Introduction

In micro-electro-mechanical systems (MEMS), glass is one of the primarily used materials due to its remarkable features that comprise high optical transparency, high chemical and thermal stability, high electrical isolation, and biocompatibility [1,2,3,4]. Borofloat glass is one of the types of glass which is being used in various MEMS applications due to its exceptional physical, chemical, and mechanical properties [5, 6]. Other qualities like dielectric characteristics and bonding with silicon are applied in the fabrication of various MEMS sensors, such as pressure sensors and several kinds of actuators [7, 8]. The technology of packaging has turned as a most crucial point for MEMS devices that consist of movable structures. Due to the presence of movable structures, like the radio frequency MEMS (RF-MEMS) switch component and the MEMS gyroscope’s resonator, they necessitate enough activity space for effective working. These structures are also required to move in the vacuum environment, which aids in minimal air damping and extended reliability in performance. Therefore, there is always a need to create a cavity surrounding for the packaging process on the substrate, which affords protection and environmental maintenance [9,10,11]. Glass is commonly used as the substrate carrier, which is crucial in packaging MEMS devices. The creation of deep cavities on the glass offers the required space for the functioning of the movable structures [6, 12].

Bulk micromachining is considered as one of the vital aspects in constantly progressing MEMS [13]. In general, three types of etching processes are considered for bulk micromachining in glass; they are mechanical, dry, and wet etching. Mechanical etching processes comprise traditional drilling with diamond-tipped drill bits, electrochemical discharge machining, ultrasonic drilling, powder blasting, and abrasive jet micromachining [1, 13,14,15]. The mentioned mechanical processes are effective in producing larger patterns; however, the processes limit their applicability when making smaller patterns < 100 µm. Also, creating smooth surfaces is another challenge in these processes [16]. Discussing dry etching, it includes plasma etching. Generating structures with a good aspect ratio is an advantage in dry etching; nevertheless, the process is expensive, and the etch rate is low (< 1 µm/min) [1, 6]. Commenting on wet etching, it is considered as an effective process in glass etching because of its benefits such as simplicity, minimal cost, low surface roughness, and excellent etch selectivity [16, 17]. Hydrofluoric acid (HF) is a commonly used chemical solution to etch the glass in the wet etching process, and it is a toxic, corrosive, and highly aggressive-natured etchant [18]. Due to the corrosive nature of HF, the selection of masking material becomes very crucial [6]. To perform selective etching and avoid aggressive etching in undesired areas of glass, several types of masking layers were exploited in the literature. One of them was photoresist, but its utilization as a masking layer was limited by poor sustainability in HF, resulting in a maximum etch depth of < 50 µm [19, 20]. Moreover, HF resistance photosensitive resist (HFPR) was also utilized as a masking layer while etching the fused silica glass in 49% HF. It resulted in an etch depth of 600 µm in 15 h of etching, but the undercutting was notably high [21]. Other than photoresists, different metal thin films were also employed as masking layers; Cr/Au is one among them. Because of the hydrophilicity of Cr/Au masking layer, HF penetrated and created pin holes on the glass surface, and the maximum reported etch depth was 250 µm [22, 23]. To overcome this issue, a multilayer of Cr/Au/Cr/Au combined with a thick photoresist was also practiced as a masking layer in 48% HF, but only etch depth of 300 µm was reported. It is a complex and time-consuming process [24]. Additionally, an etch depth of more than 500 µm was also reported by optimizing the single layer of Cr/Au thin film with different photoresists as a masking layer in 49% HF [25, 26]. Several other reports also utilized Cr/Au as the masking layer; however, Au is an expensive metal, which increases the cost of fabrication [6, 27]. To reduce the fabrication cost, some cost-effective materials were also used as masking layers. Iliescu et al. utilized Cr/Cu as a masking layer in 49% HF, but the masking layer sustained only for 15 min, and pin holes appeared as time progressed; the reported etch depth was only 100 µm [22]. In another report, Zamuruyev et al. employed a Cr thin film with polyamide tape as a masking layer for glass etching in 49% HF. An etch depth of  ~  300 µm was achieved in 1 h; however, many pinholes appeared on the glass surface. Additionally, the edges of the etched structures were not sharp, with relatively high lateral etching [28].

