Microfluidic Field-Deployable Systems for Colorimetric-Based Monitoring of Nitrogen Species in Environmental Waterbodies: Past, Present, and Future

Abstract

The biogeochemical cycling of nitrogen (N) in natural waterbodies, ranging from freshwaters to estuaries and seawater, is fundamental to the health of aquatic ecosystems. Anthropogenic pressures (agricultural runoff, atmospheric deposition, and wastewater discharge) have profound effects on these cycles, leading to widespread problems, such as eutrophication, harmful algal blooms, and contamination of drinking water sources. Monitoring of different N-species—ammonium (NH₄⁺), nitrite (NO₂⁻), nitrate (NO₃⁻) ions, dissolved organic nitrogen (DON), and total nitrogen (TN)—is of crucial importance to protect and mitigate environmental harm. Traditional analytical methodologies, while providing accurate laboratory data, are hampered by logistical complexity, high cost, and the inability to capture transient environmental events in near-real time. In response to this demand, miniaturised microfluidic technologies offer the opportunity for rapid, on-site measurements with significantly reduced reagent/sample consumption and the development of portable sensors. Here, we review and critically evaluate the principles, state-of-the-art applications, inherent advantages, and ongoing challenges associated with the use of microfluidic colorimetry for N-species in a variety of environmental waterbodies. We explore adaptations of classical colorimetric chemistry to microfluidic-based formats, examine strategies to mitigate complex matrix interferences, and consider future trajectories with autonomous platforms and smart sensor networks for simultaneous multiplexed N-species determination.

1. Introduction

Nitrogen (N) is a fundamental nutrient in limiting primary productivity in many aquatic ecosystems [1]. Its natural biogeochemical cycle involves complex transformations between various inorganic and organic forms, including atmospheric nitrogen (N₂) and ammonia (NH₃), ammonium (NH₄⁺), nitrite (NO₂⁻), and nitrate (NO₃⁻) ions, dissolved organic nitrogen (DON), and particulate organic nitrogen (PON) [2,3]. Key microbial processes, such as N fixation, nitrification, denitrification, and ammonification, govern the availability and fate of these species [4]. However, over the past century, human activities, primarily intensive agriculture (fertiliser use), combustion of fossil fuels (atmospheric N deposition), and discharge of untreated or inadequately treated wastewater (sewage) have drastically increased N loading levels into freshwater, estuarine, and coastal marine ecosystems globally [5,6] (Figure 1).

Figure 1. An illustration of the sources of anthropogenic nitrogen pollution (adapted from [7]).

Anthropogenic N enrichment has severe ecological consequences [8]. Excessive inputs of bioavailable N, particularly NO₃⁻ and NH₄⁺ ions, together with phosphorus (P) often trigger eutrophication, leading to excessive growth of algal species and aquatic plants [9,10]. In eutrophic lakes, streams, and marine waterbodies, the concentrations of N can exceed, 650, 1500, and 350 mg m³, respectively [9]. The subsequent decomposition of the organic matter by bacteria consumes dissolved oxygen (DO), creating hypoxic or anoxic conditions, commonly referred to as “dead zones,” which are detrimental to fish and other aquatic fauna [11,12]. Certain conditions of nutrient imbalance can also promote the proliferation of harmful algal blooms (HABs) that can produce toxins dangerous to wildlife and humans and cause significant economic losses [13,14]. Furthermore, elevated NO₃⁻ levels in groundwater and surface waters used for drinking can pose direct human health risks, most notably methemoglobinemia in infants [15]. Two cases of blue baby syndrome showed that infants became ill after being fed with water from private wells which contained 1.6–2.0 mM of NO₃⁻ [15]. Given these profound impacts, understanding the sources, transport, transformation, and fate of N-species in environmental waterbodies is crucial for effective environmental management, restoration efforts, and regulatory compliance (e.g., under the EU Water Framework Directive or the US Clean Water Act) [16,17]. This necessitates robust monitoring programs capable of providing accurate data on N-concentrations with sufficient sensitivity, precision, and spatiotemporal resolution to capture the dynamic nature of aquatic systems.

Traditional analytical methods for N-species, primarily spectrophotometric/colorimetric techniques, are flexible [18] and involve discrete sample collection, preservation, and transport to centralised testing laboratories. This often requires time-consuming manual or semi-automated analysis. While these methods have been developed in routine analysis because they are generally accurate and precise and respond to a wider range of analytes compared to some other traditional methods (e.g., electrochemistry, X-ray fluorescence, etc.), their logistical demands limit the frequency and spatial coverage of sampling, especially in remote or inaccessible locations [19]. This often results in an incomplete understanding of N-dynamics, as transient events (e.g., storm runoff, episodic discharges, and microbial activity) may be missed. Moreover, the cost per sample can be high, further constraining monitoring efforts. There is, therefore, a compelling need for innovative analytical tools that are reliable, rapid, accurate, precise, sensitive, cost-effective, portable, and capable of autonomous or semi-autonomous operation for in situ measurements in near-real time [20,21].

