Open Access
Add to BibSonomy
Abstract
We study the ultrafast time resolved response of 30 nm films of VO2 on a TiO2 substrate when 3.1 eV (400 nm wavelength) pump pulses were used to excite the insulator to metal transition (IMT). We found that the IMT threshold for these samples (≤30µJ/cm2) is more than 3 orders of magnitude lower than that generally reported for a more traditional 1.55 eV (800 nm wavelength) excitation. The samples also exhibited unusual reflectivity dynamics at near-threshold values of pump fluence where their fractional relative reflectivity ΔR/R initially increased before becoming negative after several hundreds of picoseconds, in stark contrast with uniformly negative ΔR/R observed for both higher 400 nm pump fluences and for 800 nm pump pulses. We explain the observed behavior by the interference of the reflected probe beam from the inhomogeneous layers formed inside the film by different phases of VO2 and use a simple diffusion model of the VO2 phase transition to support qualitatively this hypothesis. We also compare the characteristics of the VO2 films grown on undoped TiO2 and on doped TiO2:Nb substrates and observe more pronounced reflectivity variation during IMT and faster relaxation to the insulating state for the VO2/TiO2:Nb sample.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Vanadium dioxide (VO2) has been widely studied in the last several decades [1], mainly due to its insulator to metal transition (IMT) inducible by several different processes, such as electrical field, pressure, heating and ultrafast photo excitation [1–8]. The details of the ultrafast IMT dynamics in VO2 thin films triggered by ultrashort laser pulses have been explored using a variety of time-resolved techniques, including terahertz spectroscopy [9], IR transmittance measurements [10] and photoelectron spectroscopy [11], time-resolved x-ray diffraction [12,13], along with femtosecond x-ray pulses to look at atomic disordering [14]. Photon energies both above and below the VO2 bandgap energy have been used to study the IMT dynamics [5,6,12,15–21]. The majority of experiments use femtosecond pulses around 800 nm wavelength (photon energy 1.55 eV), conveniently available from commercial amplified laser systems [5–7,11,14,22,23]. The 1.55 eV pump energy is sufficient to excite the photoelectrons from the VO2 valence bands into the conduction band, which is usually reported to have a bandgap of ∼0.65 eV [24]. However, excitation of IMT in VO2 polycrystalline films with below-bandgap photo energies as low as 0.18 eV due to defect states in the gap have been reported [16].
This work explores effects induced by 3.1 eV pump pulses on the IMT in VO2 films. For such higher photon energy, photoelectrons can be excited from deep donor states into conduction bands as reported previously [17,18,25]. The main motivation for this study stems from the recently proposed application of VO2 films grown on TiO2 and TiO2:Nb substrates for efficient photodetection of blue and UV radiation with large external quantum efficiency [26]. Here we focus on the IMT dynamics in both VO2 on TiO2 and VO2 on TiO2:Nb samples and demonstrate several attractive features of high photon energy pump excitation. Compared to the traditional near-IR pump pulses, the blue pump-induced IMT has significantly lower threshold fluence requirements. Moreover, the addition of the dopant to the substrate results in larger magnitude changes in the ΔR/R responses, as well as faster rates of change for the IMT and its following recovery toward the original insulating state.
Most interestingly, in the range of pump fluences near the IMT threshold we observed dramatic fluence-dependent variation in the reflectivity, not previously reported for 1.55 eV pump pulses. Instead of the expected reduction of reflectivity for an IR optical probe, we observed an increased reflection for a few tens or hundreds of ps after the pump pulse, followed by slow continuous reduction of the reflectivity toward its expected value corresponding to the metallic form of VO2. A similar effect has been observed previously in VO2 on MgO, SiO2, and Al2O3 under the 3.1 eV excitation, and attributed to constructive or destructive interference of the optical probe reflecting off the various insulating and metallic layers of the VO2 film as the energy of the original pump pulse is dissipated throughout the thickness of the film [15,16,21]. It is important to note, however, that those experiments were conducted on much thicker (>200 nm) polycrystalline VO2 films and with almost three orders of magnitude higher pump fluences. This work is the first demonstration of such diffusion patterns for much thinner (30 nm) epitaxial film samples with uniform crystal orientation and with much lower pump fluences.
The paper is organized as follows. After a brief sample characterization, we describe the experimental setup and our main results on time-resolved studies of the VO2/TiO2 samples using frequency-doubled 3.1 eV pump pulses and 1.55 eV near-IR probe pulses. Then, using the measured optical parameters of the VO2 films in insulating and metallic states derived from ellipsometry measurements, we build a qualitative model of probe reflection taking into account spatial diffusion of the metallic state after initial pump excitation, and demonstrate good agreement with the experimental observations. Finally, we conduct a brief comparison of IMT parameters for two VO2 samples grown under the identical condition on undoped and Nb-doped TiO2 substrates and demonstrate that Nb doping enhances both the amplitude and the speed of the IMT.
