1. Introduction

To obtain realistic images, three-dimensional (3D) display technologies that have no spatial limitations are garnering attention. To implement the 3D display, stereoscopic and volumetric displays have been proposed. Among them, as one of the volumetric displays, a hologram is considered as an effective technology because it can generate the 3D display without causing additional virtual fatigue, such as eyestrain or nausea; moreover additional tools, such as glasses or screens are not required in this case [1]. The hologram can project 3D information of objects via interference and diffraction of light which are easily generated by spatial light modulators (SLMs). In addition, when diffraction angle of the SLM becomes larger, more detail 3D information of objects can be expressed. Thus, to enlarge the diffraction angle, the SLM requires smaller pixel size closed to visible light wavelength (VL-$\lambda$: 380 $\sim$ 780 nm) [2]. However, there exist a challenge when fabricating the VL-$\lambda$ sized SLM by typical fabrication technologies such as liquid crystal or micro-mirrors. It is because unit cell size of the liquid crystal and micro-mirrors requires more than few micrometers [3–7].

Therefore, we propose a novel approach that utilizes ion intercalation. When we consider the ion intercalation, the movement and intercalation of nano sized ion are the main operation mechanism. Based on this, the unit cell can be scaled down to nano scale. The size of unit cell limits only number of reacted ions. Thus, ratio of transmittance change could become smaller in nano scaled SLM, but the ion intercalation will be occurred as the same as the case of micro scaled cell [8,9]. To demonstrate the operation mechanism of ion intercalation based SLM, we investigated the feasibility of micro scaled SLM. In addition, by controlling applied bias, gradual transmittance changes can be obtained, which can facilitate the gradual modulation of the light intensity. To achieve the gradual transmittance changes, we optimized the fabrication conditions such as working pressure and oxygen proportion.

2. Experiments

The proposed approach was implemented by developed gradual transmittance-controllable device (GTC-device). The GTC-device has the structure: indium tin oxide (ITO, 240 nm thick)/WO$_x$ (100 nm thick)/lithium phosphorus oxynitride (LiPON, 40 nm thick)/ITO(240 nm thick), as shown in Fig. 1(a). Through typical radio frequency sputtering system and targets (fabricated by the TAEWON SCIENTIFIC CO., LTD., Korea), solid-state thin films were fabricated as a simple square patterned metal-insulator-metal structure. Comparing to previously published tunable photonic device, the GTC-device has simple solid-state structure composed of well-known materials [10–12]. The bottom ITO electrode was deposited by sputtering an ITO target on whole surface of bare glass substrate. The sputtering was carried out at 40 mTorr in Ar ambient gas which was supplied by a mass flow controller. Then, to fabricate the WO$_x$ layers as a lithium-ion reservoir, W target was sputtered in Ar and O$_2$ mixed ambient gas (O$_2$ reactive sputtering) at various working pressures (6, 28.5, and 35 mTorr). For the optimization of fabrication conditions, thickness of WO$_x$ layers in each condition was confirmed by a surface profiler. The same thickness of WO$_x$ layer was kept by controlling sputtering time during various fabrication conditions. For solid state electrolyte and source layer of lithium ion, the LiPON layer was formed by sputtering a Li$_3$PO$_4$ target in Ar and N$_2$ mixed ambient gas (N$_2$ reactive sputtering). Next, the top 240 nm thick ITO electrode was deposited in the same condition of the bottom ITO through patterned metal shadow mask (200 $\mu$m $\times$ 200 $\mu$m square patterns). Each apart square pattern (by 350 $\mu$m from each other) can be considered as one pixel of the SLM.

 

Fig. 1. (a) Fabrication process flow of the proposed GTC-device. For the GTC-device, Li ion intercalation was utilized. (b) Cross sectional transmission electron microscope image and energy-dispersive X-ray spectroscopy mapping analyses of the fabricated GTC-device.