Microfluidic technologies have emerged as a powerful and versatile solution to address this analytical need [21,22]. By miniaturising analytical processes into µL volume, these lab-on-chip (LOC) systems offer the potential for rapid, high-throughput analysis with significantly reduced reagent consumption and waste generation, making them highly attractive for near-real-time continuous environmental monitoring applications [23]. In this review, we compare some of the most recent strategies in microfluidic field-deployable systems for colorimetric monitoring of N-species, commented on the dis/advantages, applications in environmental analysis, and future perspectives and trends.

2. Fundamentals of Microfluidic Platforms for Chemical Sensing

2.1. Non-Droplet Microfluidic Systems

Non-droplet microfluidics include continuous-flow microfluidic systems characterised by moving fluids through connected microchannels rather than handling them as discrete droplets. These systems have attractive features, such as simplicity, portability, disposability, and rapid analysis [22,24,25,26,27,28,29,30,31,32,33,34,35]. The most significant issue in microfluidics is the method of moving fluid through the system channels. Flow-generation systems can be divided into two main groups: active and passive flow-generation systems. In active systems, valves, pumps, motors, micropumps, magnetic, and electric or thermal forces are typically included. Main types of passive systems include continuous-flow microchannels in which fluids are moved inside microchannels (e.g., Y-shaped or T-shaped channels) for controlled mixing of fluids. In capillary-driven microfluidics, the fluids are driven by capillary action without additional power supply (not by external pumps) such as in microfluidic paper-based device (µPAD) systems [36,37]. Detection on the µPAD can be carried out by comparing the colour intensity in the reaction zone, using cameras, scanners, or smartphones [38]. µPAD and a smartphone-based app allow for the simultaneous monitoring of some N-species [22]. The developed app can capture, process, and quantify the µPAD colorimetric outputs on-site. These devices have been recently used in environmental water monitoring with the development of digital communication technology [22,26,30,31,34].

2.2. Droplet-Based Microfluidic Systems

Droplet-based microfluidics involves the precise generation and manipulation of discrete droplets as immiscible fluid segments within microfabricated channels [39,40]. Typically, an aqueous phase (containing the sample and reagent/s) is dispersed as droplets within a continuous, immiscible oil phase (Figure 2). Each droplet functions as an independent, isolated microreactor. As such, droplet-based systems have the advantage of long-duration deployment with high specificity, precision, and accuracy. Common passive droplet generation techniques, relying on interfacial tension and fluid dynamics, include T-junctions, in which the aqueous phase is injected perpendicularly into a main channel carrying the continuous oil phase (Figure 2) [39,40]. In flow-focusing formation, the aqueous phase is hydrodynamically controlled by two co-flowing streams of the oil phase through a narrow orifice, resulting in droplet formation due to Rayleigh–Plateau instability exacerbated by intra-fluid shear [41,42]. This method generally offers excellent control over droplet size and monodispersion. Co-flow (or co-axial flow) assumes that the aqueous phase is introduced via an inner capillary positioned coaxially within a larger channel carrying the oil phase. Droplets detach from the capillary tip [43]. The choice of geometry and flow rates dictates droplet size and generation frequency (Hz to kHz). Surfactants are typically added to the oil phase to stabilise droplet interfaces against coalescence and to control wetting properties at the channel walls [44].

Figure 2. Schematic diagrams of (a) the continuous vs. segmented flow in microfluidic systems; (b) droplet generation strategies: capillary, T-junction, and flow-focusing formation.

Alongside these traditional droplet formation techniques, the use of peristaltic pumps has enabled phased micro-pumping, in which the sample, reagents, and carrier oil are pumped in an anti-phase approach [45]. This technique allows for very precise control of the droplet dynamics, with the droplet size and composition being governed by the size of the grooves in the peristaltic pump roller. This process allows for much more robust droplet generation, as the dynamics are independent of the overall flow rate, unlike in co-flow regimes. In our recently developed microfluidic systems suitable for environmental analyses [35,46], this approach has been applied and demonstrates stable droplet generation over long deployments.

2.3. On-Chip Droplet Operations

Once formed, droplets can be subjected to a variety of manipulations to perform complex analytical protocols. Droplets containing different reagents (e.g., salicylate and nitroprusside reagents for NH₄⁺) or water samples can be merged to initiate reactions, passively (e.g., at channel expansions) or actively (e.g., using electro-coalescence or acoustic forces) [39,47]. Complete homogeneous mixing of reagents with the sample is crucial to the precision and accuracy of the analyses. Single droplets can be divided into multiple smaller droplets for parallel processing to measure precision [48]. Rapid mixing within droplets is achieved via internal chaotic advection generated by turbulent shear at the droplet interface as it moves through curved or serpentine channels, significantly accelerating reaction rates and reducing the time to reach chemical equilibrium [49,50]. Delay lines (e.g., extended, long, serpentine channels) provide controlled residence times for reactions to proceed to completion [23]. If required, the droplets can be selectively routed based on their contents (e.g., after an optical measurement) using active sorting mechanisms like dielectrophoresis (DEP) or surface acoustic waves (SAW) [51,52].