2. Sample characterization
VO2 samples were grown by reactive DC magnetron sputtering process that is described in [26,27]. They were grown on two substrate types, TiO2 and TiO2:Nb. The latter was doped with 0.05% niobium by weight. The crystal orientation for the substrates was (002) with the TiO2 substrate being one side polished and TiO2:Nb being both side polished. Both samples were measured with a high degree of uniformity, with single crystal orientation as measured in both the XRD and RHEED images from the samples and along with the small mosaicity values of 0.0476 for VO2 on TiO2 and 0.378 for VO2 on TiO2:Nb reported in [26–28]. Both VO2 thin film samples used here were approximately 30 nm thick and had similar roughness magnitudes determined from AFM as seen in Fig. 1 with the VO2 on TiO2 having 18.8 nm RMS from the 5 × 5 µm scan and the VO2 on TiO2:Nb having 15.5 nm RMS for the 5 × 5 µm scan.
Fig. 1. AFM scans of the two-sample used. Left are 30 nm VO2 on TiO2 with 18.8 nm RMS roughness. Right is 30 nm VO2 on TiO2:Nb with 15.5 nm RMS roughness. Top images are 3D and bottom are 2D.
It is known that substrate choices [29], interface strains, and growth processes can influence the optical properties of individual VO2 thin films. Hence, to obtain accurate optical properties of our samples, we performed ellipsometry in the spectral range from 0.6 to 4.0 eV. Fitting of the optical properties was done through the WVASE program using Tauc-Lorentz oscillators [30]. The sample was measured in the insulating state and then in the metallic state via heating above the IMT critical temperature. The optical properties that were calculated also incorporate surface roughness effects. The surface roughness was determined by an AFM as shown in Fig. 1 for the VO2 on TiO2 sample. To incorporate the roughness in our model, it was assumed that the VO2 layer had grown with low surface roughness for the first 10nm approximately and subsequently roughened during the growth process (see Fig. 1) [28]. The data was fitted as a 10 nm smooth layer below an effective medium layer representation for the roughness for both the insulating and metallic states.
The dielectric functions obtained from ellipsometry in the insulating state and the metallic state are shown in Figs. 2(a) and 2(b) and help explain the sample’s response to pump-probe excitation. From Fig. 2(b) and according to the Tauc-Lorentz oscillator (TLO) model for the imaginary part of the dielectric function, ɛ2, the insulator state (blue lines) exhibits a large low energy band centered around 0.8 eV that has a low energy gap of about 0.2 eV. The three TLOs for the insulating state of VO2 are shown for both ɛ1 and ɛ2 in Fig. 2. The second bump at 2.75 eV, shown as a dotted blue light, results from two higher energy bands that overlap. Hence, 3.1 eV pump energy can excite electrons from the valence band into all three of these energy bands, while the lower 1.55 eV probe energy can only excite electrons into the lower energy band.
Fig. 2. Dielectric function ɛ of 30 nm VO2 on TiO2. Graph (a) is the real component of ɛ and graph (b) is the imaginary component of ɛ. The blue represents the insulating state and the red line represents the metallic state. The dashed blue lines are the components of the Tauc-Lorentz oscillators that make up the real and imaginary component of ɛ in the insulating state.
For the metallic state (red lines) one observes that the ɛ2 curve’s magnitude increases at low photon energies consistent with the collapse of the band gap and the emergence of free carriers indicating the presence of the metallic state. Further evidence of the metallic state is seen in negative values of ɛ1 at the low photon energy range of the curve. At the photon energy that is used as the probe, the dielectric functions for the two states that are seen in Table 1 have a larger magnitude difference then what is seen for the pump photon energy. Thus, the reflectivity difference between the two states will be more sensitive at 1.55 eV than 3.1 eV.
Table 1. Optical properties of VO2 on TiO2.
The optical properties of the insulating and metallic states at 1.5 and 3 eV are summarized in Table 1. In Table 1, n and k represent the real and imaginary components of the index of refraction, ɛ1 and ɛ2 represent the real and imaginary parts of the dielectric function, and δ is the optical penetration depth.
Finally, to characterize the thermal IMT properties of a similar sample under thermal excitation, we measured the temperature dependence of the optical reflection at both 700 nm and 400 nm light, shown in Fig. 3 (since the incoherent light source we had to use for this measurement did not reach 800 nm, we used the longest available wavelength, 700 nm; it was sufficient to verify the accuracy of our numerical modeling). As expected, we observed a reduction in reflectivity as the sample undergoes the transition. Both measurements agree in that the transition happens at approximately 50 ± 5 °C. The lower transition temperature seen for this sample is expected due to the strain that the TiO2 substrate imparts on to the VO2 film and has been reported previously [31,32]. The strain that the substrate imparts on the VO2 can lower energy requirements to cause the IMT in the VO2. ΔR/R is the relative change in the sample reflectance over its base reflectance in the VO2 insulator state. Negative value of ΔR/R shows the reduction of reflectance as the sample transitions into the metallic state. The larger ΔR/R change between the two VO2 states compared to the photoinduced transition measurements, arose from the larger incidence angle (45°) used for this measurement. As a side note, it is curious to point out that the 400 nm light reflectivity exhibits a suppressed thermal hysteresis (i.e. the difference in the curves for heating and cooling cycle are smaller), that is clearly observed in 700 nm measurements. In fact, we repeated these measurements at varying wavelengths of the optical probe, and observed complete collapse of the hysteresis below 400 nm, while the hysteresis was clearly observed for longer wavelengths, and its width was constant above 500 nm.