Download Full Size | PDF

Figure 1(a) illustrates the fabrication process flow of the GTC-device. The fabricated GTC-device was inspected by the transmission electron microscope with energy dispersive X-ray analysis, as shown in Fig. 1(b). To evaluate the developed devices, DC sweep voltages (started from 0 V to −10 V to +10 V, then went back to 0 V) were applied with 100 mV ramp voltage.

3. Results and discussion

According to previously reported researches, the WO$_x$ layer can change its transmittance when Li ion intercalates into the WO$_x$ layer. When the Li ions intercalate and form bonds with the WO$_3$, the WO$_x$ layer can be changed from transparent state (bleach state) to colored state (coloration state) [13,14]. For gradual transmittance changes, the Li ion intercalation needs to be sensitively modulated. Thus, optimization of the WO$_x$ layer is necessary to realize the GTC-device. To optimize the WO$_x$ layer, firstly, three different working pressures (6, 28.5, and 35 mTorr) were compared during deposition of the WO$_x$ layer, as summarized in Fig. 2(a). Considering that a higher working pressure leads to more porous film formation, an excessively dense (or porous) WO$_x$ films can be formed at 6 mTorr (or 35 mTorr) [15]. Therefore, in the device fabricated at 6 mTorr, insufficient Li ions permeate into the WO$_x$ layer, which leads to no changes in optical state under applied bias ($\pm$10 V). By contrast, the devices fabricated at 28.5 and 35 mTorr exhibited two optically different states. However, in rate of change respect, the device fabricated at 28.5 mTorr exhibited better performance. These results imply that a suitable working pressure which can modulate the amount of permeated Li ions is required to optimize the optical characteristics.

 

Fig. 2. Optical characteristics of the GTC-device based on different (a) working pressures and (b) oxygen proportions. At optimized conditions (28.5 mTorr and 60$\%$), the GTC-device exhibited more reliable and clear optical characteristics.

Download Full Size | PDF

Moreover, transmittance of the WO$_x$ layer can also be affected by the phase [16]. Thus, at the optimized working pressure of 28.5 mTorr, we tried to optimize the phase of WO$_x$ layer by controlling the rate of oxygen and argon ambient gases during the deposition process. The total amount of ambient gas was fixed at 20 sccm, and the oxygen ratio was changed from 20 to 80 percent, for which cases the devices are named as 20$\%$-, 40$\%$-, 60$\%$-, and 80$\%$-devices. As presented in Fig. 2(b), the 20$\%$-device initially exhibited coloration state (dot square pattern indicates ITO top electrode), and it was not changed under the applied bias. Otherwise, other devices exhibited coloration and bleach states under the applied bias. The 40$\%$-device exhibited coloration and bleach state under positive and negative bias, respectively. However, the 60$\%$- and 80$\%$-device exhibited opposite behaviors with the 40$\%$-device; the coloration state was observed under negative bias, vice versa. To clarify the operation mechanisms, we performed X-ray photo-electron spectroscopy (XPS) analyses on as-deposited layers of the 20$\%$-, 40$\%$-, and 80$\%$- devices. Note that the tendency of oxygen rate dependence was intensively investigated (Fig. 3).

 

Fig. 3. (a) XPS depth profile of the 20$\%$-device. The Li ions are initially diffused into the WO$_x$ layer. For 20$\%$-, 40$\%$-, and 80$\%$-device, (b) lithium 1s XPS spectrum in the middle of WO$_x$ layer and (c) tungsten 4f XPS spectrum at interface between LiPON and WO$_x$ layers. (A and B mean more complicated WO$_x$ bindings.)