2.4. Intrinsic Advantages of Microfluidic Droplet-Based Systems

Similar to non-droplet microfluidics, the droplet-based formats offer several key advantages for chemical analysis, including miniaturisation and reduced reagent/sample consumption (typically 10³ to 10⁶-times less than conventional methods), lower costs, and minimised chemical waste, which is a significant benefit for field deployments where reagent transport and waste disposal can be problematic [21,53]. Enhanced mixing and short diffusion distances within droplets accelerate reaction kinetics, often reducing analysis times from h to min or even sec [49,54].

High throughput and temporal resolution enable the ability to generate and analyse droplets at high frequencies, which then allows for rapid sequential measurements enabling high-temporal-resolution monitoring of dynamic processes [23]. Each droplet acts as an isolated microreactor; their separation prevents cross-contamination between sequential samples and eliminates axial dispersion (i.e., Taylor dispersion), which is common in continuous-flow systems, leading to sharper signals [23]. In continuous flow, due to in-tube mixing and diffusion, signal smearing occurs and therefore sensitivity is low, as is temporal resolution. In droplet-based flow, there is no mixing or diffusion, nor smearing, and better resolution and temporal discretisation are obtained. The small scale of microfluidic devices facilitates the development of compact, automated, and portable analytical systems suitable for on-site and field deployments [20,55].

3. Colorimetric Detection in Microfluidic Systems

3.1. Principles of Colorimetric Measurement

Colorimetric assays, which are based on the formation of a coloured reaction product whose absorbance is proportional to analyte concentration, are widely used due to their simplicity, reliability, stability, and relatively low cost. The relationship between light absorbance (A) and analyte concentration (c) is governed by the Beer–Lambert law:

$$A=\epsilon \times l \times c$$

(1)

where A is the absorbance, ε is the molar absorptivity of the chromophore, l is the optical pathlength through the sample, and c is the analyte concentration [18]. For colorimetry-detection-based systems, this means that the intensity of the colour developed is positively correlated with the analyte concentration, provided the other variables are constant. As an example, our recently tested microfluidic droplet-based system showed an excellent calibration regression across a broad range covering three orders of magnitude of ammonium ion concentrations (Figure 3) [56].

Figure 3. Results of the calibration of an automated droplet microfluidic system for ammonium ion concentrations (by modified Berthelot’s reaction) in water [56].

3.2. Optical Detection Setups for Microfluidic Analysis

Measuring the absorbance in field-deployable microfluidic systems requires precise optical integration with the microfluidic chip [57]. For the light source, the light-emitting diodes (LEDs) are favoured for their low-cost, stability, small size, and availability across various wavelengths suitable for different chromogenic reactions [58]. As detectors, photodiodes (PDs) or photomultiplier tubes (PMTs) are commonly used to measure the transmitted light intensity. PDs are compact and cost-effective, while PMTs are more costly and bulkier but offer higher sensitivity for low light levels, which can be important when dealing with diluted environmental samples or short optical pathlengths [59]. Image sensor technologies such as charge-coupled devices (used in digital cameras) can capture images for colorimetric analysis but may have lower throughput for quantitative absorbance of individual droplets compared to PD/PMT setups, unless specialised high-speed imaging is used [60]. Optical fibres are frequently used to deliver light to, and collect light from, the microchannel at the droplet interrogation point, ensuring precise alignment [58,61]. Microlenses can be integrated to focus the light beam onto the droplet or to collimate the transmitted light.

A significant challenge in microfluidic absorbance measurements is the inherently short optical pathlength (often equivalent to the droplet diameter or channel height, typically 50–200 µm), which can limit sensitivity. Strategies to increase effective pathlength in droplet-based systems include removing oil and converting the droplets into a single-phase aqueous flow which can be measured within a U-shape channel [62], designing channels that flatten droplets in the detection zone, or employing more complex optical configurations like liquid-core waveguides [63] or multi-pass cells (e.g., internal reflectance), although these are more challenging to implement robustly for droplet systems in the field (Figure 4). In our recently developed systems, a flow cell with UT7 polytetrafluoroethylene (PTFE) tubing is commonly used with its optimal pathlength of 0.7 mm [45].

Figure 4. Strategies for the design of different optical configurations of pathlengths: cross-sectional view of the U-shape channel with membrane and absorption pad [62].

As droplets pass the detector, a series of transient absorbance signals (peaks) are generated. Sophisticated algorithms are required for peak detection, baseline correction, noise filtering, and calculation of absorbance values, which are then correlated with analyte concentrations using calibration curves established with standard solutions [58]. Baseline correction requires monitoring blanks, such as oil separation droplets. Ideally, calibration standards could be introduced frequently to measure accuracy (Figure 5).

Figure 5. Screenshot displaying raw light intensity data for nitrate ion standards obtained for calibrating the sensor. The raw light intensity signal shows the oil levels, and the droplet plateaus correspond to different concentration of nitrate standard solutions in increasing order (0.03–2 mM).