Fig. 3. Characterization of the thermally induced IMT of the VO2 on TiO2 film using the reflectivity of (a) 400 nm and (b) 700 nm incoherent light. The red dots indicate the measurements taken while heating the sample, while the blue dots correspond to the cooling.
3. Pump-probe measurements of IMT dynamics in a VO2 film on TiO2 substrate using 400 nm excitation
The experimental arrangement of the pump-probe experiment to measure the changes in the optical properties of VO2 thin films is similar to the one we have used previously [33]. We used a Ti:sapphire oscillator and regenerative amplifier laser system that generates ∼150fs pulses with a wavelength that was centered at 1.55 eV (800 nm) and energy of 1mJ per pulse with 1kHz repetition rate. The laser output was split into two beams using an 80/20 beam splitter, and the more powerful beam was doubled to 3.1 eV (400 nm) via the use of a 0.3mm-thick β-barium borate crystal to serve as a pump beam. A combination of a polarizer and a hot mirror efficiently removed any residual 1.55eV light from the pump line to a maximum level of 100nJ/cm2 which is an order of magnitude lower than the probe line fluence. The weaker 800 nm beam, serving as a probe, was redirected to a delay line that introduces an adjustable delay up to 4 ns between probe and pump pulses. The powers of both beams were further adjusted using neutral density filters. The probe beam was kept at a fluence of 4.3µJ/cm2 for all measurements which was significantly below the IMT threshold of 100µJ/cm2 for the 1.55eV photon energy for these samples.
Cross-polarized pump and probe beams were focused on a VO2 sample, such that their diameters on the surface were correspondingly 130 µm and 70 µm. The smaller probe size ensured that the pumped region was sampled across a uniform area. We used a silicon photodetector to measure the power of the reflected probe beam, with an additional polarizer, placed before the detector, helping to reject any scattered pump beam from reaching the silicon photodetector. To directly measure the change in probe reflectivity, we used a lock-in amplifier. The relative change of reflectivity ΔR/R was found by dividing the change in reflectively with the baseline reflectivity.
Figure 4 shows the time resolved ΔR/R response of the VO2 on TiO2 sample for different values of the blue pump fluence. For higher fluences (>25 µJ/cm2) the general time dependence is similar to the IMT dynamics produced by 800 nm pump pulses [9,33,34]: an initial few ps rapid reduction in reflectivity due to IMT produced by a pump pulse, is followed by a “slow-growth” stage, in which the sample continues to transition to metallic stage for several hundreds of ps via slower phonon-mediated interactions. The slow-growth response behavior has been described for photoinduced VO2 when pumped at 1.55 eV [9,33], although this type of behavior was typically observed in the range of several tens of ps. The slow-growth response in our measurements lasted up to 1000ps, which can be related to differences in the surface roughness or grain size in the samples, since it has been shown that larger particles or grain sizes can slow down this transition process [35]. After the slow growth the film began to relax back to the insulating state, causing the relative reflectivity to reverse direction and begin to increase toward zero again. This thermal recovery relaxation rate measured here was substantially faster than the previously measured value for VO2/TiO2 samples [29], potentially due to the smaller thickness of our sample and substantially lower pump powers and is further discussed and quantified in section 5.
Fig. 4. 30 nm VO2 on TiO2 time response for various fluences from 0 to 4000ps. The insert at top right shows times from 0 to 40ps which corresponds to the first 1% of the data and the area of the insert is represented by the small box seen at t=0. The vertical scale of the insert directly aligns with the corresponding main ΔR/R scale.
However, unlike previous experiments, we encountered an unexpected variation in ΔR/R signal for low pump fluences, as shown in Fig. 4. Specifically, we observed an immediate increase in reflectivity after the initial pump excitation, with the positive ΔR/R signal lasting up to a few hundreds of ps, depending on the pump fluence. The lowest fluency used, 18µJ/cm2 (blue line), was too low to cause the IMT, and thus the negligible ΔR/R response indicates little change in the optical properties of the VO2. However, even a slight increase in pump fluence to 22µJ/cm2 (red line) produced noticeable optical changes: an increase in reflectivity to ΔR/R ≈ 0.002, followed by a slow linear decrease, crossing zero at approximately 400-500ps and finally reaching negative ΔR/R ≈ -0.002 at 2-3 ns after the initial pump pulse. When the pump fluence was increased further to 30µJ/cm2 (purple line), the system’s response had a negative ΔR/R for all time delays, reaching ΔR/R = -0.0125 after about 1 ns. Further increase of fluence to 43µJ/cm2, (green line), did not change the overall temporal dependence, but allowed a faster drop toward a lower reflectivity change ΔR/R=-0.02.
