Paper is an inexpensive and ubiquitous resource that has been used in various applications for a long time (L Santana and Angela A Meireles, 2014; Kumar Gupta et al., 2019). Its properties (e.g., porosity, chemical composition, and wetting performance) are readily adjustable for different purposes (Glavan et al., 2013; Böhm et al., 2014; Gao et al., 2019b). Like other porous materials, the porous nature and high surface-to-volume ratio of the paper promote passive fluid driving and control. The fiber chemical composition (e.g., degree of polarity) can also be modified to enhance sample-paper interactions and plays a key role in device design and operation (Chitnis et al., 2011; Lim et al., 2019; Ma et al., 2019; Soum et al., 2019). Owing to a wide variety of paper types commercially available in the market, the correct property selection also saves time and labor for material treatments (Tomazelli Coltro et al., 2014; Xia et al., 2016). For example, nitrocellulose paper serves as a good substrate for covalent immobilization of the biomolecules due to the strong binding capability to proteins originated from the nitrate groups on their surfaces. Filter paper and chromatography paper can outperform other paper types in terms of uniform thickness and pore size (Tang et al., 2022).

Note that the paper material per se only provides the backbone of the devices, while analytical analysis taking place on papers is realized through incorporation of various sensing or detection methods (Adkins et al., 2015; Nishat et al., 2021). Existing detection methods can be categorized into several types including colorimetric, fluorescent, chemiluminescent, electrochemical, electro chemiluminescent and Raman sensing (Fu and Wang, 2018; Kaneta et al., 2019; Zheng et al., 2021a; Li et al., 2021). Review articles for in-depth discussions on advances in µPADs can be found in the following references: (Carrell et al., 2019; Kaneta et al., 2019; Lim et al., 2019; Soum et al., 2019).

The µTADs are another successful application of porous materials for environmental monitoring and general analytical analysis (Agustini et al., 2016; Tan et al., 2021). Similar to µPADs, these devices are good candidates for low-cost applications. The existing industry worldwide also promotes the applications without complex material modifications (Farajikhah et al., 2019; Weng et al., 2019). Currently, a variety of threads are available for different applications, including natural (e.g., silk, wool, linen, etc.) and synthetic (e.g., polyester, polyether-polyurea, acrylic, etc.) threads (Oliveira et al., 2022). The flow characteristics and the detection methods employed in threads are similar to those employed in paper, since both are porous (Berthier et al., 2017; Tan et al., 2021). However, compared to paper devices, thread-based devices are more suitable for wearable applications since threads can be used to create clothing either by directly sewing, or having walls patterned onto cloth (Xiao et al., 2019; Tan et al., 2021; Xia et al., 2021).

The detection methods used in µTADs are similar to those used in µPADs. Conventional detection methods (e.g., fluorescence, electrochemical, Raman, etc.) are applicable to thread-based devices as well (Weng et al., 2019; Agustini et al., 2021). Moreover, distance and barcode-based detection are another two possible low-cost detection strategies (Tan et al., 2021). Distance-based detection relies on the fact that disparate wetting performances can be induced by different analytes for identification (Alsaeed and Mansour, 2020; Jarujamrus et al., 2020; Shimazu et al., 2022). Moreover, barcode detection can provide results of multiple analyte reactions (e.g., blood typing) that otherwise are difficult to achieve (Nilghaz et al., 2014). For a comprehensive review on thread devices, the readers are encouraged to read the suggested references: (Farajikhah et al., 2019; Weng et al., 2019; Tan et al., 2021; Xia et al., 2021).

Similar to paper and thread, cloth also has a porous structure, thus most fabrication and analytical approaches used in the aforementioned porous materials can also be extended and exploited (Zheng et al., 2021b; Xu et al., 2021). Colorimetric method is the most popular method used in µCADs due to its simplicity and independence on external analysis tools (Bagherbaigi et al., 2014; Nilghaz et al., 2015; Li et al., 2018; Tasaengtong and Sameenoi, 2020). Electrochemical and chemiluminescence methods and their combination were explored as well (Jiang et al., 2020; Shang et al., 2020; Zheng et al., 2021b; Shang et al., 2022). Readers are encouraged to read more detailed review papers that summarizes fabrication, detection methods and performances of µCADs (Nilghaz et al., 2013; Zhang et al., 2020; Agustini et al., 2021). In addition, other low-cost materials have also been reported. For example, sponge was used for thedetection of heavy metal ions in environmental samples (Ding and Lisak, 2019), leveraging the strength of sponge structure and the coupling with other materials for better mechanical properties (Hu et al., 2022; Silva et al., 2022). A few examples of the applications of low-cost porous materials are shown in Figure 3.

