Information integrated glass module fabricated by integrated additive and subtractive manufacturing

Abstract

In this letter, we report a novel integrated additive and subtractive manufacturing (IASM) method to fabricate information integrated glass module. After a certain number of glass layers are 3D printed and sintered by direct CO₂ laser irradiation, a microchannel will be fabricated on top of the printed glass by integrated picosecond laser, for intrinsic Fabry-Perot interferometer (IFPI) optical fiber sensor embedment. Then glass 3D printing process continues for the realization of bonding between optical fiber and printed glass. Temperature sensing up to 1000°C was demonstrated using the fabricated information integrated module. In addition, the long-term stability of the glass module at 1000°C was conducted. Enhanced sensor structure robustness and harsh temperature sensing capability makes this glass module attractive for harsh environment structural health monitoring.

Compact, robust temperature sensors are increasingly necessary in industrial applications, such as engine turbines, downhole and boiler tubes, for temperature measurement to ensure consistent processes [1,2]. Precise and in situ temperature monitoring is of great importance in these fields. With compact size and accurate reading capability, electrical thermometers like thermocouples and resistance temperature devices (RTD) have been widely applied in various applications. However, electrical thermometers cannot survive under harsh conditions (e.g., high temperature >600°C, high pressure and chemical corrosions). In addition, for temperature sensing in harsh conditions, extreme condition survivability and stability of sensors should be critical. As such, the entire sensing module should be built using high temperature materials for stable and reliable performance in harsh conditions.

Fiber-optic sensors have been widely demonstrated and applied for harsh environment applications. Compared with electronic sensors, fiber-optic sensors have the unique advantages of light weight, chemical resistance, high temperature resistance and distributed sensing capability [1–5]. Additionally, because of their immunity to electromagnetic interference (EMI), fiber optic sensors can be applied in strong EMI harsh environments [2–9]. Over the past years various optical fiber sensors have been proposed and demonstrated for temperature sensing, such as fiber interferometers, long period fiber gratings, fiber Bragg gratings and optical fiber directional couplers [6,8–11]. These sensors generally operate by detecting the temperature-induced optical parameter changes such as intensity, wavelength and phase.

Phase/Wavelength-modulated fiber optic temperature sensors are more stable, easy to fabricate and interrogate [5,6,8]. Compared with the phase/wavelength-based temperature sensors, intensity-based fiber-optic sensors often have limitations due to their sensitivity to environmental changes such as the bending of optical fiber induced intensity change and source power vibrations [12]. Typically, a reflection mode probe-type configuration is preferred for temperature sensing with the advantage of easy installation during applications, enhanced robustness and improved signal stability. Among the many demonstrated reflection-mode based probes, the interferometer and grating based structures are among the most popular choices for temperature sensing and have been successfully developed and proven to be capable of operating in extreme conditions [5,8,12,13].

Although fiber-optic sensors have shown quite effective temperature measurements, it is observed that the implementation is difficult due to the fragile nature of optical fiber, and the difficulty of mounting optical fibers into harsh environments like turbines and downhole. Compared with the conventional sensor packaging method, which is realized by attaching or mounting sensors on structures after being fabricated, additive manufacturing approach in sensor development has been initiated recently by directly embedding sensors into functional parts or smart structures to realize in situ measurement of parameters of interests. Optical fiber sensors have been demonstrated to be embedded in 3D printed ceramic and metal composites and shown the potential for high temperature sensing applications [11,14,15]. However, considering different material properties, especially the large difference of thermal expansion coefficient (CTE) between embedded structure and optical fiber, extra protective coating layers on fiber or external protectors may be necessary to avoid embedded fiber damage [11,14,15].