4. Optical response diffusion model
Such an unexpected increase of reflectivity soon after the pump pulse has been reported previously when thicker samples on different substrates were pumped with 3.1 eV photons, but at much higher fluences [17,18]. In these previous studies the fluence-dependent reflectivity dynamics was attributed to constructive or destructive interference of the probe pulse reflecting off the evolving boundaries between various VO2 phases inside the film. As we show below, this model can qualitatively capture the essence of our experimental results. The key difference between our work and that of Lysenko et al. [17] is that there the pump power deposited onto the VO2 film was high enough to induce the IMT thermally in addition to direct photoexcitation. Thus, according to Ref [17], the thicker VO2 film did not homogeneously undergo a thermal IMT across the thickness as the energy of the absorbed pump pulse diffused through the 100-nm thick film. In our case, however, the energy required for initiation of the IMT is orders of magnitude lower which is attributed to the TiO2 substrate-induced strain in epitaxial VO2 films. Strain has been theorized to modify the bands and the effect has been modelled [36]. Furthermore, modifying the bands via doping the VO2 has been shown to further lower the fluence requirements of the IMT [37]. In our experiment the expected temperature change is not sufficient for inducing IMT by itself. The overall heating of the 30nm sample was calculated to be less than 3° C which is well below the thermal IMT value for this sample starting at room temperature at the maximum fluence of 43 µJ/cm2. This calculated temperature change was a conservative estimate based on the specific heat of VO2 found in Ref [38], the pump spot size, sample thickness of the VO2, and the assumption of roughly 60% absorption from the pump energy as predicted by the optical model of the beam energy entry the VO2. At these low energy densities, the system was nevertheless able to photoinduce the IMT in a manner that produced a ΔR/R zero crossing that roughly mimicked the thermal IMT process proposed in [17].
To model the varying material states throughout the VO2 film, we treated the VO2 material to be locally either an insulating layer, a metallic layers or a layer that is a combination of nucleated metallic area intermixed with insulating material [39]. As the film was probed in time, the fraction of metallic to insulator material evolved through the thickness of the sample. As a first order approximation, we modeled the changes in the VO2 optical properties throughout the film thickness as temporally evolving processes based on a first order diffusion equation in one dimension in order to describe the VO2 sample’s atypical ΔR/R zero crossing response at the fluences used on these samples. For a delta function surface excitation, representing a sub-ps pump pulse, the solution to the diffusion equation is of the form:
As time progresses, the sample composition evolves with time following the diffusion equation Eq.(1), causing deeper and deeper insulating layers to convert into metallic layers. We used the 4 × 4 optical matrix method [40] to calculate the probe beam’s propagation through each layer, taking into account their proper optical properties and layers dynamics. Figure 5 shows the film layer distribution at 10ps and 200ps. The color map shows the modeled depth ranges with insulating (blue), intermixed (rainbow), and metallic (red) optical properties. The black lines in Fig. 5 show the plot of the real component of the index of refraction. One can see that for the small delays, the front surface undergoes the IMT while deeper regions are unaffected, as seen in Fig. 5(a). As time progresses, the metallic VO2 fraction appears to be diffusing into deeper regions of the film increasing the depth of the IMT region, as seen in Fig. 5(b). The interference of the probe beam reflects off various boundaries within the evolving film and the overall optical reflectivity can briefly increase before finally decreasing when the sample becomes mostly metallic.
Fig. 5. Simulation of diffusion of the insulating VO2 into metallic VO2 layers. Red layers are metallic VO2. Blue layers are insulating. The colors between represent intermixed values of the optical properties of the VO2 layers. The black line shows the values of the real component of the index of the reflection. The two graphs (a) and (b) are for 10ps and then 200ps in time steps for the diffusion of the system.
The predictions of this model, shown in Fig. 6, qualitatively agree with the dynamics observed in the experiments. Here we chose a wide range of relative pump powers to elucidate the general time behavior seen in our measurements. The temporal ΔR/R response of this model mimics the experimental observations ΔR/R (Fig. 4) in these samples, where the sample’s ΔR/R response is initially positive and then goes negative over time, ultimately reaching the negative value of about −0.02 for the full transition to the metallic state, in full agreement with experimental results for higher fluence. The time at which the reflectivity switches from positive to negative shows similar strong dependence on the pump fluence, with lower pump powers resulting in longer crossover times. It is easy to see that this simple model overestimates the initial positive ΔR/R spike even for higher pump fluence that stems from the oversimplified assumption of the initial delta function excitation. Since such a model does not deal with pump beam penetration into the sample, it always predicts a positive spike in the beginning. Since we did not study the IMT dynamics at short times, this model was sufficient to explain the long-term reflectivity evolution. However, if more accurate calculations are desired, one needs to account properly for the beam penetration and the lateral spread of metallic state at the VO2 surface after the pump pulse that should damp the initial fast ΔR/R rise that is seen Fig. 6.