Polymer is a type of material that consists of large molecules called repeating units (or mer) arranged in a periodic manner within the structure (Strobl and Strobl, 1997; Young and Lovell, 2011). Nature has generously provided us many polymeric materials such as wood and rubber (Carothers, 1936; Mark et al., 2004). The paper cellulose described above is indeed a type of polymer composed of glucose units (El Seoud and Heinze, 2005; Rose and Palkovits, 2011). Moreover, many synthetic polymers have been recently invented for various purposes (Hacker et al., 2019). For example, plastic is a large family of polymers, including polycarbonate (PC), polyethylene (PE), polypropylene (PP), polyethylene terephthalate (PETE or PET), polyvinyl chloride (PVC), acrylonitrile-butadiene-styrene (ABS) and many others (Young and Lovell, 2011; Maitz, 2015; Hacker et al., 2019). Oftentimes, to create microfluidic devices, the associated cost does not come from the materials themselves since they are cheap, instead, the fabrication methods such as photolithography that creates polymeric structures are responsible for the high cost (Chan et al., 2015; Faustino et al., 2016). In particular, PDMS is a popular polymeric material used in microfluidics (Haubert et al., 2006; Li et al., 2012; Raj M and Chakraborty, 2020). It offers several advantages over other materials such as cost-effectiveness, good biocompatibility and transparency, favorable elasticity and flexibility, inertness to chemicals and permeability to gases (Raj M and Chakraborty, 2020; Miranda et al., 2021). To create PDMS based devices, soft lithography has been considered as a gold standard. Specifically, a mold with desired pattern is created first, and then the PDMS mixture is poured onto the mold allowing the curing over time to create PDMS replicas with identical patterns. Though the method itself is low-cost, the molds are made from complex conventional photolithography, for which a cleanroom is indispensable (Tomazelli Coltro et al., 2014; Barocio et al., 2021). To reduce the cost and eliminate the needs of a cleanroom, other fabrication methods such as 3D printing and milling have been explored for mold manufacturing (Gale et al., 2018; Ruiz et al., 2020).

PMMA is another popular polymeric material used in microfluidics (Chen et al., 2019; Ma et al., 2020; Razavi Bazaz et al., 2020). As a thermoplastic, PMMA becomes pliable when heated up above the glass transition temperature. Therefore, similar to PDMS, PMMA devices can be made by molding, thus holding promise for mass production (Trotta et al., 2018; Ma et al., 2020). In addition, PMMA can be used as an UV-sensitive material on which the structures are created by the UV radiation (Fan et al., 2012). Thin plastic films such as the double sided tapes, PET films are also explored to create lab-on-a-foil devices (Focke et al., 2010; Bertana et al., 2018). Unlike the porous materials described above, the devices made on thin films are much similar to regular PDMS devices, on which microchannels can be created and active fluid and particle manipulation technologies can be integrated (Gale et al., 2018). Another important polymer that has been widely used nowadays are the photosensitive resins used in 3D printing techniques. Note that although traditional 3D printing resins possess good mechanical and physical properties, limitations still exist in terms of the molding performance if used as the molds and the biocompatibility for biology and medical purposes (Heuer et al., 2021). Other issues such as flow control issues, channel dimensional accuracy, solvent compatibility, surface roughness and low wettability are still the major concerns for broader applications (Gale et al., 2018; Mehta and Rath, 2021), though several studies have reported novel photopolymer formulations (resins) capable of potentially addressing these issues (Pranzo et al., 2018; Mehta and Rath, 2021). More comprehensive reviews of 3D printing materials for microfluidic devices can also help the readers understand the current status and future perspectives for this hot field (Gale et al., 2018; Weisgrab et al., 2019).