Sharing the same material properties (e.g., CTE) as optical fiber, fused silica is the natural choice of material for fabricating fiber-optic sensor embedded structures. Recently, glass 3D printing utilizing CO₂ laser direct melting method has been demonstrated. It offers great flexibility and simplicity to produce desired 3D structures with no post-treatment and high temperature resistance [16]. And an all-glass pressure sensor has been reported utilizing this glass 3D printing method for high temperature application [5]. In this letter, an integrated additive and subtractive manufacturing (IASM) process is employed to embed intrinsic Fabry-Perot interferometer (IFPI) optical fiber sensor into 3D printed glass components for harsh environment application. The glass 3D printing technique allows rapid fabrication of desired glass structures. The integrated picosecond laser subtractive machining process, which presents higher fabrication accuracy, helps to create channel to place optical fiber sensor at desired position. And 3D printing process continues to seal optical fiber inside the channel and realize fabrication of the information integrated glass module. IFPI in-situ monitoring of glass 3D printing during embedding sensor process is presented. High temperature response and long-term stability at high temperature of this glass module is studied.

Figure 1 schematically illustrates the IASM fabrication process. Several glass layers are 3D printed on a fused silica substrate first, as shown in Fig.1(a). During each printing layer, glass pastes [16] are extruded following a line trace through an extruder (eco-Pen300, Preeflow). And CO₂ laser irradiation (with wavelength of 10.6 μm, ti100W, Synrad) is conducted with optimized output power, scanning speed and spot size, for paste melting both in printing layer and between adjacent layers [5,16–18]. The glass 3D printing assisted with CO₂ laser direct melting procedure was described in details in our previous publications [5,16]. After several layers of glass have been printed, an integrated picosecond laser subtractive micromachining system (Olive-1064-10; Attodyne, Inc. Toronto, ON, Canada, with wavelength of 1064nm) is applied for fabrication of the channel to place optical fiber. In this IASM system, glass 3D printing process presents the dimension accuracy on the order of tens or hundreds of microns. The addition of laser micromachining allows the fabrication of structures with micron dimension accuracy, showing the unique advantage of IASM system in high dimensional accuracy compared with traditional 3D printing process. More importantly, the high dimensional accuracy is necessary in embedding sensor process for sensors with micron size features. After the channel is fabricated, optical fiber IFPI sensor fabricated by femtosecond (fs) laser is placed and fixed inside the channel [19], as shown in Fig.1(c). Fig.1(d) shows the glass 3D printing for embedding optical fiber process. To ensure the fiber sensor is not damaged during the CO₂ laser melting process, multi-layers of glass pastes will be extruded on top of the embedded fiber sensor first. Then CO₂ laser irradiation with the same scanning parameters in previous glass 3D printing process [5,16] will be applied to melt paste layers and bond the optical fiber with the surrounding glass, to ensure optical fiber being fixed and embedded.

Figure 1.

IASM for information integrated glass structure fabrication process. (a) Glass 3D printing on top of glass substrate. (b) integrated ps laser subtractive micromachining for micro-channel fabrication. (c) fiber-optic sensor placed and fixed inside the channel. (d) glass 3D printing conducted on top of the channel with optical fiber inside.

Fig.2 illustrates the fabricated information integrated glass structure. After five glass layers (with a thickness of ~500 μm, and length of 2 cm) have been printed and melted on top of glass substrate, ps laser micromachining was applied to fabricate sensor placement channel with a dimension of 150 μm x 300 μm. Fig.2(a) shows the IFPI sensor, which was fabricated by femtosecond laser micromachining, was placed and fixed inside the ps laser fabricated V-groove channel, as shown in Fig.2(b), where the width at top was ~150 μm. To ensure IFPI region was placed inside the channel, a red laser pointer connecting to the sensor was used for IFPI location visualization. As shown in Fig.2(a), two red light spots are observed to indicate the locations of two IFPI reflection mirrors. Then two layers of glass paste were extruded on top of channel with fixed sensor. Finally, CO₂ laser irradiation process was conducted following the channel direction for the top paste melting, the melting between top paste and adjacent layers, and the bonding between paste and fiber sensor. Fig.2(c) shows the SEM image of cross-section of the glass module, where optical fiber is surrounded by 3D printed glass. Additionally, the large void underneath fiber was the unfilled region of the channel, which can be avoided through optimized ps laser ablation parameters. As shown in Fig.2(d)–(e), the interface between the optical fiber and 3D printed glass and no gap between them are clearly observed, indicating that the optical fiber was undamaged and maintained its integrity. Additionally, porous glass structure surrounding optical fiber was observed.