Fig. 6. Optical model response of 30 nm VO2 TiO2 for various relative fluence of light with using a 4 × 4 optical matrix method and the ellipsometry values with a hybrid layer of VO2 between the insulating and metallic states. The relative fluence at level 1 represents the power to fully transition the film in 200ps with blue light.
5. Effects of Nb doping of the TiO2 substrate on the photoinduced VO2 IMT dynamics
Recent studies demonstrated the strong effect of doping TiO2 substrate with Nb on the photoelectric properties of VO2 in blue and UV spectral ranges [26] thanks to a more efficient hole diffusion process and band bending at the VO2/TiO2 interface. To investigate the possible effects of such doping on the photoinduced IMT, we repeated the pump-probe measurements using 400 nm pump and 800 nm probe for a 30 nm VO2 sample, grown on the volume-doped TiO2:Nb substrate, as described in Sec. 2. Figure 7 shows the data for VO2 on TiO2 doped with Nb. The overall behavior of the optical reflection is very similar to that of the VO2/TiO2 sample. At the lowest fluence used on this sample, 17µJ/cm2 (blue line), the ΔR/R response changes very little and displays a positive change only with a slow recovery to zero after ∼2 ns, indicating little change in the optical properties of the VO2. At a slightly higher fluence of 20µJ/cm2 (orange line) one can see small changes in the ΔR/R throughout the time scale: the reflectance ratio crosses zero from a positive ratio of ∼0.002 to a negative ratio of −0.002 at about 2 ns. Even a very modest increase in fluence by just a few percents to 21µJ/cm2 (red line) yields a more drastic response: the material now exhibits a large positive change in reflectivity, followed by a significant negative ΔR/R after about 30ps with a sharp downward slope. The graph bottoms out at about a ΔR/R of about −0.03 in magnitude at around 1-2 ns, then starting to relax slowly back to zero. At the highest fluence level of 28µJ/cm2, (purple line), it is notable that the response is even stronger in magnitude and only trends negative with a peak drop of ΔR/R to −0.065, while the rate of the response is similar to the 21µJ/cm2 fluence.
Fig. 7. 30 nm VO2 on TiO2:Nb time response for various fluences from 0 to 4000ps. The insert at top right shows times from 0 to 40ps which corresponds to 1% of the total data and is represented by the small box seen at t=0 in the main graph. The vertical scale of the insert is equivalent to the main ΔR/R scale.
Despite many similarities, the careful comparison of Figs. 4 and 7 reveals some clear differences between the optical properties of the two samples. For example, it is obvious that the maximum magnitude of the reflectivity change is about three times higher for the samples with the doped substrate. The peak ΔR/R signal seen in Fig. 6 for the VO2 on TiO2:Nb substrate is ∼ −0.065 vs. VO2 on plain TiO2 substrate, which is ∼ −0.02. From Table 1, the optical properties correctly predicted the ΔR/R for VO2 on TiO2 substrate. In order to obtain a value of ΔR/R of a magnitude ∼ −0.065 for VO2 on TiO2:Nb substrate, the index of refraction of the TiO2:Nb substrate would need to be lowered from about 2.5 to ∼2 (assuming the VO2 thin film’s optical properties are constant). Such effect of doping reducing the refractive index has been reported for thin films of TiO2 for which the 5% Nb doping resulted in a reduction of the index of refraction by 5% [41].
The other distinct feature of the VO2 samples with the Nb-doped substrate is an even lower fluence threshold (21µJ/cm2), required to accomplish the IMT and the faster evolution during both the “slow growth” IMT stage and recovery stage, compared to the undoped substrate. Figure 8 insert shows examples of the ΔR/R data for VO2 on TiO2 and VO2 on TiO2:Nb, plotted from 0 to 4000ps and shows the slow growth exponential fit seen as the orange lines along with the start of the exponential recovery seen as the black lines.
Fig. 8. Time constants for the slow growth and recovery periods of the ΔR/R for the for the two VO2 samples. The blue line is for VO2 on TiO2 and red is for VO2 on TiO2:Nb. The insert is the graph of ΔR/R response of the 30 nm VO2 on TiO2 at 30µJ/cm2 (green line) and 30 nm VO2 on TiO2:Nb at 28µJ/cm2 (purple line). The time constant fits for both samples are shown as orange for the slow growth fit and black for the recovery fit.
The rates of these time responses are strongly fluence dependent. The responses for the lowest pump fluences were not fitted as they were not amenable to simple exponential time constants. For VO2/TiO2, increasing the fluence from 22µJ/cm2 to 30µJ/cm2 resulted in a 83% rate increase in the slow growth rate, while an increase from 22µJ/cm2 to 42µJ/cm2 resulted in about 84% which corresponds to an additional 1% increase in rate suggesting that slow growth rate is reaching a plateau for that sample and higher fluences, which is seen in Fig. 8. In the case of the VO2 on TiO2:Nb, increasing the fluence from 20µJ/cm2 to 21µJ/cm2 did appreciably increased the slow growth rate by approximately 80%. Increasing from 20µJ/cm2 to 28µJ/cm2 only resulted in an 88% rate which was an additional 8% increase suggesting that the VO2 on TiO2:Nb slow growth rate may be leveling off just like the VO2/TiO2 but at lower fluence. For the present samples, the fastest slow growth seen was 145ps for the VO2 on TiO2:Nb and 224ps for the VO2 on TiO2. For comparison with previously published work, the slow growth component seen in Fig. 8 is slower than what has been reported previously [33] which reported rates of 6 to 15ps.