Since µPADs, µTADs and µCADs have similar structures, they do share similar fabrication approaches (Dou et al., 2015; Tan et al., 2021). One of the straightforward approaches is to cut paper into strips with desired dimensions, followed by loading essential reagents (Adkins et al., 2015; Akyazi et al., 2018; Weng et al., 2019). Here, the capillary action serves as the driving pump to spread samples from one end to another (Xie et al., 2019; Nishat et al., 2021). Hydrophobic fluid barriers can also be used to control the fluid flow in porous devices, turning a single piece of paper into a fluid managing platform (Soum et al., 2019; Liu et al., 2020; Liu et al., 2021). Different printers can be used to create hydrophobic walls that confine the fluid transport in between, most printers are simple, inexpensive, and suitable for large volume creation (mass production) (Tomazelli Coltro et al., 2014; Adkins et al., 2015). Inkjet printing is a popular printing fabrication technology, consisting of two main categories: powder based, or photopolymer based (Loo et al., 2019; Olmos et al., 2019; Aladese and Jeong, 2021; Khorsandi et al., 2021). Using this technique, hydrophobic inks are used to create the channel walls on paper-based devices (Fan et al., 2018; Lim et al., 2019; Wang et al., 2019). Note that inkjet printing is applicable for multiple types of paper, while laser printers provides rapid and large volume printing processes (Bamshad and Cho, 2021; Nishat et al., 2021).

In addition, flexographic printing has also been used for creating µPADs. This method provides a continuous nature of fabrication, which is critical in mass production (Caetano et al., 2018; Qian et al., 2022). The screen printing process has also been used yet it requires multiple steps and has low resolution (Morbioli et al., 2019; Soum et al., 2019; Dixon, 2020; Yehia et al., 2020). Moreover, wax screen printing is a technology that combines the advantages of wax printing and screen printing, offering a simple 2-step process at much lower costs than traditional wax printing technology (Morbioli et al., 2019). Wax printing has been widely used for creating channel walls in paper-based devices (Liu et al., 2015; Kamnoet et al., 2021), however, there are still some limitations for this technology such as the difficulty to create smaller size channels (Gale et al., 2018; Lin et al., 2022; Tesfaye and Hussen, 2022). The solvents may also soak into the wax and paper boundaries, thus compromising the functionality of the device (Szabo and Hess-Dunning, 2021; Chen et al., 2022a; Ruiz et al., 2022; Tran et al., 2022). Using stamps (ink imprinting) and pen writing (handwriting) are easy yet non-precise techniques to pattern 2D channels (Dornelas et al., 2015; Xia et al., 2016; Noviana et al., 2020). Plasma treatment can also be used to pattern channels using hand held corona treater (He et al., 2015; Zhu et al., 2016). The cross-sectional area of porous devices can be adjusted to control the flow motion (Soum et al., 2019; Modha et al., 2021), it is possible to cut paper and cloth with different inexpensive tools (i.e., scissors, razor blade) (Tomazelli Coltro et al., 2014; Xia et al., 2016). Those tools already exist in commercial versions, coupled to CNC machines and are able to execute a predefined cutting path based on the drawings (Gosset et al., 2018; Cortes-Medina et al., 2020; Guo et al., 2021). Similarly, xurography (digital craft cutter) can be used to cut other materials (e.g., polymeric sheet), as long as the material thickness is small (Speller et al., 2019; Caffiyar et al., 2020; Guo et al., 2021).