4. Application of Microfluidics to Nitrogen Species Analysis in Environmental Waterbodies

Unlike pure aquatic matrices, adapting standard colorimetric chemistries for N-species to microfluidic platforms for environmental water analysis involves optimising reactions for small volumes and rapid timescales and, critically, addressing the diverse and often challenging matrices. These introduce challenges and uncertainties in maintaining precision, sensitivity, and accuracy. Table 1 summarises developed microfluidic systems based on colorimetric or spectrophotometric analyses of N-species in environmental waterbodies with their analytical performances, including detection time, reagent consumption, detection range, and limits of detection (LOD) for each species.

Table 1. Review of microfluidic analysis for in situ monitoring of ammonium, nitrite, and nitrate ions in environmental waterbodies.

Sample	Detection Method	Chip Material	Detection Time	Reagent Consumption	Detection Range	LOD	Reference
Ammonium ions
Natural waters: river water samples, ammonium standard solutions	Colorimetric/FIA	D-ABS/graphene PLA	1.6 min	100 µL	28–280 µM	8.4 µM	[24]
Water	Colorimetric (Berthelot’s)	Paper	15 min	1.2 mL	0.55–11 mM	<0.55 mM	[25]
Industrial wastewater	Colorimetric (Nessler’s)	Paper	2 min	15 µL	0–280 µM	187 µM	[26]
Freshwater: park pond, garden lake	Colorimetric	Paper BTB-based NY based	1 min 5 min	8 µL	28–1680 µM 112–560 µM	17.9 µM 26.3 µM	[27]
River	Colorimetric (Berthelot’s)	PLA	10 s	0.3 µL	10–110 µM	6.1 µM	[56]
Water	Colorimetric (Berthelot’s)	PMMA	205 s	100 µL	21–286 µM	19 µM	[64]
Nitrite ions
Groundwater, tap water	Colorimetric	Paper	10 min	20.5 µL	1.09–217 µM	651 nM	[30]
Aqueous samples	Colorimetric	Paper/PMMA	15 min	0.5 µL	3.9–1000 µM	14.8 µM	[36]
Aqueous sample	Colorimetric	Paper	5 min	11 µL	5–500 µM	1.3 µM	[37]
Aqueous samples	Spectrophotometric	PDMS/PMMA	N/A	<1.25 mL	4.3–97.6 µM	1.54 µM	[38]
River and lake water, tap water	Colorimetric	PMMA	45 min	10 µL	0–0.28 mM	9.35 µM	[65]
Seawater	Colorimetric	PMMA	30 min	12 µL h−1	0–5 µM	15 nM	[66]
Aqueous samples	Spectrophotometric	PDMS	N/A	2.5 mL	2.17–217 µM	1.17 µM	[67]
Water solutions	Differential Colorimetric	PDMS	>90 s	N/A	0.25–4 mM	0.33 µM	[68]
Nitrate ions
Sargasso Sea	Colorimetric	Paper	21 min	95 µL	0–80 mM	8.48 µM	[32]
Atlantic	Colorimetric	PMMA	11 min 20 s	0.66 mL	0.25–750 µM	0.03 ± 0.01 µM	[33]
Mineral water	Colorimetric	PMMA	~14 min	N/A	0.02–0.3 mM	1.25 µM	[69]
Water solutions	Colorimetric	PMMA	<30 min	0.4 µL min−1	0.25–50 µM	20 nM	[70]
Seawater	Colorimetric	PMMA	17 min	4.71 mL	N/A	<0.1 µM	[71]
Water	Spectrophotometric	PMMA/PDMS	115 s (per sample)	26.8 µL (per sample)	0.1–30 µM	0.05 µM	[72]

ABS—acrylonitrile butadiene styrene. BTB—bromothymol blue. NY—nitrazine yellow. PLA—polylactic acid. PDMS—polydimethylsiloxane. PMMA—polymethyl methacrylate.

4.1. Ammonium (NH₄⁺) Species

The Berthelot reaction (or its modified version with a less toxic salicylate variant) is the most common colorimetric method for NH₄⁺ detection [24,25,56,73,74,75]. NH₄⁺ reacts with hypochlorite to form monochloramine, which then reacts with phenol (or salicylate) in an alkaline medium, typically catalysed by nitroprusside, to form a blue indophenol dye (λ_max ≈ 630–660 nm) [18,76]. In our recently developed system for in situ monitoring of NH₄⁺ ions in natural waters (Figure 6), the droplet-based microfluidic system showed the main advantages of high measurement frequency (10 s per measurement) and low reagent consumption (0.3 µL per measurement) (Table 1).

Figure 6. Schematic diagram of the microfluidic droplet-based system designed for continuous monitoring of ammonium ions in water samples (adapted from [56]). From left to right: The aqueous sample (via a filter), was pumped and mixed with reagents and oil, stored in fluid pouches, by a peristaltic pump into a microfluidic chip to generate droplets. The droplets reacted in a temperature-controlled the heater unit forming the coloured (blue) compound detected by an in-line photo detector (with an LED light source).