The recovery rate of the VO2 ΔR/R responses from the Fig. 8 insert is $ 34.35_{ - 0.02}^{ + 0.01} $ns for the VO2 on TiO2:Nb (28µJ/cm2). For the 30µJ/cm2 VO2 on TiO2, its long-time scale is still growing and produces a larger error with values of -$49.32_{ - 0.51}^{ + 0.66} $ns. The 43µJ/cm2 scan of VO2 on TiO2 had a slower recovery time constant of $44.37_{ - 0.05}^{ + 0.05}$ns. That fit is not shown in the insert in Fig. 8 but the change in the rate can be observed in Fig. 4 between the 43µJ/cm2 scan and 30µJ/cm2. These recovery rates are faster than what has been reported in Ref [29]. The difficulty with these measurements reflects the limited available delay time to truly evaluate the long recovery times, resulting in large confidence limits.
6. Conclusion
Our studies demonstrated that 3.1 eV ultrafast pulses can efficiently photo induce IMT in thin film samples of VO2 grown on TiO2 and TiO2:Nb substrates at fluence values orders of magnitude lower than what is typically required by the IR pump. Our pump probe measurements of the relative reflectivity ΔR/R revealed the unexpectedly rich dynamics of the film’s optical response due to the evolution of the VO2 metallic fraction through the film thickness, indicating qualitative agreement between the experiment and a simple model based on a diffusion type transition dynamics coupled with a 4 × 4 optical matrix method. Due to its strong pump fluence dependence, this mechanism can provide additional tools for all-optical control of VO2-based photonic devices. We also found that the addition of Nb dopant to the substrate improves the magnitude of the ΔR/R response of the thin film VO2, as well as increasing the rate of change for both in the slow growth and relaxation stages and increases the sensitivity to the pump fluence. These properties illustrate the advantage of using the Nb-doped TiO2 substrates not only for VO2-based photodetectors but also for ultrafast IMT applications.
Funding
Defense Threat Reduction Agency (HDTRA 1-16-1-0056); National Science Foundation (IIP-1827536).
Disclosures
The authors declare no conflicts of interest.
References
1. F. J. Morin, “Oxides which show a metal-to-insulator transition at the neel temperature,” Phys. Rev. Lett. 3(1), 34–36 (1959). [CrossRef]
2. G. Stefanovich, A. Pergament, and D. Stefanovich, “Electrical switching and mott transition in VO2,” J. Phys.: Condens. Matter 12(41), 8837–8845 (2000). [CrossRef]
3. B. A. Kruger, A. Joushaghani, and J. K. S. Poon, “Design of electrically driven hybrid vanadium dioxide (VO2) plasmonic switches,” Opt. Express 20(21), 23598–23609 (2012). [CrossRef]
4. L. Bai, Q. Li, S. A. Corr, Y. Meng, C. Park, S. V. Sinogeikin, C. Ko, J. Wu, and G. Shen, “Pressure-induced phase transitions and metallization in VO2,” Phys. Rev. B 91(10), 104110 (2015). [CrossRef]
5. A. Cavalleri, C. Tóth, C. W. Siders, J. A. Squier, F. Ráksi, P. Forget, and J. C. Kieffer, “Femtosecond structural dynamics in VO2 during an ultrafast solid-solid phase transition,” Phys. Rev. Lett. 87(23), 237401 (2001). [CrossRef]
6. A. Cavalleri, T. Dekorsy, H. H. W. Chong, J. C. Kieffer, and R. W. Schoenlein, “Evidence for a structurally-driven insulator-to-metal transition in VO2: A view from the ultrafast timescale,” Phys. Rev. B 70(16), 161102 (2004). [CrossRef]
7. A. Cavalleri, M. Rini, H. H. W. Chong, S. Fourmaux, T. E. Glover, P. A. Heimann, J. C. Kieffer, and R. W. Schoenlein, “Band-Selective Measurements of Electron Dynamics in VO2 Using Femtosecond Near-Edge X-Ray Absorption,” Phys. Rev. Lett. 95(6), 067405 (2005). [CrossRef]
8. S. E. Madaras, J. Creeden, S. Kittiwatanakul, J. Lu, I. Novikova, and R. A. Lukaszew, “Insulator to metal transition induced by surface plasmon polaritons in VO2/Au thin films,” Opt. Express 26(20), 25657–25666 (2018). [CrossRef]
