Laser-welded elastomeric valves for the rapid fabrication of flow control in microfluidics

1 Introduction

Laser transmission welding (LTW) using a nanosecond pulsed laser has been investigated for welding and sealing microfluidic devices, demonstrating its effectiveness (Boglea et al., 2007; Bishayee et al., 2025). Conventional bonding methods, such as thermal bonding and ultrasonic bonding, cause heating of a larger area, leading to bonding over a larger surface area. In this study, LTW of cyclic olefin copolymer (COC) is applied to manufacture and seal a representative microfluidic device, specifically a pumping valve, to evaluate the suitability of the LTW process for producing such devices. A demonstrator is fabricated in which LTW is used to bond an elastomeric valving membrane (substrate) to a black COC substrate. Subsequently, this valve is functionally tested. To that end, Section 2 first presents and discusses the design of a valving microfluidic device. Next, in Section 3, the materials and technology used to manufacture the device are presented. Moreover, the setups and methods used to characterise and verify the functionality of the pumping valve are presented. Finally, the performance and characteristics of the device sealed using LTW are assessed, and the results are discussed.

2 Microfluidic valve

A valve is a mechanical device used to regulate, direct, or stop the flow of liquids, gases, or other materials within a system. It works by opening, closing, or partially obstructing a passageway, allowing control over the amount of material (liquid, gas, or other materials) that flows through and the direction of flow. Microfluidic devices rely on precise fluid handling; regulating flow within the device channel is essential to their proper functionality. Valves act as control points to ensure that the flow is managed smoothly, safely, and efficiently (Kim et al., 2018; Li and Dong, 2024; Stone and Microfluidics, 2001; Pandey et al., 2018; Yeo et al., 2011). In microfluidic systems, flow control can be achieved through capillary action, the use of pumps to drive fluids, and valves to regulate fluid pathways by toggling them on and off (Li et al., 2021; Lee et al., 2026; Guo et al., 2010). Valves are also used to control the amount and direction of flow through microfluidic channels (Li et al., 2021; Guo et al., 2010).

Microfluidic valves fabricated in glass and silicon commonly rely on deformable polymeric materials to achieve flow control. For example, polydimethylsiloxane (PDMS) is frequently employed as a compliant layer that deforms under applied pressure to provide valving functionality. Alternatively, latex membranes sandwiched between glass layers can be pneumatically actuated to regulate fluid flow, as demonstrated by Wang et al. (2019) and Han et al. (2008). However, these approaches require the integration and bonding of dissimilar materials, which increases fabrication complexity and may affect manufacturing scalability. Hydrogel-based microvalves have also been reported, such as the design presented by Obst et al. (2020), in which hydrogel beads are locally heated using a printed circuit board (PCB). The hydrogel contracts upon heating and expands during cooling, enabling active control of fluid flow within a microfluidic channel. However, the need for integrated heating elements further increases device complexity. In addition, the temperature changes required for valve actuation may restrict the range of applicable fluids and biological or chemical processes, as many systems contain temperature-sensitive reagents, cells, or biomolecules that must remain compatible with the imposed thermal conditions.

For thermoplastic-based microfluidic devices, elastomeric membranes are among the most used valving strategies and are typically integrated through lamination, thermal bonding, or ultrasonic bonding (Childers et al., 2024). Although these approaches are compatible with thermoplastic manufacturing, the bonding process requires careful optimization of temperature, pressure, and bonding time to prevent membrane sagging, wrinkling, or deformation. Consequently, significant process development is often required to achieve reliable and reproducible valve performance, increasing both prototyping time and manufacturing complexity.

The valving mechanism demonstrated and studied in this study is based on an elastomeric membrane, which is expanded using pneumatic actuation from one side (Kim et al., 2012). Figure 1a shows a schematic cross-section of the valve under consideration when the elastomeric membrane is not expanded (open state), allowing flow through the device. When the elastomeric membrane is expanded, it closes the flow path (closed state) and, thus, shuts off the microfluidic device (see Figure 1b). Upon switching off the actuation, the elastomeric membrane relaxes and the flow (re)starts, as shown in Figure 1a. Figure 2 shows the schematic top-view of the valving device, with a circular valve seat (pocket) of 5 mm diameter and the inlet/outlet holes for the fluid flow.