As mentioned in most of the earlier reports, high concentrated HF (40%, 48%, and 49%) was used for the wet etching of glass [17, 22, 24,25,26,27,28]. Due to its corrosive nature, high concentrated HF damages masking layers and causes severe surface defects, due to which the sustainability of the masking layer decreases. As an alternative, a low concentration of HF can be utilized for wet etching of glass to achieve a smooth surface without creating any surface imperfections or defects. Also, the sustainability of the masking layer can be enhanced in the lower concentrations of HF to achieve deep cavities in glass wafers. In this perspective, Ceyssens et al. used Mo (cost-effective metal) thin film with photoresist as a masking layer for the etching of D263 glass in 25% HF, which resulted in an excellent etch depth of 1.2 mm in 3.5 h, but D263 glass cannot be anodically bonded with silicon due to mismatch in coefficient of thermal expansion [29]. So, this glass cannot be used for packaging but only for microfluidic applications. This study demonstrated that metals other than Au can also be used as a masking layer material to etch glass in low HF concentrations. In this perspective, Cr, a cost-effective metal, can also be used as the masking layer material, and its sustainability can be further explored in low HF concentrations.

Therefore, in the current work, Cr is selected as one of the masking layer materials and low concentrated HF (10%) as an etchant solution. Here, Cr thin film is deposited through DC sputtering at room temperature, 200 °C, and 400 °C, respectively, on three different glass wafers. A thorough study is conducted to analyse the sustainability of the respective Cr thin film combined with positive photoresist as the masking layer during wet etching of the Borofloat glass wafers (HK9L) in 10% HF. The surface of glass wafers with and without masking layer, and edges of the etched cavities are examined further through microscopic studies. Additionally, etching time, etch depth, lateral etching, etch rate, and surface roughness of the glass wafer are inspected.

Experimental

In the present work, 4 inch Borofloat glass wafers (HK9L) (University Wafer Inc.) of thickness 500 µm are utilized. The chemical composition of these glass wafers is 69.13% SiO2, 10.75% B2O3, 10.40% Na2O, 6.29% K2O, 3.07% BaO, and 0.36% As2O3. Here, a Cr thin film with positive photoresist (AZ1512HS) is served as a masking layer. Initially, the glass wafer is placed in the DC magnetron sputtering chamber for Cr thin film deposition, and vacuum is continuously generated. When a pressure of ~ 4.4 × 10–6 torr is attained in the chamber, argon gas is purged with a flow rate of 30 sccm. Then, a Cr thin film of 500 nm is deposited using a 3 inch Cr target. The deposition of Cr thin film is performed in five steps; each step comprises 100 nm deposition and 15 min relaxation time to obtain a thin film of good quality [26]. During the deposition, the maintained working pressure is 20 mtorr, power is 74 W, and deposition rate is 0.7 Å/s, respectively. The thickness of Cr thin film is consistently traced through the in-situ thickness crystal monitor. Cr thin film deposition is performed at various temperatures, namely, room temperature, 200 °C, and 400 °C, respectively, on three different glass wafers. The microstructural features of Cr thin films deposited at various temperatures are inspected through field emission scanning electron microscopy (FESEM, Jeol, JIB-4700F).