9. T. L. Cocker, L. V. Titova, S. Fourmaux, G. Holloway, H. C. Bandulet, D. Brassard, J. C. Kieffer, M. A. El Khakani, and F. A. Hegmann, “Phase diagram of the ultrafast photoinduced insulator-metal transition in vanadium dioxide,” Phys. Rev. B 85(15), 155120 (2012). [CrossRef]
10. V. R. Morrison, R. P. Chatelain, K. L. Tiwari, A. Hendaoui, A. Bruhács, M. Chaker, and B. J. Siwick, “A photoinduced metal-like phase of monoclinic VO2 revealed by ultrafast electron diffraction,” Science 346(6208), 445–448 (2014). [CrossRef]
11. D. Wegkamp, M. Herzog, L. Xian, M. Gatti, P. Cudazzo, C. L. McGahan, R. E. Marvel, R. F. Haglund, A. Rubio, M. Wolf, and J. Stähler, “Instantaneous band gap collapse in photoexcited monoclinic VO2 due to photocarrier doping,” Phys. Rev. Lett. 113(21), 216401 (2014). [CrossRef]
12. M. Hada, K. Okimura, and J. Matsuo, “Characterization of structural dynamics of VO2 thin film on c-Al2O3 using in-air time-resolved x-ray diffraction,” Phys. Rev. B 82(15), 153401 (2010). [CrossRef]
13. M. Hada, K. Okimura, and J. Matsuo, “Photo-induced lattice softening of excited-state VO2,” Appl. Phys. Lett. 99(5), 051903 (2011). [CrossRef]
14. S. Wall, S. Yang, L. Vidas, M. Chollet, J. M. Glownia, M. Kozina, T. Katayama, T. Henighan, M. Jiang, T. A. Miller, D. A. Reis, L. A. Boatner, O. Delaire, and M. Trigo, “Ultrafast disordering of vanadium dimers in photoexcited VO2,” Science 362(6414), 572–576 (2018). [CrossRef]
15. M. F. Becker, A. B. Buckman, R. M. Walser, T. Lépine, P. Georges, and A. Brun, “Femtosecond laser excitation of the semiconductor-metal phase transition in VO2,” Appl. Phys. Lett. 65(12), 1507–1509 (1994). [CrossRef]
16. M. Rini, Z. Hao, R. W. Schoenlein, C. Giannetti, F. Parmigiani, S. Fourmaux, J. C. Kieffer, A. Fujimori, M. Onoda, S. Wall, and A. Cavalleri, “Optical switching in VO2 films by below-gap excitation,” Appl. Phys. Lett. 92(18), 181904 (2008). [CrossRef]
17. S. Lysenko, A. J. Rua, V. Vikhnin, J. Jimenez, F. Fernandez, and H. Liu, “Light-induced ultrafast phase transitions in VO2 thin film,” Appl. Surf. Sci. 252(15), 5512–5515 (2006). [CrossRef]
18. S. Lysenko, V. Vikhnin, G. Zhang, A. Rua, F. Fernandez, and H. Liu, “Insulator-to-metal phase transformation of VO2 films upon femtosecond laser excitation,” J. Electron. Mater. 35(10), 1866–1872 (2006). [CrossRef]
19. C. Kübler, H. Ehrke, R. Huber, R. Lopez, A. Halabica, R. F. Haglund, and A. Leitenstorfer, “Coherent Structural Dynamics and Electronic Correlations during an Ultrafast Insulator-to-Metal Phase Transition in VO2,” Phys. Rev. Lett. 99(11), 116401 (2007). [CrossRef]
20. A. Pashkin, C. Kübler, H. Ehrke, R. Lopez, A. Halabica, R. F. Haglund, R. Huber, and A. Leitenstorfer, “Ultrafast insulator-metal phase transition in VO2 studied by multiterahertz spectroscopy,” Phys. Rev. B 83(19), 195120 (2011). [CrossRef]
21. M. Nakajima, N. Takubo, Z. Hiroi, Y. Ueda, and T. Suemoto, “Photoinduced metallic state in VO2 proved by the terahertz pump-probe spectroscopy,” Appl. Phys. Lett. 92(1), 011907 (2008). [CrossRef]
22. M. van Veenendaal, “Ultrafast photoinduced insulator-to-metal transitions in vanadium dioxide,” Phys. Rev. B 87(23), 235118 (2013). [CrossRef]
23. D. Wegkamp and J. Stähler, “Ultrafast dynamics during the photoinduced phase transition in VO2,” Prog. Surf. Sci. 90(4), 464–502 (2015). [CrossRef]
24. C. N. Berglund and H. J. Guggenheim, “Electronic Properties of VO2 near the Semiconductor-Metal Transition,” Phys. Rev. 185(3), 1022–1033 (1969). [CrossRef]
25. V. S. Vikhnin, S. Lysenko, A. Rua, F. Fernandez, and H. Liu, “The model of ultrafast light-induced insulator-metal phase transition in vanadium oxide,” J. Phys.: Conf. Ser. 21, 44–49 (2005). [CrossRef]
26. J. A. Creeden, S. E. Madaras, D. B. Beringer, I. Novikova, and R. A. Lukaszew, “Growth and Characterization of Vanadium Dioxide/Niobium Doped Titanium Dioxide Heterostructures for Ultraviolet Detection,” Adv. Opt. Mater. 7(23), 1901143 (2019). [CrossRef]