Several other research groups have also developed non-droplet optical-based sensors for NH₄⁺ monitoring in aquatic media. For instance, Khongpet et al. (2019), though focused on freshwater samples, demonstrated an inexpensive microfluidic hydrodynamic sequential injection method with a detection limit of 19 µM, using a sensitive Berthelot’s method [64]. Systems for environmental waters must often achieve lower detection limits (µg/L or sub-µM levels) compared to wastewater applications [77]. Cho et al. (2018) fabricated small colorimetric ammonia gas sensors by filtration of modified Berthelot’s reagent on porous paper for the analysis of wastewater or seawater [25]. Nxumalo et al. (2020) described a µPAD for continuous quantification of NH₄⁺ in industrial wastewater using Nessler’s reagent (potassium tetraiodomercurate(II) and potassium hydroxide), which has a high environmental impact and, hence, is not preferred for field use [26]. Peters et al. (2019) developed a more suitable field screening system based on the colour change of bromothymol blue or nitrazine yellow for total NH₄⁺ monitoring in freshwater [27]. Salee et al. (2025) recently developed a new gel-based sensing platform for LED-based colorimetric NH₄⁺ flow-analysis detection in water using Berthelot’s reagents, highlighting stability, accuracy, and high sensitivity (LOD = 1.5 µM) using a higher sample/reagent volume range (50–200 mL) [78].

Challenges in environmental waters may arise because of very low ammonium levels in natural unpolluted waters, demanding high sensitivity of the developed methods. In estuarine and marine waters, high salt (saline and brackish) concentrations can affect reaction rates, reagent solubility, and optical properties (refractive index changes) and can cause salt precipitation in microchannels if evaporation occurs [28,29]. Specific calibration matrices matching sample salinity are essential. Dissolved organic matter (DOM) such as humic and fulvic acids can impart colour to the water (interfering with absorbance readings) or react with reagents, causing chemical interference [79]. Moreover, the natural buffering capacity of some waters can influence the optimal pH for Berthelot’s reaction.

4.2. Nitrite (NO₂⁻) Species

The Griess reaction is the standard method for determination of total NO₂⁻ ion levels. Nitrite reacts with an aromatic amine (e.g., sulphanilamide) under acidic conditions to form a diazonium salt, which then couples with a second aromatic compound (e.g., N-(1-naphthyl) ethylenediamine dihydrochloride, known as NEDA·2HCl) to produce a stable, intensely coloured azo dye (pink, λ_max ≈ 540–550 nm) [18,80,81,82].

The rapid kinetics and high molar absorptivity of the Griess reaction make it well-suited for microfluidics. Man et al. (2024) demonstrated a good concentration gradient colorimetric detection chip designed for nitrite in environmental water samples (river, lakes, etc.) with good sensitivity and high recovery (>95%) [65]. Portable systems, sometimes smartphone-based, have also been developed for rapid nitrite field measurements [22,30]. A recent strategy for NO₂⁻ detection is based on a smartphone app using photooxidation [83]. This reaction, operated on an Android phone, was proven by the injection of sodium nitrite solution as the real sample.

Some microfluidic systems are valuable for tracking nitrification hotspots or anoxic zones where nitrite can accumulate. Beaton et al. (2011) developed an early in situ microfluidic colorimetric sensor for the determination of total NO₂⁻ in seawater samples [66]. This microfluidic device based on a poly (methyl methacrylate) (PMMA) microchip was operated in situ for 57 h and made 375 discrete measurements with low reagent consumption. Although the developed systems require the on-site addition of some reagents after loading the sample onto the device, they have several advantages in terms of speed, simplicity, selectivity, and sensitivity. With some further modifications, the developed system can be a good candidate for the simultaneous detection of other inorganic N-species, such as NH₄⁺ and NO₃⁻.

Low-cost fabricated non-droplet-based microfluidic devices (µPADs with or without integrated smartphones) were employed for quantitative colorimetric analysis of NO₂⁻ across a broad concentration range with satisfactory accuracy and reproducibility [36,37,38,67]. Shi et al. (2018) demonstrated a robust method for the determination of NO₂⁻ ions by colorimetry in the microfluidic network [68]. It was shown that for lower concentrations of NO₂⁻ ions, the errors measured were lower than 1.16% and the method is therefore suitable for different application scenarios with complex water quality and testing environments.

Like the monitoring of NH₄⁺, high salinity can affect the kinetics of the Griess reaction and optical baselines. Therefore, salinity-matched calibrations are crucial in these systems [28]. Strong oxidants or reductants, if present, can also interfere, though this is generally less common in typical surface waters than in some industrial discharges.