The valve seat depth and the diameter of the valve seat are design parameters of this microfluidic valve design. The valve seat depth determines how much the elastomer needs to stretch to actuate and close the flow. If this dimension is too large, it may exceed the stretchability of the elastomer and lead to failure to block the flow due to the elastomer’s permanent (inelastic) deformation. In the current design, a valve seat depth of 0.3 mm was chosen (see Figure 1a).

The valve seat diameter also defines the diameter of the elastomeric membrane, which (again) contributes to stretching to close the valve and relaxing to open the valve. The effective spring constant of the membrane is inversely proportional to this diameter, thus impacting the force required to deform the membrane. If this diameter increases, the membrane may sag, and if it is too small, the pressure or force required to actuate the membrane increases to an extent that may lead to a tear in the membrane. In the current design, a valve seat diameter of 5 mm was chosen, as can be observed in Figure 2.

3 Materials and fabrication

3.1 Materials

The microfluidic valve under consideration consists of several substrates (see Figure 1). It consists of a COC plate, an elastomeric membrane, and a layer of adhesive tape. The fluidic circuit and the valve seat of the microfluidic device were fabricated from cyclic olefin copolymer (COC) plates (TOPAS 6013, TOPAS Advanced Polymers GmbH, Germany) measuring 130 mm × 130 mm and 1 mm thickness. Typically, black COC, referred to hereafter as COC_black, contains 2 wt.% carbon. These carbon additives are highly absorptive at the laser wavelength of 1,064 nm used to weld the substrates (Bishayee et al., 2025). The channels on the back face of the device were sealed using 25 µm-thick transparent transferable adhesive tapes (ARcare 8026, Adhesives Research Ltd., Ireland). The elastomeric valving membrane consists of 100 μm-thick elastomeric COC (eCOC) foils (TOPAS E140, TOPAS Advanced Polymers GmbH, Germany). This membrane was pneumatically actuated to open and close the flow.

3.2 Fabrication

COC_black substrates measuring 15 mm × 30 mm were cut from the larger sheets to facilitate easier handling and to ensure that they fit properly into the LTW clamp tool. For this study, a computer numerical control (CNC)-milled (Datron Neo, Datron AG, Germany) was used to fabricate the valve seats, the channels, and the inlet and outlet holes in the 15 mm × 30 mm COC_black parts. The inlet and outlets are 1 mm in diameter, and the valve seat is 5 mm in diameter and 300 µm deep and was milled on the top face of the substrate. The channels, being 1 mm wide and 200 µm deep, were milled on the back face, as shown in Figure 1a. The backside of the chip, containing the channels, is sealed using transparent adhesive tape ((ARcare 8026, Adhesive Research Ltd., Ireland), while the valve-seat side is covered with eCOC elastomeric foil and welded by laser transmission welding (LTW). Adhesive tape was chosen for sealing the channels because it is a well-established and reliable method in microfluidic device fabrication.

Figure 3 shows microscopy images of the produced valving microfluidic device. The laser processing conditions—i.e., the processing window—for nanosecond pulsed LTW were studied in our earlier work (Bishayee et al., 2025). Based on that study, the laser processing parameters employed to weld the eCOC and the COC_black were set to a laser wavelength of λ = 1,064 nm, laser pulse energy of E _p = 23.4 mJ, laser pulse repetition rate of f _p = 100 kHz, focal spot diameter of d _laser = 50 μm, and a scan velocity of the laser spot relative to the substrates of v = 200 mm/s. To improve structural stability and reduce shear stress on the sealed weld during handling or insertion into the clamping setup, an additional laser welding contour is applied. This contour was fabricated using a pulse energy of E = 11.7 mJ, ensuring that it does not contribute to the sealing performance of the microfluidic device and serves solely as mechanical reinforcement, as shown in Figure 3c.

For comparison, a valving device was fabricated by the conventional bonding method of thermal bonding. A COLLIN Press P200S (COLLIN Lab & Pilot Solutions GmbH, Germany) was used, in which the black COC with the valve seat and the eCOC foil placed over the valve seat were stacked and sandwiched between two aluminium plates. The pressure was set to 20 N/cm², and the temperature of the plates was set to 103 °C with a hold time of 20 min.

The net processing time associated with LTW is substantially lower than that of thermal bonding for valve fabrication, with LTW requiring approximately 2 s compared with approximately 20 min for thermal bonding.

4 Analysis tools

4.1 Confocal microscope

A Keyence 3D laser scanning microscope (Keyence VK-X3000, Keyence International, Belgium) was used to acquire topographical information on the welded transparent elastomeric COC valve after laser welding.