27. J. A. Creeden, S. E. Madaras, D. B. Beringer, M. R. Beebe, I. Novikova, and R. A. Lukaszew, “Structural and Photoelectric Properties of Epitaxially Grown Vanadium Dioxide Thin Films on c-Plane Sapphire and Titanium Dioxide,” Sci. Rep. 9(1), 9362 (2019). [CrossRef]
28. J. A. Creeden, S. E. Madaras, D. B. Beringer, I. Novikova, and R. A. Lukaszew, “Intrinsic anomalous scaling of epitaxial vanadium dioxide thin films on titanium dioxide,” AIP Adv. 9(9), 095045 (2019). [CrossRef]
29. M. Rodriguez-Vega, M. T. Simons, E. Radue, S. Kittiwatanakul, J. Lu, S. A. Wolf, R. A. Lukaszew, I. Novikova, and E. Rossi, “Effect of inhomogeneities and substrate on the dynamics of the metal-insulator transition in VO2 thin films,” Phys. Rev. B 92(11), 115420 (2015). [CrossRef]
30. T. J. Huffman, P. Xu, A. J. Hollingshad, M. M. Qazilbash, L. Wang, R. A. Lukaszew, S. Kittiwatanakul, J. Lu, and S. A. Wolf, “Modification of electronic structure in compressively strained vanadium dioxide films,” Phys. Rev. B 91(20), 205140 (2015). [CrossRef]
31. J. Cao, E. Ertekin, V. Srinivasan, W. Fan, S. Huang, H. Zheng, J. W. L. Yim, D. R. Khanal, D. F. Ogletree, J. C. Grossman, and J. Wu, “Strain engineering and one-dimensional organization of metal–insulator domains in single-crystal vanadium dioxide beams,” Nat. Nanotechnol. 4(11), 732–737 (2009). [CrossRef]
32. E. Radue, E. Crisman, L. Wang, S. Kittiwatanakul, J. Lu, S. A. Wolf, R. Wincheski, R. A. Lukaszew, and I. Novikova, “Effect of a substrate-induced microstructure on the optical properties of the insulator-metal transition temperature in VO2 thin films,” J. Appl. Phys. 113(23), 233104 (2013). [CrossRef]
33. E. Radue, L. Wang, S. Kittiwatanakul, J. Lu, S. A. Wolf, E. Rossi, R. A. Lukaszew, and I. Novikova, “Substrate-induced microstructure effects on the dynamics of the photo-induced metal–insulator transition in VO 2 thin films,” J. Opt. 17(2), 025503 (2015). [CrossRef]
34. M. R. Beebe, J. M. Klopf, Y. Wang, S. Kittiwatanakul, J. Lu, S. A. Wolf, and R. A. Lukaszew, “Time-resolved light-induced insulator-metal transition in niobium dioxide and vanadium dioxide thin films,” Opt. Mater. Express 7(1), 213–223 (2017). [CrossRef]
35. S. Lysenko, V. Vikhnin, A. Rúa, F. Fernández, and H. Liu, “Critical behavior and size effects in light-induced transition of nanostructured VO2 films,” Phys. Rev. B 82(20), 205425 (2010). [CrossRef]
36. B. Lazarovits, K. Kim, K. Haule, and G. Kotliar, “Effects of strain on the electronic structure of VO2,” Phys. Rev. B 81(11), 115117 (2010). [CrossRef]
37. N. Émond, A. Ibrahim, B. Torriss, A. Hendaoui, I. Al-Naib, T. Ozaki, and M. Chaker, “Impact of tungsten doping on the dynamics of the photo-induced insulator-metal phase transition in VO2 thin film investigated by optical pump-terahertz probe spectroscopy,” Appl. Phys. Lett. 111(9), 092105 (2017). [CrossRef]
38. H. Kizuka, T. Yagi, J. Jia, Y. Yamashita, S. Nakamura, N. Taketoshi, and Y. Shigesato, “Temperature dependence of thermal conductivity of VO2thin films across metal–insulator transition,” Jpn. J. Appl. Phys. 54(5), 053201 (2015). [CrossRef]
39. M. M. Qazilbash, M. Brehm, B.-G. Chae, P.-C. Ho, G. O. Andreev, B.-J. Kim, S. J. Yun, A. V. Balatsky, M. B. Maple, F. Keilmann, H.-T. Kim, and D. N. Basov, “Mott Transition in VO2 Revealed by Infrared Spectroscopy and Nano-Imaging,” Science 318(5857), 1750–1753 (2007). [CrossRef]
40. D. W. Berreman, “Optics in stratified and anisotropic media: 4×4-matrix formulation,” J. Opt. Soc. Am. 62(4), 502–510 (1972). [CrossRef]
41. A. Shah, A. Mahmood, U. Aziz, R. Rashid, Q. Raza, and Z. Ali, “Influence of Nb content on the structural and optical properties of anatase TiO2 polycrystalline thin film by e-beam technique,” Mater. Chem. Phys. 180, 383–389 (2016). [CrossRef]