Later, hexamethyldisilazane (HMDS) is spin-coated on all three Cr-deposited glass wafers for 30 s at 2000 rpm to promote the adhesion between Cr thin film and positive photoresist (AZ1512HS). Further, the mentioned photoresist is spin-coated five times on HMDS using the same parameters. Afterwards, these glass wafers are baked at 90 °C for 30 min in an oven. For patterning the required structure, the respective glass wafer is aligned under the mask aligner, which comprises the required mask pattern (with 400 and 500 µm diameter circles) and subsequently exposed for 14 s to UV light. Afterwards, the glass wafers are immersed in the developer solution to remove the exposed photoresist and develop the structures. Later, these glass wafers are baked at 120 °C for 1 h in the oven. The thickness of the photoresist layer after baking is 1.8 µm. Then, the glass wafers are cooled to room temperature and further used for Cr selective etching. The Cr etchant solution is made using 4.9 g of ceric ammonium nitrate (H8N8CeO18) and 0.43 ml of perchloric acid (HClO4) dissolved in 50 ml of deionized water. The glass wafers are dipped in the Cr etchant solution and rinsed with deionized water after Cr selective etching. Afterwards, the individual glass wafer is submerged in 10% HF for the glass etching. While etching, the edges and rear side of the wafers are protected from HF solution by using a custom-made sample holder. The respective glass wafer is immediately removed from the HF solution if any deterioration is noticed on the masking layer. Later, the entire photoresist coating is removed by dipping the wafers in acetone. Next, the whole Cr thin film is removed by dipping the wafers in the Cr etchant solution. The stated experimental procedure is schematically represented in Fig. 1. The surface of the glass wafers with and without masking layer, the dimensions of etched patterns, and the etch depth are measured through a 3D laser microscope (OLYMPUS OLS4000) of laser wavelength 405 nm. The surface roughness of the glass wafers after removing the masking layer is inspected through the above-mentioned 3D laser microscope. Also, the surface of the glass wafers after removing the masking layer is examined under scanning electron microscopy (SEM) (Zeiss, Supra 40) to trace the surface defects. A Cr thin film of thickness  ~  10 nm is deposited as a conductive layer for SEM analysis. Hereafter, the glass wafer on which the mentioned studies are conducted using Cr thin film (deposited at room temperature) and photoresist as a masking layer is referred to as sample S1. Similarly, the glass wafer on which the studies are conducted using Cr thin film (deposited at 200 °C) and photoresist as a masking layer is referred to as sample S2; also, the glass wafer on which the studies are conducted using Cr thin film (deposited at 400 °C) and photoresist as a masking layer is referred to as sample S3.

Fig. 1
figure 1

Schematic representation of the experimental procedure

Full size image

Results and discussion

The FESEM images of the deposited Cr thin films deposited at room temperature, 200 °C, and 400 °C are presented in Fig. 2. It is observed that with an increment in the deposition temperature, there is an increment in density and uniformity of Cr thin film. It is dedicated to the fact that the increment in the deposition temperature provides energy to adatoms, thereby increasing their mobility, which leads to an increment in the surface diffusion. The enhanced surface diffusion of adatoms leads to a reduction in intergranular voids, resulting in a dense film with a uniform morphology [30,31,32,33].

Fig. 2
figure 2

FESEM images showing surface of Cr thin film deposited at a room temperature, b 200 °C, and c 400 °C

Full size image

This work demonstrates the sustainability of the masking layer that contains Cr thin film (deposited at various temperatures) with photoresist towards the deep wet etching of Borofloat glass wafers in 10% HF. In this regard, primarily, sample S1 is investigated. While etching sample S1, no masking layer peel-off is observed, which confirms the good adhesion between the masking layer and glass. However, after 70 min of etching, the deterioration of the masking layer is detected. After observing the deterioration, sample S1 is removed from the HF solution and analysed. The optical images of the etched cavities in sample S1 with and without the masking layer are shown in Fig. 3. It is observed that the defects are present on the entire surface of the masking layer (Fig. 3 a and b), and it approves its deterioration. To further analyse the surface of the glass wafer, the masking layer is removed. After the removal, it is perceived that the surface of the glass wafer is equipped with many pinholes and the edges of the etched cavities are deteriorated (Fig. 3 c and d). The attained etch depth is found to be  ~  65 µm. These outcomes conclude that the demonstrated Cr thin film (deposited at room temperature) with photoresist as a masking layer is not suitable for performing deep etching of glass, as it displayed relatively low etch depth and lack of sustainability.

Fig. 3
figure 3

Optical images of the etched cavities a,b with and c,d without the masking layer in sample S1

Full size image

In the next test, a similar etching study in 10% HF is performed on sample S2. In this case, the deterioration of the masking layer is discerned after 210 min of etching, which is higher than in the previous experiment, and it confirms the increment in sustainability. Here too, no peel-off of the masking layer is noticed during etching, that assures decent adhesion between the glass wafer and the masking layer. The optical images of etched cavities of sample S2 with and without the masking layer are given in Fig. 4. The presence of defects is observed on the surface of the masking layer in this case as well; nevertheless, the number of defects is considerably reduced in contrast to the earlier event (Fig. 4 a and b). The surface of the glass wafer is also examined after removing the masking layer, and a relatively minimal number of pinholes are observed; also, a significant decrement in the pinhole formation is noticed compared to sample S1. Furthermore, the edges of the cavities are sharper and well-defined in contrast to the previous test (Fig. 4 c and d). The attained etch depth is  ~  165 µm. The obtained results ensure that the current masking layer behaves relatively better in terms of sustainability and protection than the previous masking layer; correspondingly, it indicates that the increment in the deposition temperature of Cr thin film has a crucial influence on the attained relatively better performance.