4.2 Microfluidic test setup

Only the microfluidic valve, sealed by LTW, is subjected to a microfluidic performance test. The performance of the microfluidic device is being assessed based on its durability and reliability by actuating the valve pneumatically for more than 300 cycles. Figure 5 shows a schematic overview of the fluidic circuit depicting all the elements and components for testing the microfluidic device. Here, a flow unit (Fluigent Smart Microfluidics, Germany) is integrated into the setup to monitor and confirm the switching of fluid flow during each cycle. The recorded data from the flow sensors provides insight into the device’s functional longevity, precision, and consistency over repeated use.

As a test fluid, deionised water (DI water) is used. A pressure control unit (Microfluidic Flow Control System, MFCS™ series, Fluigent Smart Microfluidics, Germany) maintains a constant pressure of 200 mbar in the fluid reservoir. The same unit is used to create an actuating pressure of 700 mbar to close the elastomeric membrane and stop the flow of the microfluidic circuit. A flow unit sensor connected to the output of the device records the flow exiting the microfluidic circuit. The device holder, as shown in Figure 4, is used to clamp the laser-welded valve and the PDMS pneumatic gasket to maintain pneumatic sealing.

As depicted in Figure 5, a computer is connected to the pressure control unit and the flow unit, which maintain the set pressure in the reservoir and control the flow sensor, respectively. The computer is used to set different pressures for flow and actuation. The reservoir is pressurised to maintain a constant flow pressure and is connected to the inlet of the microfluidic valving device. The valve in the microfluidic device is actuated pneumatically with a high pressure of 700 mbar to stop the flow and a low pressure of 0 mbar to open the fluidic circuit. The outlet of the device is connected to a waste reservoir via a flow unit that measures the flow in the outlet. The output of the flow sensor is recorded along with the flow pressure and the fluctuating actuating pressure for the valve.

5 Results and discussion

5.1 Microscopy and confocal measurement

Microscopy images of the laser-welded and thermally bonded joints showed no significant visual differences in appearance. However, LTW welds only along a circular laser trajectory, whereas thermal bonding welds across the entire COC_black substrate (see Figure 6). In addition, laser welding is considerably faster (about 600 times), requiring only 2 s of processing time compared to approximately 20 min for thermal welding of the valve.

Confocal microscopy of the laser-welded valve (Figure 7) distinctly reveals the weld seams and membrane, with no evidence of significant thermal deformation or membrane wrinkling. The concentric LTW welds appear in the confocal scanning of the central cross-section as bumps (protruding from the membrane surface) measuring ∼80 μm in height; the surface roughness of the membrane is between 4–11 µm.

5.2 Microfluidic performance

Five valving LTW devices were actuated repeatedly to evaluate their performance over 360 cycles (each) of pneumatic actuations. In other words, the actuation of the valve(s) was performed pneumatically by alternating pressurisation and depressurisation using air, with each cycle including a 30-s hold time per cycle. The actuations were continued for 3 h, resulting in 360 actuations. The valve was found to be fully functional after more than 360 cycles, indicating good structural and mechanical durability. In other words, none of the five devices tested failed. To verify the latter, the same valves were used to control fluid flow across eight cycles, during which actuation pressure, fluid pressure, and flow rate were recorded. The results of one of these measurements are shown in Figure 8. As can be observed from this figure, when the valve was in the open state (no actuation pressure applied), an average flow rate of`∼3,500 μL/min was recorded by the flow unit, with a standard deviation of ±1% of the peak flow rate, driven by a flow pressure of 200 mbar. This is represented by the blue and orange curves in Figure 8. Upon applying a pneumatic actuation pressure of 700 mbar (grey curve), the valve closed completely, and the flow sensor detected 0 μL/min, confirming effective flow blockage. The primary limitation in extending the duration of these actuations was the reservoir’s 20 mL capacity (see Figure 5). In the first actuation cycle, the measured peak flow rate (blue) and the corresponding peak actuation and flow pressures (orange) are noticeably higher than in cycles 2–7. This behaviour is commonly observed with the type of flow and pressure sensors used during the first actuation. Here, the flow meter often exhibits an initial transient overshoot before the signal stabilises. This can occur as the sensor’s internal measurement elements settle and the fluid pathway reaches steady-state conditions. As a result, the recorded flow and pressure signals during the first cycle tend to appear elevated relative to those during subsequent cycles. A similar, but smaller, transient effect also occurs at the start of each new actuation and can therefore be interpreted as sensor-related noise rather than a true change in system performance.