3D printing technology has proven to be a cost-effective method for prototyping and engineering studies (Mehta and Rath, 2021; Jin et al., 2022). With the improvements of 3D printers, filaments and CAD technologies, 3D printing has emerged as a great tool to create microfluidic devices (Bhattacharjee et al., 2016; Gale et al., 2018; Mehta and Rath, 2021). Additive manufacturing constructs three-dimensional objects directly from the CAD designs using techniques such as fused deposition modeling (FDM) and stereolithography (SLA) (Gale et al., 2018; Weisgrab et al., 2019; Mehta and Rath, 2021). Moreover, this technique allows for the fabrication of the final enclosed device directly from the resin and also for the development of PDMS molds using specific resins (He et al., 2016; Razavi Bazaz et al., 2020; Mehta and Rath, 2021). The printed parts may also be bonded to other substrates or 3D printed parts using adhesive tapes or treatments such as UV bonding (Bressan et al., 2019; Razavi Bazaz et al., 2020; Wei et al., 2022). Owing to the fact that 3D printing does not require a cleanroom setting nor the skilled personnel, this method holds great promise for low-cost microfluidics, especially when non-conventional designs and multi-layered structures are needed (Raoufi et al., 2020; Razavi Bazaz et al., 2020; Su et al., 2020). However, current 3D printing techniques still have limitations such as clogging of the channels, poor quality of the surfaces and low resolution (He et al., 2016; Mehta and Rath, 2021). Despite having lower resolution than conventional cleanroom techniques, the resolution of 3D printers is already suitable for multiple microfluidic applications (He et al., 2016; Mehta and Rath, 2021). Although the selection of resins for printing transparent parts and molds is limited, with the rapid advances in this technology, 3D printing resins that promote better resolution, surface finishing and transparency would further enhance the capabilities of microfluidic devices; in fact, there are resins currently being developed with the specific purpose of fabricating microfluidic devices (e.g., Figure 4E) (He et al., 2016; Nielsen et al., 2020). For further discussion on 3D printing technologies applied to microfluidic devices manufacturing, the readers are encouraged to review the following references: (Chen et al., 2016; He et al., 2016; Enders et al., 2019; de Almeida Monteiro Melo Ferraz et al., 2020; Gonzalez et al., 2020; Nielsen et al., 2020; Mehta and Rath, 2021).

Unlike 3D printing, micromilling is a subtractive manufacturing technique that removes the materials from the bulk to create the desired structures. The prepared parts can be bonded to a substrate to create the final enclosed microfluidic device (Hossain and Rahman, 2018; Rahim and Ehsan, 2021), or it can be used as a mold for PDMS (Jiménez-Díaz et al., 2019; Javidanbardan et al., 2021). Similar to many other low-cost fabrication methods, micromilling does not require a cleanroom and is relatively fast, greatly expediting the manufacturing processes especially for prototyping tests (Faustino et al., 2016; Nguyen et al., 2019). Currently, many materials have been explored to create microfluidic devices using micromilling, among which PMMA and aluminum are two most popular materials (Jiménez-Díaz et al., 2019; Nguyen et al., 2019; Behroodi et al., 2020; Javidanbardan et al., 2021; Saptaji et al., 2021). The micromilled molds made of aluminum can be used for casting multiple times, which could further reduce the cost of the final device (Guckenberger et al., 2015; Nguyen et al., 2019).

On the other hand, micromilling has several limitations that should be considered. For example, the milling bits used in micromilling are prone to breaking especially when high resolution (e.g., 25 µm) is required (Charles et al., 2018; Leclerc, 2021). In addition, complex 3D features and designs may not be suitable for micromilling, even though customized milling bits may be able to create structures with preset shapes (Ku et al., 2018; Javidanbardan et al., 2021). Micromilling only removes the materials from external surfaces, therefore bonding with other substrates is inevitable to create enclosed microchannels. The bonding could be done mechanically (i.e., using screws), thermally (i.e., bonding two PMMA plates when heated up), or using surface treatments and adhesives such as the tapes (Kosoff et al., 2018; Owens and Hart, 2018; Madureira et al., 2019; Gonçalves et al., 2021).