Fig. 4
figure 4

Optical images of the etched cavities a,b with and c,d without the masking layer in sample S2

Full size image

Furthermore, the etching study on sample S3 is performed in similar conditions as denoted above; it is found that the deterioration of the masking layer occurred after 380 min of etching, which demonstrates that the etching time has greatly improved compared to the earlier two cases. Here too, the peel-off of the masking layer is unnoticed, which certifies the good adhesion between the masking layer and the glass wafer. The optical images of etched cavities of S3 with and without the masking layer are shown in Fig. 5. A very minimal surface defects are perceived on the masking layer, which are way lesser than in the previous experiments, and the surface of the masking layer is almost defect-free (shown in Fig. 5 a and b). When the surface of the glass wafer is investigated after the removal of the masking layer, it is observed that the surface of the glass wafer is completely free of pinholes. Moreover, the etched structures are smooth, and their boundaries are very sharp and distinct in contrast to the former experiments (shown in Fig. 5 c and d). Also, this masking layer presented an excellent etch depth of  ~  245 µm. Thus, it is evident that the current masking layer offered an improvement in all aspects, such as etching time, sustainability, boundaries of the etched structures, and protection to the glass wafer, as compared to the other two masking layers. From this, it is apparent that the increment in the deposition temperature of Cr thin film impacted the quality of the masking layer and its persistence towards the wet etching of glass in 10% HF to form deep cavities in this current investigation.

Fig. 5
figure 5

Optical images of the etched cavities a,b with and c,d without the masking layer in sample S3

Full size image

In the succeeding study, SEM analysis is also conducted to further scrutinise the surface of the three samples (S1, S2, and S3) and the features of the developed etched cavities; and their corresponding images are displayed in Fig. 6. It is evident from Fig. 6 a and b that the surface of S1 comprises numerous pin holes which are spread all over the surface. Additionally, the edges of the etched cavities are deteriorated, which proves the incapability of the respective masking layer that contains room temperature-deposited Cr thin film with photoresist. Discussing the surface of sample S2, it is apparent that the pinholes are very few, and their number is significantly decreased in contrast to sample S1. Also, the edges of the etched cavities are relatively better in terms of shape and sharpness, as shown in Fig. 6 c and d. It validates the enhanced sustainability of its masking layer, which has Cr thin film deposited at 200 °C and photoresist. In the case of sample S3, it is detected that the surface is completely free of pinholes, and the edges of the etched cavities turned sharper and more well-defined compared to samples S1 and S2, as presented in Fig. 6 e and f. Therefore, it authenticates that its masking layer, comprising Cr thin film deposited at 400 °C and photoresist, acted superiorly against the etchant solution. The accomplished outcomes in SEM images are in decent agreement with laser microscopy images.

Fig. 6
figure 6

SEM images displaying the surfaces and etched cavities of samples S1 a ,b, S2 c,d, and S3 e,f, respectively

Full size image

In this study, it is confirmed that the sustainability of the masking layer is highly dependent on Cr thin film deposition temperature and it increased with an increase in the deposition temperature. It is dedicated to the increment in the density of Cr thin film with deposition temperature (derived from Fig. 2). The relatively high dense Cr thin film effectively serves as a strong barrier, preventing the HF penetration during etching, thereby improving the sustainability of the masking layer, and providing enhanced protection to glass wafer in HF. Therefore, the glass wafer with masking layer comprising Cr thin film deposited at 400 °C (sample S3) exhibited comparatively improved sustainability and defect-free surface.

Additionally, the resulted variation of etching time, etch depth, lateral etching, etch rate, and average surface roughness with respect to the three samples is schematically illustrated in Fig. 7 to represent the effect of Cr thin film deposition temperature on the mentioned parameters in a better manner. Figure 7 (a) shows the increment in the etching time from sample S1 to sample S3. In detail, it states that the etching time (which is directly related to sustainability) increased with the increase in the deposition temperature of Cr thin film. Additionally, Fig. 7 (b) displays the enhancement in the etch depth from sample S1 to sample S3, which shows this enhancement is highly related to the increment in Cr deposition temperature. Thus, it proves that the enhancement in the sustainability of the masking layer ensures a higher etch depth.