Download Full Size | PDF

Figure 3(a) illustrates the atomic concentration of the 20$\%$-device via XPS depth profiling; the Li ions already diffused into, and reacted with the WO$_x$ layer at the interface. Therefore, Li$_x$WO$_3$ bonds can be initially formed which can exhibit a blue color, as shown in Fig. 2(b) [6]. For detailed mechanisms, the XPS spectrum was analyzed in two regions: (i) in the middle of the WO$_x$ layer and (ii) at the interface between LiPON and WO$_x$ layers. For the region (i), chemical reaction of the Li ions with WO$_x$ was analyzed, as illustrated in Fig. 3(b). The XPS spectrum related to Li 1s is divided into two peaks at binding energies of 51.0 and 55.9 eV, indicating the existence of Li and Li-O bonds, respectively [17]. Considering that the XPS spectrum was measured at the middle of the WO$_x$ layer, the Li-O bond peak indicates Li-WO$_x$ binding that is the origin of coloration state [14,18–21]. When the oxygen ratio increased, more Li-O bond peaks were observed, while Li peaks decreased. It means that more Li-WO$_x$ bonds which is an origin of coloration state can be formed as the oxygen ratio increases. Thus, more clear transmittance changes were observed in the devices having higher oxygen proportion (Fig. 2(b)).

In addition, in region (ii), the WO$_x$ phases were analyzed, as shown in Fig. 3(c). On comparison of the three conditions, when the oxygen ratio was low, stoichiometric WO$_x$ layer consisting of WO$_3$ and WO$_2$ was formed. However, for the 80$\%$-device, more complicated phases were observed, which means the non-stoichiometric WO$_x$ layer. These results mean that the Li-WO$_x$ bonds of 20$\%$-device can be considered as Li$_x$WO$_3$ bonds. It is because the 20$\%$-device exhibited initial coloration state and involved the stoichiometric WO$_x$ layer [19–22].

Moreover, in the 80$\%$-devices, the non-stoichiometric WO$_x$ layer can have more oxygen vacancies or mobile oxygen ions, which implies that the phases of WO$_x$ can be modulated by applied bias [23–26]. Based on these results, operation mechanisms of the 60$\%$- and 80$\%$-devices exhibiting coloration state under negative bias can be explained by oxygen migrations. Considering that only lithium and oxygen ions can be migrated by the applied bias, and the Li-WO$_3$ bonds are necessary for the coloration state, the oxygen ions can be migrated from WO$_x$ layer to ITO layer under the negative bias. Then, more Li-WO$_3$ bonds are generated by the phase modulation of the WO$_x$ layer. Therefore, the 60$\%$- and 80$\%$-devices exhibited coloration and bleach states under the applied negative and positive bias, respectively. In particular, among the evaluated devices (20$\%$-, 40$\%$-, 60$\%$-, and 80$\%$-devices), the 60$\%$-device exhibited clearer and consecutive optical transitions.

Based on these results, gradual transmittance changes were achieved in the optimized device (60$\%$-device), as shown in Fig. 4(a). Under higher negative bias, more oxygen ions of the WO$_x$ layer can be migrated, which leads to stronger coloration state. To demonstrate the basic SLM operations: light interference and diffraction, an optical experiment was conducted based on the GTC-device, as shown in Fig. 4(b-c). Two GTC-devices were employed as pixels for three cases: all bleach, all coloration, and bleach and coloration (named odd). Then, we prepared a dot slit on the GTC-devices, and performed irradiation using a 530 nm laser. In Fig. 4(c), interference and diffraction patterns were successfully observed when the device states changed as all bleach, odd, and all coloration. These results clearly demonstrate that the SLM based on gradual transmittance changes can cause light interference and diffraction, which also strongly supports the feasibility of the GTC-device as a micro SLM.

 

Fig. 4. (a) Achieved gradual transmittance modulation of the GTC-device under variable bias (−10 $\sim$ −5 V). (b) Simple schematic of optical dot slit experiment. (c) Light interference and diffraction patterns depending on GTC-device states: all bleach, bleach and coloration (named odd), and all coloration. These results strongly mean that the developed GTC-device can occur light interference and diffraction by the transmittance modulation.

Download Full Size | PDF

4. Conclusion

In this study, the gradual transmittance change based SLM (GTC-device) was developed by optimization of the Li ion intercalation. Considering that nano sized ions are the origin of transmittance change, the unit cell of SLM can be scaled down to nano-scale. The developed GTC-devices were evaluated in the dot slit experiment, and clearly showed the basic SLM operations: light interference and diffraction. These results successfully demonstrated the feasibility of GTC-device as the SLM.