Laser micromachining has also been employed for low-cost microfluidics (Mohammed et al., 2016; Persson et al., 2022). For example, CO₂ laser is a widely used microfabrication method (Chen et al., 2019; Buchroithner et al., 2021; Nishat et al., 2021; Shin and Choi, 2021). During the fabrication process, the laser energy is focused on the region of interest of the workpieces, causing the materials to melt and evaporate. Typically, a CO₂ laser with a wavelength of 10.6 µm are used (Mohammed et al., 2016; Persson et al., 2022). Indeed, sophisticated laser machine or reduced wavelength (e.g., femtosecond lasers) can be applied to further improve the cutting resolution, yet these methods are not suitable for low-cost microfluidics since extra costs are inevitably required (Elgohary et al., 2020; Saadat et al., 2020; Andriukaitis et al., 2022). When it comes to the materials used in laser micromachining, both hard materials such as glass and soft materials such as PMMA, cyclic olefin copolymer (COC) and even paper could be used (Islam et al., 2018; Lin et al., 2019b). Note that to avoid the cracks caused by thermal stress, surface coating could be applied on the glass slides (Chung et al., 2010). In addition, laser micromachining can be used to create both molds and final devices after bonding (Mahmud et al., 2018; Gao et al., 2019a; Ma et al., 2019). The bonding and assembly techniques used for laser cut devices are similar to those used for micromilled devices (Faustino et al., 2016; Mohammed et al., 2016; Nguyen et al., 2019; Persson et al., 2022).

Thin plastic films can also be directly made into final devices via screen printing technology, or as simple as hand cutting (Gale et al., 2018; Nishat et al., 2021). Films and thin plastics can be fabricated at large scale using laminate manufacturing or roller imprinting (Focke et al., 2010; Su et al., 2016). Note that PMMA has been widely used in low-cost microfluidics, it has been used to create devices by micromilling, laser ablation, and by the injection molding (Kotz et al., 2020; Ma et al., 2020), thus holding promise in mass production (Trotta et al., 2018; Ma et al., 2020). It is also worth mentioning that the methods such as roller imprinting, injection molding and hot embossing do require a high resolution mold, which increases the initial cost but eventually can compensate towards low unit price (Shiu et al., 2008; Zhang et al., 2018; Zhang et al., 2019). Indeed, there are other fabrication methods explored for microfluidics, for example, microwire has been used to create devices but the performance is not as high as that of 3D printing (Jia et al., 2008; Kuo et al., 2013). Interested readers are encouraged to read the references (Tomazelli Coltro et al., 2014; Faustino et al., 2016; Raj M and Chakraborty, 2020). Table 1 compares the previously mentioned fabrication methods.

Heavy metal pollution in water has received increasing attention over the past decades (Almeida et al., 2018; Santangelo et al., 2019; Kinuthia et al., 2020). It is reported that even at low concentrations, these contaminants can pose a great threat to the aquatic environment, ecosystem, and human health (Snyder et al., 2020; Barocio et al., 2021). Given such growing concerns, low-cost microfluidic devices can be an affordable tool for continuous water monitoring regarding heavy metal contamination worldwide. The burgeoning advancements are distinct as evidenced by continuous developments made over the past years, with many applications built on top of paper microfluidics (Almeida et al., 2018). To name a few, Wang et al. developed a μPAD with high detection accuracy and selectivity for lead ions (Pb²⁺) in drinking water. The device realized rapid visual quantitative detection by examining the extension length of the color bar in the particle dam (Wang et al., 2022). A similar device was designed to quantify silver (Ag⁺) contamination in freshwater, and it was reported to have a detection limit of 453.7 nM, high selectivity, and a high recovery rate of 96.8% (Wang et al., 2020). Jarujamrus et al. developed a μPAD to detect Mercury (Hg²⁺) in various water samples with the ability to instantly report Hg²⁺ concentration on-site by using a smartphone. The smartphone analyzer is responsive and user-friendly, which has enabled unskilled users to use this device to conduct sample analysis (Jarujamrus et al., 2018). Similar applications used μPADs for the detection of Cu²⁺ (Quinn et al., 2018; Sharifi et al., 2020).

Besides the detection of a single type of heavy metal, μPADs were also developed for the identification of multiple heavy metals simultaneously. Khoshbin et al. developed a paper-based aptasensor to detect Ag⁺ and Hg²⁺ within 10 min based on conformational changes of Ag⁺ -and Hg²⁺ specific aptamers. The concentration of the ions can be indicated by fluorescence recovery rate, with a limit of detection of 1.33 p.m. for Hg²⁺ and 1.01 p.m. for Ag⁺ (Khoshbin et al., 2020). Idros et al. used a μPAD to detect several major heavy metals, including Hg²⁺, Pb²⁺, Cr³⁺, Ni²⁺, Cu²⁺, and Fe³⁺ by applying different ligands loaded onto the test paper (Idros and Chu, 2018) (Figure 5A). Similarly, Kamnoet et al. capitalized on the colorimetric assays to identify multiple heavy metals including Cu²⁺, Co²⁺, Ni²⁺, Hg²⁺, and Mn²⁺ with a corresponding limit of detection of 0.32, 0.59, 5.87, 0.20, and 0.11 mg/L, respectively (Kamnoet et al., 2021).

