Numerous developing countries in Asia (e.g., India) and Africa (e.g., Nigeria) offer subsidized liquid fuels such as kerosene to support lighting and cooking needs of the rural poor. [1][2][3] The low cost of kerosene compared to market-rate fuels often results in fuel adulteration, i.e., the unauthorized addition of foreign substance into fuel. For example, about 40% of the kerosene sold in India frequently gets blended with gasoline and diesel. [4][5][6] Such fuel adulteration can be as high as 35-50%, and it significantly alters the desired properties of the fuel and leads to substantial economic and environmental concerns. 4,[6][7][8][9][10] Thus far, onsite (e.g., at fuel dispensing stations) detection of fuel adulteration with kerosene has been difficult to quantify because the existing technologies for detecting fuel adulteration (e.g., gas chromatography-mass spectrometry, Fourier transform infrared spectrometry, microcontroller sensor, long period fiber grating, hydrometer, etc.) are either time consuming, expensive, not sensitive enough or require well-equipped analytical laboratories. 6,[11][12][13] So, there is a critical need for inexpensive and easy-to-use devices for rapid and onsite detection of fuel adulteration, especially in developing economies, where fuel quality is a major concern. In this work, we developed an inexpensive and easy-to-use device for rapid and onsite detection of fuel adulteration using smooth and non-polar surfaces. Our device detects fuel adulteration by sensing changes in the mobility of a fuel droplet on a solid surface that in turn depends on the surface tension of the fuel. We have rationally designed our device so that it can detect changes in the mobility of the fuel droplet even at concentrations an order of magnitude below typical adulteration concentrations. We envision that our inexpensive, easy-to-use, and field-deployable device as well as the design strategy will pave the way for novel fuel quality sensors.
The working principle of our fuel adulteration detection is identifying the changes in mobility of fuel droplets on a solid surface. One simple measure of the mobility of a fuel droplet on a solid surface is the sliding angle ω (i.e., the minimum angle by which the surface must be tilted relative to the horizontal for a liquid droplet to slide off from the surface). Sliding angle ω can be estimated based on a balance between work done by gravitational force and work expended due to adhesion as [14][15][16][17][18][19][20][21] gV D sin (cos cos )
lv TCL rec adv (1) Here, D TCL is the width of solid-liquid-vapor contact line perpendicular to the sliding direction, γ lv and ρ are the surface tension and density of the liquid, respectively, g is the acceleration due to gravity, V is the volume of the liquid droplet, θ adv is the advancing (maximum) contact angle, and θ rec is the receding (minimum) contact angle. Typically, droplets of low surface tension liquids (e.g., alkanes, alcohols etc.) display low sliding angles on non-textured (i.e., smooth) and non-polar solid surfaces. 22,23 Now, let us consider alkane droplets of a constant volume V on non-textured and non-polar solid surfaces.
An alkane droplet with a lower surface tension γ lv,low displays lower sliding angle ω low , and an alkane droplet with a higher surface tension γ lv,high displays higher sliding angle ω high (i.e., for γ lv,low < γ lv,high , ω low < ω high ). When the droplets of both lower and higher surface tension alkanes are placed on a solid surface, for a tilt angle α < ω low , both the alkane droplets remain adhered to the surface with no motion; for ω low < α < ω high , the alkane droplet with lower surface tension γ lv,low slides past the surface, but the alkane droplet with higher surface tension γ lv,high remains adhered to the surface; and for α > ω high , both the alkane droplets slide past the surface. In a similar manner, consider droplets of a fuel blend (predominantly consisting of alkanes) with a constant volume V. The sliding angle of a fuel blend droplet decreases with decreasing surface tension, which in turn depends on its chemical composition. When a market-rate fuel (e.g., diesel with γ lv = 25.8 mN m -1 ) is adulterated by blending with another subsidized fuel (e.g., kerosene with γ lv = 20.1 mN m -1 ), its chemical composition and consequently its surface tension changes. Correspondingly, the sliding angle of the adulterated fuel blend changes compared to the unadulterated fuel. Leveraging such differences in sliding angles of fuel blend droplets, we designed our device to detect fuel adulteration.
To systematically design a fuel adulteration detection device, we first studied the mobility of droplets on non-textured (i.e., smooth) substrates with different non-polar (hydrocarbon and fluorocarbon) chemistries. To begin with, we chose silicon wafers (R rms = 1.4 ± 0.5 nm) as our non-textured substrates because they display low surface roughness, and they can be easily modified via silanization to impart non-polar surface chemistry. 24 We modified the surface chemistry of silicon wafers by covalently attaching hydrocarbon brushes (with octadecyltrichlorosilane, OTS, thickness ≈ 1 nm) or fluorocarbon brushes (with heptadecafluoro-1,1,2,2-tetrahydrodecyl trichlorosilane, FDTS, thickness ≈ 1 nm) via liquid phase silanization; the surface roughness R rms of silicon wafers did not change significantly upon surface modification with OTS or FDTS (see Supporting Information, Section S1). We chose two different non-polar surface chemistries to investigate because we anticipate that the differences in chain flexibility of hydrocarbon and fluorocarbon brushes will influence the mobility of droplets, 22,25 which plays a critical role in designing our fuel adulteration detection device. Typically, hydrocarbon brushes display higher chain flexibility compared to their fluorocarbon analogues, resulting in higher droplet mobility on a solid surface. [26][27][28][29][30] We characterized the surface chemistry of our unmodified, OTS-modified, and FDTS-modified silicon wafers using X-ray photoelectron spectroscopy (see Supporting Information, Section S2).