4.3. Nitrate (NO₃⁻) Species

Direct colorimetric determination of nitrate is challenging. Most methods involve the reduction of nitrate to nitrite, followed by nitrite quantification using the Griess reaction [18]. Several reduction methods are typically used, among which cadmium (Cd) reduction by passing the sample through a column of copper-plated Cd granules is traditional, but Cd is toxic and the efficiency of the column is reduced over time [84]. Nevertheless, miniaturised Cd reactors have been integrated into microfluidic chips (often continuous flow upstream) [85]. Jayawardane et al. (2014) used zinc microparticles (University of Melbourne, Australia) for on-chip reduction of NO₃⁻ in a µPAD [31]. Vanadium-(III) chloride (VCl₃) reduction is based on the reduction of nitrates to nitrites at elevated temperatures (60–80 °C) [32,86,87]. While avoiding Cd, integrating controlled and stable heating for VCl₃ reduction in portable microfluidic systems can be complex. Using smartphone-based detection, a simple method based on the VCl₃ reduction-Griess reaction was established for on-site NO₃⁻ analysis in natural water samples with various salinities [88]. The proposed method showed a good linearity for concentrations of NO₃⁻ up to 20 µM. While on-chip enzymatic reactors have been explored, enzyme stability, cost, and potential inhibition and hence reduction in reaction stability of matrix components are challenges and sources of uncertainty [89].

Microfluidic portable devices often involve an upstream microreactor for reduction (e.g., a packed bed of Cd or Zn) followed by generation for the Griess reaction [31,85]. Ensuring consistent reduction efficiency over time, especially with natural water particulates that can clog reducers, is a major challenge for long-term autonomous deployments. For example, Plant et al. (2023) developed an in situ UV spectrophotometry-based method for NO₃⁻ analysis in seawater, with updated temperature correction as an important instrumental component in the experimental methodology [90].

Khanfar et al. (2017) developed microfluidic chips to detect nitrate in different water samples [69]. A novel method with a designed PMMA device showed reduced power consumption (~40%) for NO₃⁻ using the Griess reagent at 35 °C [70]. The use of a nutrient sensor deployed within an autonomous underwater vehicle (AUV) to collect NO₃⁻ data in a shelf sea environment showed the possibility of capturing high-frequency events of spring phytoplankton [71]. Rapid NO₃⁻ determination was realised using a portable device based on innovative 3D double micro-structured associated reactors [72] (Figure 7). This system allowed for different water samples’ analysis with low reagent consumption (26.8 µL per sample), good reproducibility, and low RSD (0.5–1.38%).

Figure 7. Portable device for nitrate detection with the main interface of the App on a smartphone and the main body assembled with PDMS chip, printed circuit board (PCB), and PMMA detection chip suitable for the long-term spectrophotometric measurements. Reprinted with permission [72].

Challenges in environmental water analysis comprise reducer longevity and fouling due to particulates and biofilm growth that can rapidly deactivate or clog micro-reducers. Reduction efficiency ensuring complete reduction of NO₃⁻ to NO₂⁻ without over-reduction to NH₄⁺ or N₂O, especially across varying sample salinities and DOM content, is challenging. This is especially the case for low concentrations, achieving sensitive NO₃⁻ detection after dilution by reagent addition given the potential inefficiencies in reduction.

4.4. Total Dissolved Nitrogen (TDN) and Total Nitrogen (TN)

Total dissolved nitrogen (TDN) (e.g., for filtered sample) and total nitrogen (TN) (e.g., for unfiltered sample) analysis involves converting all N (organic and inorganic) compounds into a single measurable form, usually NO₃⁻ or NH₄⁺. This typically requires an initial, aggressive oxidative digestion step [18]. Digestion methods can be performed by alkaline persulphate after heating with potassium persulphate (K₂S₂O₈) and sodium hydroxide (NaOH) often under pressure at higher temperatures (100–120 °C) that converts N to NO₃⁻ [91]. UV irradiation in the presence of an oxidant (e.g., persulphate or titanium dioxide, TiO₂) can oxidise organic N to NO₃⁻ [92,93]. This is often preferred for in situ systems due to less aggressive reagent requirements.

Integrating high pressure and temperature and persulphate digestion into conventional polydimethylsiloxane (e.g., PDMS) microfluidic devices is highly problematic due to material incompatibility and sealing [94]. In contrast, UV-based micro-digesters are more amenable to microfluidic integration [33,95]. Often, digestion is performed off-chip or in a separate robust microfluidic module upstream of NO₃⁻/NO₂⁻ detection. For example, Lin et al. (2021) developed a flow injection analysis (FIA) system using on-line UV digestion for TDN in natural waters [96]. Seamlessly coupling such digesters with downstream droplet systems for aqueous environmental matrices remains an active research area. The major areas of challenge, and hence the primary hurdle in microfluidic systems, are the robust, efficient, and compact on-chip digestion processes for diverse environmental samples and preventing interference from the digestion reagents in subsequent colorimetric steps.

4.5. Organic Nitrogen

Dissolved organic nitrogen (DON) and PON are significant components of the organic N pool in many ecosystems. Direct colorimetric methods for bulk DON/PON in microfluidics are rare. Typically, DON is determined by the difference in TDN (NH₄⁺ + NO₂⁻ + NO₃⁻). Recently, Uhlikova et al. (2024) developed a µPAD for the speciation of dominant inorganic N-species (NH₄⁺ and NO₃⁻) in environmental waters [34]. This system showed stability during 24 h at ambient temperature. Therefore, the development of accurate microfluidic methods for medium- and long-term TDN and dissolved inorganic nitrogen (DIN) monitoring are crucial for DON estimation.