The porous nature of papers and capillary driving can limit associated fluid and particle manipulation. Herein, other materials such as polymers are also applied to fabricate microfluidic devices for more accurate heavy metal detection. For example, an epitaxial graphene sensor was combined with a 3D-printed microfluidic chip to detect Pb²⁺ and Cd²⁺ (Santangelo et al., 2019). In another study, a porous conductive carbon cloth was integrated with a microfluidic device to desalinate and recover valuable metal ions (Cu²⁺, Zn²⁺, Ni²⁺, Ag⁺, and Zn²⁺/Cu²⁺ mixtures) from wastewater samples (Allioux et al., 2018). Ding et al. successfully conducted heavy metal analysis by using a sponge-based microfluidic device that was integrated with ion-selective electrodes for sampling heavy metal ions (Cd²⁺ and Pb²⁺) and non-metal clinically related chemical ions, namely K⁺, Na⁺, and Cl⁻ (Ding et al., 2021). Furthermore, a combined μCPAD was developed to detect Mercury and lead ions in water samples. The groups used cloth’s ductility and durability to endure the oscillation during fabrication to improve the producibility and life span of the device (Wang et al., 2022).

Non-metal substances are more abundant pollutants in water and are highly complex by nature. Nowadays, portable microfluidic devices are playing a critical role in water quality analysis for a large variety of toxins, such as pharmaceutical residues, due to their many advantages (Barocio et al., 2021). Scala-Benuzzi et al. developed an electrochemical paper-based immunocapture assay (EPIA) to assess Ethinylestradiol quantitively in water samples. It was reported the test achieved a low detection limit of 0.1 ng/L and a linearity range of 0.5–120 ng/L (Scala-Benuzzi et al., 2018). In another study, chlorpyrifos pesticide was detected by using a lipase-embedded paper-based device (Sankar et al., 2020). The limit of detection and limit of quantification was found to be 0.065 mg/L and 0.198 mg/L, respectively. Interestingly, the wash water of cauliflower, grapes, coriander leaves, brinjal, and bitter guard could be used as samples (Sankar et al., 2020). Jemmeli et al. developed a highly sensitive paper-based electrochemical sensor to detect bisphenol A (BPA) in drinking water (Jemmeli et al., 2020). Mako et al. developed a μPAD to detect nitrite levels in drinking water (Mako et al., 2020). Peters et al., developed a μPAD to monitor total ammonia levels in freshwater (Peters et al., 2019). Similarly, μPAD was used for detecting phosphate in water samples (Sarwar et al., 2019; Racicot et al., 2020). Besides paper-based devices, Carvalho et al. developed a fully 3D printed thread-based microfluidic device to detect Nitrite in well water samples with high precision (Carvalho et al., 2021). Caetano et al. developed a textile thread-based microfluidic device combined with an electrochemical biosensor to detect phenol concentration in tap water (Caetano et al., 2018).

It is worth noting that, among all the non-metallic pollutants, microplastics have been drawing lots of research attention recently. Microfluidic devices can benefit microplastic-related research in many ways, such as microplastic identification and separation. However, only a few low-cost microfluidic devices have been developed for these applications. Pollard et al. developed a low-cost and high-throughput three-dimensional printed microfluidic resistive pulse sensor for characterizing algae with spherical and rod structures as well as microplastics from tea bags. The device can rapidly screen liquids at a volume rate of 1L/min in the presence of microplastic and algae (Pollard et al., 2020) (Figure 5C). Mesquita et al. developed a 3D printed microfluidic device for microplastic identification that improved the Nile Red staining process (Mesquita et al., 2022). It is suggested that researchers use the full potential of low-cost microfluidic devices to achieve reproducible and reliable long-term assessment of environmental microplastics.

