. Specifically, the laser treatment induced both an activation of Sn dopants and a modification of the oxygen vacancy concentration throughout the depth of the film. The modulation of the oxygen vacancies was deemed to be the likely cause for the ambient dependent modulation of the carrier concentration and gave rise to the ability to tune the carrier concentration by controlling the mixture of oxidizing or reactive gasses within the pressure cell during ReLA. Hereby, we demonstrate how ReLA of low-$\textrm{N}$ ITO films can tune the carrier concentration of ITO between $({0.3 - 1.27} )\times {10^{20}}\;\textrm{c}{\textrm{m}^{ - 3}}$ and modulate the ENZ wavelength of the top layer between $3.07 - 11.7\;\mathrm{\mu}\textrm{m}$ (with a single laser pulse) while simultaneously enhancing the figures of merit for SPP, LSPR, near-field enhancement and photothermal conversion. Vitally, the modulation range is significantly larger than has been reported for ReLA of high-$\textrm{N}$ ITO films $1.91 - 1.73\;\mathrm{\mu}\textrm{m}$ [25], opening the mid-infrared spectral range for low-loss plasmonics and photonics.