Fig. 7
figure 7

Respective plots showing a the variation of etching time, b the variation of etch depth and lateral etching, c the variation of etch rate and d the variation of average surface roughness with respect to three different samples (S1, S2, and S3)

Full size image

Moreover, glass etching is isotropic in nature, due to which lateral etching or underetching also takes place. Lateral etching is defined as the change in the dimensions of the etching window before and after wet etching. It takes place under the masking layer. Also, lateral etching is related to the quality of the masking layer in terms of adhesion between the substrate and the masking layer. The poor adhesion of the masking layer with the substrate results in high lateral etching than etch depth. Ideally, the lateral etching should be equal to the etch depth. From Fig. 7 (b) it is noticed that the lateral etching increases as the etch depth increases, which assures the isotropic etching behaviour in this current work. Furthermore, it is found that the lateral etching is more than the etch depth in sample S1, whereas, it is less than the etch depth in the case of samples S2 and S3. The underetch ratio, which is the ratio of etch depth to lateral etching (one side), is < 1 in sample S1, and it is slightly > 1 in the case of samples S2 and S3 [23, 29]. These results agree with the previous analysis, which demonstrates the improvement in the quality of the masking layer with the increment in Cr thin film deposition temperature.

Moreover, Fig. 7 (c) shows the decrement in the etch rate from sample S1 to sample S3. The reason for the decrement in the etch rate is due to the formation of insoluble by-products during wet etching of the glass. The by-products start settling down inside the etched cavities, and as the etch depth increases, the chances of these by-products coming out from the cavities decrease. As a result, these insoluble by-products remain in the cavities, act as a barrier, and limit the penetration of etchant through these by-products, which consequently causes a low etching rate [16, 27, 34].

The average surface roughness of the etched glass wafers after removal of the masking layer is measured, and its variation with respect to three different samples (S1, S2, and S3) is presented in Fig. 7 (d). It is observed that the average surface roughness decreased from sample S1 to sample S3. This validates that the increment in Cr thin film deposition temperature contributed to the reduction of average surface roughness, depicting the dependency of average surface roughness on Cr thin film deposition temperature. It is observed in Fig. 2 that the density of Cr thin film is enhanced with the deposition temperature. The enhancement in the density of Cr thin film leads to a reduction in the penetration of HF and consequent interaction of HF with the underlying glass surface. This restricts the formation of surface defects (such as pinholes) on the glass surface, and subsequently, the surface roughness decreases, resulting in a smooth glass surface. Therefore, in this study, sample S3 offers a relatively low surface roughness compared to samples S1 and S2.

The current work offers a facile approach by employing a low-cost masking layer (Cr thin film along with photoresist) for the fabrication of deep cavities in glass wafers without any surface defects in low concentrated HF (10%). The fabricated cavities in glass wafers can be used for packaging and encapsulation of various MEMS devices, e.g., micropumps, mechanical inertial sensors, capacitive silicon resonators, and fluidic devices for micro total analysis system (µ-TAS) [35,36,37,38]. The projected method can also be explored for the fabrication of deep microchannels, which can be used in the fabrication of glass-based microchips for chemical synthesis [39] and electrophoresis applications [40]. Further, this method can be used for the fabrication of glass-based microfluidic devices that have vast biological applications such as cell analysis, point-of-care diagnostic, drug delivery, etc. [41].

Conclusion

In this contribution, Cr thin film of 500 nm (deposited at room temperature, 200 °C, and 400 °C, respectively) combined with positive photoresist (AZ1512HS) layer is used as a masking layer for deep wet etching of Borofloat glass wafers in 10% HF solution. The deposition temperature of Cr thin film plays a pivotal part in the sustainability of the masking layer in the etchant solution. The sustainability displays an increment trend with the enhancement in deposition temperature of Cr thin film. Consequently, the glass wafer comprising Cr thin film deposited at 400 °C and photoresist as the masking layer exhibits relatively higher sustainability that out-turns into longer etching time (of 380 min) and deeper etching (etch depth of  ~  245 µm). Moreover, the stated masking layer yields defect-free and smooth surface, sharp and well-defined etched cavities in the glass wafer and stands as the best masking layer among all the three masking layers employed in this work. The current work opens a new window for producing deep cavities in glass wafers by utilizing the demonstrated cost-effective masking layer with low-concentrated HF etchant solutions.