The presence of waterborne pathogens can cause severe illnesses. Continuous monitoring and in-situ studies of waterborne microorganisms are another rapidly growing research interest. In a recent study, Yin et al. developed a 3D-printed integrated microfluidic chip for colorimetric detection of SARS-CoV-2 and other human enteric pathogens in wastewater. The sensitivity of detection was reported to be 100 genome equivalent (GE)/mL for SARS-CoV-2 and 500 colony-forming units (CFU)/mL for other targeted human enteric pathogens (Yin et al., 2021). Schaumburg et al. designed a μPAD for waterborne bacteria detection which consists of two sequential pre-concentration steps. The detection limit of concentration was as low as 9.2 CFU/ml in laboratory samples and 920 CFU/ml in apple juice samples within ∼90 min (Schaumburg et al., 2019). Several studies successfully used μPAD and 3D printed microfluidic devices to detect E. coli in various water samples and achieved low detection limit, high sensitivity, and quick analysis (Sweet et al., 2019; Lin et al., 2020b; Snyder et al., 2020; Alonzo et al., 2022).

Airborne metal particles are one of the most representative harmful elements. Several studies have used μPAD to detect some typical airborne metal particles. Sun et al. developed a μPAD that realized on-site multiaxial quantification of airborne trace metals by implementing unmanned aerial vehicle in-air sampling (UAV). Data can be easily processed by a smartphone within 30 min (Sun et al., 2018). The same group later applied graphene oxide (GO) coating onto the paper and improved the detection limits for Fe, Cu, and Ni to 6.6, 5.1, and 9.9 ng respectively, which is comparable to the commercial coupled plasma (ICP) instruments (Sun et al., 2019) (Figure 6A). Jia et al. successfully used a μPAD to detect cobalt (Co), copper (Cu), and iron (Fe) in ambient air and street sediments with detection limits of 8.2, 45.8, and 186.0 ng (Jia et al., 2017).

Unlike water applications, only a few devices used low-cost microfluidic devices for the detection of non-metallic airborne pollutants. For example, Guo et al. developed a smartphone-based microfluidic sensor to detect gaseous formaldehyde in the ambient. The microfluidic chip consists of two reagent reservoirs, a reaction reservoir, and a mixing column. The PTFE membrane was used to prevent the fluid from flowing out while the gas molecules enter. The system showed great selectivity against other ambient gas (Guo et al., 2018). Zhao et al. developed a 3D printed based microfluidic impactor for particular matter classification and concentration detection (Zhao et al., 2016). However, most of these applications used costly fabrication techniques, such as photolithography, and the integration with multiple sensors also increased the overall cost (Poenar, 2019).

Dias et al. used a μPAD to detect levoglucosan concentration in the ambient with a colorimetric method. The linear detection range is 0–64.8 μg m/L and the detection limit is 2 and 6 μg m/L. The device showed selectivity for levoglucosan with variation in colorimetric signal intensity lower than 8% (Dias et al., 2019). Seok et al. developed a μPAD combined with a 3D printed analysis kit for detection of airborne bacteria by collecting aerosols (Seok et al., 2021) (Figure 6C).

Ding et al. used an acidified μPAD integrated with potentiometric sensors for the detection of multiple heavy metal ions in the soil, street run-off, and multiple environmental samples (Ding et al., 2021). Similarly, an eco-friendlier metal-modified μPAD was developed for the same purpose (Silva et al., 2022). Xi et al. developed a centrifugal microfluidic system for pyrene extraction from soil (Xi et al., 2010). A similar centrifugal microfluidic device was also used for the detection of pesticide residues in vegetables and soil (Duford et al., 2013). Other soil nutrients can be detected using colorimetric microfluidic devices, most of these applications used μPAD, 3D printed devices, and a combination of low-cost fabrication techniques (Cheng et al., 2021).

With the advantages of microfluidic platforms, the development of soil-on-a-chip has been growing to study soil biofilms and microorganisms’ ecological and biological impacts (Stanley et al., 2016; Wu et al., 2022). However, challenges and limitations still exist, such as the controlling of hydrophilic and hydrophobic surfaces in PDMS based devices, which highlighted the potential benefits of using porous membrane microchannels, which are normally fabricated with low costs.