Figure 2.

(a) IFPI sensor fixed inside the (b) ps laser micromachined channel on 3D printed glass structure, where two red dots indicate the IFPI location. SEM image of (c) the cross-section of information integrated glass module (d) the embedded optical fiber region and (e) zoomed-in image of the rectangular area in (c) showing no gap and interface between the optical fiber and 3D printed glass.

During the CO₂ laser melting process, the embedded optical fiber was connected to one optical sensing interrogator (SM125, Micron Optics) for in-situ spectra measurement. SM125 was controlled by a LabVIEW program to record spectra and calculate the IFPI cavity length in every 1.2 seconds. Fig.3(a) shows the cavity length change during CO₂ laser sintering process, where the output power of CO₂ laser was set to 11 W, and the scanning speed for sintering process was 1 mm/s. It is observed that IFPI cavity length was varying rapidly during sintering process, indicating the interaction between the pastes and laser irradiation (e.g., volume shrinkage of the glass pastes and melting of polymers of the glass paste), as well as the interaction between the paste under laser irradiation and optical fiber. After CO₂ laser sintering process and allowing the glass module to return to room temperature, cavity length was almost the same as the one before sintering process, showing that the IFPI was not damaged after the CO₂ irradiation process.

Figure 3.

(a) In situ monitoring of the cavity length variation of embedded IFPI sensor during CO₂ laser melting process. (b) Interference spectra of the IFPI sensor before (red) and after (blue) sensor fabrication process, acceptable optical loss but no distortion of optical signal is observed.

After the first CO₂ laser irradiation process, to make sure that the glass pastes were melted and good bonding was realized between optical fiber and melted fused silica, the second laser sintering process was conducted, with a lower laser output power (10 W) but the same scanning speed and spot size. Similar to the phenomena existing in the first sintering process, cavity length decreased first, then increased above the initial cavity length, and finally returned to almost same cavity length as the one before first sintering. Additionally, Fig.3(a) shows that the largest cavity length deviation during second sintering process was smaller compared to the one during the first sintering process. This observation should be related to a lower laser output power, smaller printed glass volume change in the second laser irradiation process after the formation of denser fused silica glass and the accomplished bonding between glass and optical fiber realized in the first one.

Fig.3(b) shows the spectra of embedded IFPI before and after laser sintering processes. After the glass module fabrication procedure, acceptable optical loss and no distortion of the optical signal were observed. The loss should be related to optical fiber micro-bending that may exist during the laser irradiation process. Considering the constant laser output during laser irradiation process, optical fiber would be under abrupt temperature change at the laser scanning starting and ending points, where optical fiber micro-bending may be easily formed. And scanning areas except starting and ending points were under same CO₂ laser irradiation profiles and were without sudden temperature change. As such, IFPI region located in the center of scanning path was not damaged and IFPI patterns were not destructed, as shown in Fig.3(b). In addition, rough surface of printed porous glass surrounding optical fiber may contribute to optical fiber micro-bending and thus the optical loss.

The optical losses could be further decreased by optimizing the CO₂ laser scanning parameters and processes. For example, two-step laser irradiation processes can be conducted, that is, preheating the pastes with a low laser output power first and repeating the process with increased laser power. Nevertheless, the clean interference pattern with a fringe visibility of about 16 dB shown in Fig.3(b) is adequate for most sensing applications.