Across the entire sequence of eight cycles, a gradual increase in peak actuation pressure is observed. This progressive increase is likely related to cumulative effects within the system—such as minor changes in fluid resistance, pressure propagation, or sensor drift—rather than a change in the underlying actuation mechanism. After eight rounds of actuation, the fluid level in the reservoir was significantly depleted. However, no delamination of the laser-transmission-welded eCOC valve was observed following the actuations.

6 Discussion and conclusions

In this study, LTW of black COC polymer was used to fabricate a demonstrator for a valving microfluidic device, where the flow of fluids can be controlled by pneumatically actuating an eCOC membrane over a valve seat. In other words, the elastomeric membrane was bonded over the valve seat using laser transmission welding. For comparison, the same valve seat and the elastomeric foil were also bonded using thermal bonding, which is an established process for bonding polymer microfluidic devices. Next, the laser-welded valving device was characterised to verify its functioning with 360 pneumatic actuation, which exceeds the typical test requirement of actuations from a valving device.

It was shown that the use of LTW for bonding elastomeric thin films is a suitable method for fabricating a functional valving microfluidic device. The net processing time for laser welding an eCOC membrane with a 5 mm diameter valve seat takes less than 2 s. Except for protrusions along the welded contour, the laser welding process causes no visible (thermal) deformation of the elastomeric thin film.

The confocal analysis indicates that weld-induced surface protrusions (∼80 µm) are predominantly located outside the active sealing region defined by the 5 mm valve seat. The measured membrane roughness (4–11 µm) is small compared to the seat depth (∼0.3 mm) and is therefore not expected to significantly affect sealing performance. This is consistent with the low leakage observed during valve closure. However, the potential influence of these features on long-term sealing behaviour under extended cyclic loading warrants further investigation.

The results of this study demonstrate the potential of LTW as an effective manufacturing approach for the rapid fabrication of thermoplastic microfluidic valves. Compared with conventional valving technologies that often require multiple materials, complex assembly steps, and extensive process optimisation, LTW enables fast and precise device fabrication using a streamlined manufacturing workflow. This significantly reduces prototyping time and accelerates process parameter optimisation, thereby shortening the development cycle from laboratory-scale proof-of-concept devices to scalable manufacturing. From a lab-to-fab perspective, this facilitates the rapid translation of new microfluidic designs into production-ready devices. Although the initial investment required for laser welding equipment remains a challenge for industrial adoption, the reduction in fabrication complexity and development time can provide substantial benefits in high-volume manufacturing environments. Furthermore, the significance of these findings extends beyond the specific valve design investigated in this work. Since the LTW process is applicable to a wide range of thermoplastic polymers, provided that a suitable laser-absorbing counterpart is incorporated, the approach offers a versatile platform for the fabrication of diverse microfluidic components and systems.

To further strengthen the claim regarding the functionality of the LTW-fabricated microfluidic valve, a more in-depth comparison with thermally bonded devices would be beneficial. Thermal bonding is a well-established and widely used method for fabricating and sealing microfluidic valves; however, a rigorous, head-to-head quantitative comparison would require the design and fabrication of fully equivalent devices to ensure a fair and controlled evaluation. Such an approach would involve additional effort in device optimisation, fabrication consistency, and standardised testing protocols, which were beyond the scope of the present study. The primary objective of this study was to establish LTW as a rapid and viable fabrication method and to validate the functionality and operational performance of LTW-based valves. Consequently, systematic benchmarking against thermally bonded counterparts—using metrics such as bond strength, leakage, burst pressure, or long-term reliability—was not performed. Although stable operation was observed up to 360 cycles, extended high-cycle testing to failure, coupled with detailed analysis of membrane fatigue, weld integrity, and material creep, will be required to fully define the operational envelope and long-term reliability of LTW-fabricated valves.

Future work will, therefore, need to focus on a comprehensive comparative analysis to more rigorously assess the relative advantages of LTW over conventional thermal bonding techniques and incorporate failure mode analysis for the valves.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Stefan Köstler, elyte diagnostics GmbH, Austria

Reviewed by: Siqin Dong, Zepto Life Technology, United States

Wonhyung Lee, Kumoh National Institute of Technology School of Mechanical Engineering, Republic of Korea