4.6. Multiplexed Detection of Nitrogen Species

Environmental studies often require simultaneous quantification of multiple N-species to understand transformation pathways and nutrient stoichiometry. Strategies for this are to use parallel microfluidic channels, each dedicated to a specific analyte with its distinct colorimetric reaction and detector [97]; sequential reagent addition if chemistries are compatible; or spectral deconvolution if chromophores have distinct absorbance maxima [98]. Nightingale et al. (2019) developed a system for simultaneous monitoring of NO₃⁻ and NO₂⁻ in natural waters [35] (Figure 8). This droplet-based microfluidic sensor device showed high measurement frequency and low fluid consumption, at a rate of 2.8 mL per day.

Figure 8. Photograph of a field-deployable microfluidic sensor (slightly larger than a pen) with integrated fluidics, heater, flow cells, and control electronics for simultaneous monitoring of total nitrate and nitrite in situ [35].

Altahan et al. (2022) demonstrated robust sensors for long-term simultaneous measurements of NO₃⁻, NO₂⁻, phosphate (PO₄³⁻), and silicic acid in seawater [99] (Figure 9). The analyser performed well in the field during a 46-day deployment on a pontoon with a water supply from a depth of 1 m.

Figure 9. Deployment setup of the AutoLAB analyser (Envirotech LLC, Chesapeake, VA, USA) in a weather-proof aluminium container for simultaneous determination of nitrate plus nitrite in seawater (adapted from [99]).

Developing compact, robust microfluidic multiplexed sensors capable of delivering near-real-time and accurate results in field conditions remains a key but challenging technological and scientific goal. Achieving this requires overcoming barriers in fluid control, signal cross-talk, and miniaturised optical integration, all within a compact, low-power platform suitable for decentralised deployment.

5. Advantages and Overcoming Limitations in Environmental Water Analysis

5.1. Enhanced Analytical Capabilities for Environmental Science

Portable and potentially autonomous systems enable denser sampling in space and time, capturing episodic events, diel cycles, and fine-scale heterogeneity in nutrient distributions crucial for ecological modelling [20,100]. Reduced logistical burden for small-scale fieldwork, low power consumption, and minimal reagent requirements simplify deployment in remote locations, e.g., on buoys, AUVs, or for manual transects [55,101]. Lower reagent costs per analysis and potential for automation can make extensive monitoring programs more feasible [21].

5.2. Addressing Complex Environmental Matrix Effects

As explained, applying microfluidic colorimetry to diverse environmental waters offers substantial benefits but also faces unique matrix-specific challenges. Environmental waters are far more variable than controlled laboratory solutions or even treated wastewater [26]. Algae, detritus, and sediments in rivers, lakes, and coastal waters can clog microchannels and cause significant light scattering, interfering with absorbance measurements [79]. DOM, including humic and fulvic acids which are ubiquitous in natural waters, can impart colour (background absorbance) and fluorescence, complex with analytes or reagents, or foul surfaces [79,102,103,104]. Background correction includes using a reference wavelength, sample pre-treatment (e.g., solid-phase extraction for DOM removal, though this adds complexity), development of colorimetric chemistries less susceptible to DOM interference, or advanced chemometric data processing.

Salinity is a major challenge for estuarine and marine applications. High and variable salt content can affect reaction kinetics and equilibria; alter reagent solubility and stability; change the refractive index of the aqueous phase, impacting optical measurements if not properly accounted for in microfluidic systems (e.g., lens effects); or lead to salt precipitation and channel clogging, especially if evaporation occurs or if reagents are incompatible with high salt concentrations. Use of salt-matched calibration standards corresponding to the sample salinity range [28], development of salt-tolerant reagent formulations, careful material selection to avoid corrosion, and robust fluidic design to prevent clogging are some of the most recent issues developed.

Many pristine or oligotrophic waters have very low nutrient levels (nM to low µM range). Achieving adequate sensitivity with short optical pathlengths is demanding. Mitigation of this can be achieved by using highly sensitive chromogenic reagents (high molar absorptivity), optimisation of reaction conditions for maximum colour yield, on-chip pre-concentration techniques (e.g., solid-phase extraction, electrophoresis) prior to coloured-compound formation, or exploration of alternative detection methods with higher intrinsic sensitivity (e.g., fluorescence, though this review focuses on colorimetry).

Biofouling for long-term in situ deployments and biofilm formation on chip surfaces and in tubing can alter flow characteristics, consume analytes, or release interfering substances [105]. This is often mitigated by the use of antifouling coatings, periodic flushing with cleaning agents or biocides, optimised flow regimes, or integration of UV sterilisation.

5.3. Robustness and Field Deployment Considerations

Device durability is important as devices must withstand physical handling, temperature fluctuations, and potentially corrosive environments. Materials like glass or robust polymers such as cyclic olefine copolymer (COC) or polyetheretherketone (PEEK) are often preferred over PDMS for field systems [106]. Critically, power management for autonomous deployments should minimise the power consumption of pumps, valves, detectors, and control electronics [45,55]. Essentially, calibration stability implies maintained calibration accuracy over time and across varying environmental conditions (e.g., temperature changes affecting reaction rates or detector performance). Automated on-chip recalibration routines or temperature compensation mechanisms are needed. Long-term reagent stability and storage of liquid reagents on-chip or in portable cartridges can be an issue. Lyophilised reagents reconstituted on-demand or in situ reagent generation are potential solutions.