To demonstrate the feasibility for high temperature sensing applications, the fabricated information integrated glass module was placed into an electrical tubular furnace and the interference spectra were monitored when temperature varied from room temperature to 1000°C, at a step of 100°C. At each step, 10 min waiting time was set to ensure that furnace’s temperature was stabilized. As the temperature increased, the cavity length increased, and the spectra were recorded by SM125. Cavity length shift as a function of temperature change was plotted in Fig.4. Fig.4 also presents the cavity length change during two temperature test cycles. Linear regression was applied to fit each response curve and the slopes were calculated as the temperature sensitivity, to be 1.46nm/°C during the two temperature cycles. Experimental results show good repeatability of the fabricated glass module for high temperature measurement.

Figure 4.

Temperature response of the information integrated glass module during two temperature cycles.

When information integrated glass module is under temperature variations, both the refractive index of the core (n_core) and length (L) of embedded IFPI will change due to the thermo-optic effect and thermo-expansion effect, respectively. The change in optical path difference (OPD) of the IFPI can be expressed as [8]:

$$\Delta OPD=OPD\left({\alpha }_{CTO}+{\alpha }_{CTE}\right)\Delta T$$ (1)

where α_CTO and α_CTE are the coefficient of thermo-optic (CTO) and coefficient of the thermo-expansion (CTE) of the optical fiber with typical values of α_CTO=8.3x10⁻⁶°C⁻¹ and α_CTE=0.55x10⁻⁶°C⁻¹, separately. Taking the IFPI’ s cavity length of 148123.5 nm at 100°C into consideration, the bare IFPI temperature sensitivity is calculated according to Eq. (1) as 1.31 nm/°C, showing a small difference from the measurement sensitivity of 1.46 nm/°C. This difference should be related to thermal properties variation between the bare IFPI and embedded IFPI. Considering α_CTO of silica fiber should not change during the embedment process, the difference between measured and calculated temperature sensitivity corresponds to α_CTE difference between the bare silica fiber and 3D printed glass.

Considering the forming of an integrated part realized by the good bonding between printed glass and optical fiber as shown in Fig.2(b) and (c), CTE of the embedded fiber and printed glass should be the same. And the dominated CTE during tests is the CTE of printed glass that surrounds fiber, which is 1.01x10⁻⁶°C⁻¹. The comparatively larger α_CTE of 3D printed glass should be related to the porous structure due to the smaller laser irradiation power and the possible composite difference between the commercial glass/glass fiber and 3D printed glass [16]. The porosity of glass can be flexibly tuned by controlling thickness of top covering paste layers and precise laser processing parameter settings [17,18]. The linear and repeatable temperature responses indicate good bonding between 3D printed glass and silica fiber was realized and the information integrated all-glass module is functional and suitable for high temperature measurements.

To evaluate long-term stability, temperature response of this module has been continuously monitored for 35 days at 1000°C. As shown in Fig.5, after quick settlement, cavity length deviation continued to drift to −60nm (corresponding to ~40°C deviation) in 35-day period. About 8dB optical loss was observed after 35-day test. The interference fringes were still in good qualify as shown in the inset of Fig.5. This phenomenon could be related to combined effects including property change of 3D printed fused silica at high temperature (e.g., annealing), thermal stability of IFPI fabricated by fs laser [20]. Detailed mechanism deserves future investigation.

Figure 5.

35-day stability test of glass module under 1000°C. (Inset: IFPI waveform after 35 days at 1000°C)

In summary, an information integrated structure with an embedded IFPI sensor was fabricated by the novel Integrated Additive and Subtractive Manufacturing (IASM) method. Glass 3D printing assisted by CO₂ laser direct melting process was applied for flexible 3D structure fabrication. An integrated picosecond laser micromachining system was used for sensor embedded microchannel fabrication with high dimension accuracy. IFPI sensor was embedded in the glass module during fabrication process for in situ temperature monitoring. SEM images and repeatable high temperature responses indicate the bonding between the printed glass and optical fiber, and the feasibility of sensor embedment utilizing IASM method. The proposed information integrated glass module shows the enhanced mechanical robustness, good high temperature response, long term survivability at high temperature and is miniature in size, all of which make it attractive for high temperature harsh environment applications.