6. Future Perspectives and Emerging Trends

6.1. Advanced Materials and Manufacturing

The field of microfluidic colorimetry for environmental N analysis is continually evolving, driven by technological advancements and the growing demand for better environmental intelligence. Beyond PDMS, the use of glass, thermoplastics like COC, and fluoropolymers will increase for field-deployable systems due to better chemical resistance, optical properties, and mechanical robustness [106,107]. Additive manufacturing (3D printing) is enabling rapid prototyping of complex 3D fluidic networks and integrated components [108].

6.2. Enhanced On-Chip Integration and Autonomy

The trend is towards fully autonomous “sample-to-answer” systems. This includes an integrated upstream sample preparation system comprising robust on-chip modules for filtration, particulate removal, dilution, and, importantly for environmental samples, analyte pre-concentration. Autonomous operation includes integration with microcontrollers, low-power pumps and valves, and telemetry for unattended long-term deployments on buoys, moorings, or AUVs [55,101]. This includes self-diagnosis and automated recalibration. “Smart” reagent cartridges are disposable cartridges containing stable reagents, calibration standards, and waste reservoirs to simplify field operation and reduce contamination risks.

6.3. Data Science and Sensor Networks

On-board data processing (edge computing) and machine learning (ML) algorithms allow for real-time quality control, advanced signal processing, interference correction, and predictive alerts for unusual N levels. Data integration includes wireless communication (e.g., satellite) for transmitting data from distributed sensor networks to central databases, enabling large-scale, real-time environmental monitoring and modelling [20,109].

Continued development of low-cost, user-friendly smartphone-based readers are in development for citizen science applications and rapid field screening by non-specialists [22,83].

6.4. Novel Chemistries and Detection Schemes

More sensitive, selective, and matrix-tolerant chromogenic reagents are needed to be specifically tailored for microfluidic analysis of environmental samples (e.g., stable at varying salinities, less prone to DOM interference). Although N-species are dominant in freshwater ecosystems, some other important chemical parameters, in particular P-ions (e.g., PO₄³⁻) or C are limiting nutrients in eutrophication processes and should be analysed simultaneously using a specially designed multiplex microfluidic system. Currently, one of the goals of our research group is the design of new systems that would allow simultaneous monitoring of two or more nutrients (e.g., different N- and P-ions), which is of great importance when it comes to rivers and lakes. While this review focuses on colorimetry, coupling microfluidics with other sensitive detection modes (e.g., electrochemical sensors or highly sensitive fluorescence assays) for ultra-trace analysis of N-species will expand the capabilities of such hybrid technologies in the future [110].

6.5. Standardisation and Commercialisation

Robust, validated, and commercially available microfluidic N sensors are needed for widespread adoption by environmental agencies and researchers. This requires standardised performance metrics and validation protocols ensuring the comparability and reliability of data from different devices and laboratories. For example, the use of internal standards is likely to be required to meet environmental regulatory compliance. Cost-effectiveness implies further reductions in the cost of components, e.g., chips, sensors, and consumables. User-friendly interfaces make devices easy to operate and maintain by non-specialists. Several companies are emerging in the broader field of microfluidic water quality sensors, and N-specific systems are gradually becoming more mature [111,112].

7. Conclusions

Microfluidic colorimetric analysis offers a promising technology that can revolutionise the way we monitor N-species in diverse environmental waterbodies. The inherent advantages of miniaturisation include rapid analysis, dramatically reduced sample and reagent consumption, enhanced portability, and potential for high-throughput automation directly address many limitations of traditional laboratory-based methods. Significant advances have been made in adapting classical colorimetric chemistries for species of NH₄⁺, NO₂⁻, and NO₃⁻ to non-droplet and droplet-based microfluidic platforms, and pioneering efforts are tackling the challenges of on-chip digestion for total N and pre-concentration for ultra-trace analysis.

However, the transition from laboratory prototypes to widely deployed, robust, and autonomous field sensors for complex and variable environmental matrices is ongoing. Key challenges remain, particularly in mitigating interference from suspended solids, DOM, and salinity; ensuring long-term stability and antifouling capabilities for in situ deployments; and achieving the extremely low detection limits often required for oligotrophic waters.

Continued innovation in materials science, microfabrication, integrated sample preparation, reagent stability, data analytics, and smart system design will be crucial in overcoming these hurdles. As microfluidic colorimetric sensors become more sensitive, reliable, cost-effective, and user-friendly, they will empower scientists, environmental managers, and policymakers with unprecedented capabilities for high-resolution spatiotemporal monitoring. This will lead to a deeper understanding of N-species biogeochemistry, improved detection of environmental pollution events, more effective management of aquatic ecosystems, and ultimately, better protection of our vital water resources in a rapidly changing world.