We investigated the mobility of droplets on OTS-modified and FDTS-modified silicon wafers using multiple liquid systems [i.e., water-sodium dodecyl sulfate (SDS) mixtures, waterethanol mixtures, and homologous series of alkanes] by measuring sliding angles as a function of droplet volume. Our OTS-modified silicon wafers displayed high droplet mobility (see Figure 1a-c and see Supporting Information, Section S3) with both high surface tension liquids (e.g., water, γ lv = 72 mN m -1 , see Figure 1d) and low surface tension liquids (e.g., ndecane, γ lv = 20 mN m -1 , see Figure 1e). For each liquid, as the droplet volume decreased, the sliding angles increased (i.e., droplet mobility decreased), in accordance with eq 1. 31,32 More importantly, for any two liquids within each liquid system (i.e., water-SDS mixtures, water-ethanol mixtures, or homologous series of alkanes), as the droplet volume decreased, the difference in droplet sliding angles increased (see Figure 1a-c and Table S1). This indicates that lower droplet volumes allow for more noticeable differences in droplet mobility between any two liquids studied (even with small changes in surface tension or composition) within a liquid system. For example, on OTSmodified silicon wafers, at 20 μL droplet volume, the sliding angles of water-SDS system spanned across a range of 4°(i.e., difference in droplet sliding angles of water and water + 6 mM SDS). When the droplet volume decreased to 2 μL, this range in
Article sliding angles increased to 61°(see Figure 1a). Similar increasing trend in the range of sliding angles with decreasing droplet volume was also observed with the water-ethanol system (see Figure 1b) and alkanes (see Figure 1c). Such higher ranges in sliding angles indicate that the changes in droplet mobility are more distinct at lower droplet volumes for each liquid system. This in turn allows sensing changes in the liquid composition at finer levels of adulteration enabling reliable detection.
Our FDTS-modified silicon wafers also displayed high droplet mobility (see Figure 2a-c and see Supporting Information, Section S3) with both high surface tension liquids (e.g., water, see Figure 2d) and low surface tension liquids (e.g., n-decane, see Figure 2e). As anticipated, all liquids also displayed increasing sliding angles (i.e., decreasing droplet mobility) with decreasing the droplet volume. The sliding angles on FDTS-modified silicon wafers were slightly higher compared to OTS-modified silicon wafers. For example, a 20 μL water droplet displayed a sliding angle of 16°on FDTS-modified silicon wafer (see Figure 2a) and 10°on OTS-modified silicon wafer (see Figure 1a). This increase in the sliding angle (i.e., decrease in droplet mobility) is due to the lower chain flexibility of FDTS brushes compared to OTS brushes on modified silicon wafers. 22,[25][26][27][28]30 Similar to the OTS-modified silicon wafers, FDTS-modified silicon wafers also displayed an increasing trend in the range of sliding angles with decreasing droplet volume for each liquid system (see Figure 2a-c and Table S1). This reaffirms that difference in droplet mobility is more distinct at lower droplet volumes, which enables reliable fuel adulteration detection. Overall, our droplet mobility results on OTSmodified silicon wafers (see Figure 1) and FDTS-modified silicon wafers (see Figure 2) indicate that either of these nonpolar surface chemistries is suitable for designing a fuel adulteration detection device when lower droplet volumes are used. For example, for 2 μL droplets on OTS-modified silicon wafers, comparing the sliding angles of water + 50% ethanol (15°) with those of water (85°) and ethanol (6°) clearly indicates the presence of ethanol in water (see Figure 1b). Similarly, for 2 μL droplets on FDTS-modified silicon wafers, comparing the sliding angles of water + 50% ethanol (47°) with those of water (no sliding) and ethanol (20°) clearly indicates the presence of ethanol in water (see Figure 2b).
While we could have used FDTS-modified substrates to design our device for detecting fuel adulteration, we chose to use OTS-modified substrates because hydrocarbon ligands are environmentally friendly compared to their fluorocarbon analogues, which have raised worldwide environmental concerns. [33][34][35] To develop a device that can detect the adulteration of diesel with kerosene, we first characterized the droplet mobility of diesel-kerosene blends on OTS-modified silicon wafers as a function of blend concentration and droplet volume (see Figure 3a). As the kerosene concentration in the diesel-kerosene blend varied from 0% (only diesel) to 100% (only kerosene) by volume, the surface tension of the blend varied from 25.8 to 20.1 mN m -1 (see Supporting Information, Section S4). As anticipated, our OTS-modified silicon wafers displayed lower sliding angles with higher kerosene concentrations (i.e., lower surface tension blends; see Figure 3a and Supporting Information, Section S4), indicating higher droplet mobility with increasing kerosene concentration in the blend. However, the sliding angle differences across the dieselkerosene blends were too low. For example, 5 μL droplets of
Article diesel and diesel + 10% kerosene displayed sliding angles of 11 and 10°, respectively, on OTS-modified silicon wafers. Such low differences in sliding angles make it practically difficult to detect changes in mobility due to changes in the composition, thereby leading to unreliable detection.
To obtain larger differences in droplet sliding angles across diesel-kerosene blends on OTS-modified surfaces, we used glass slides, which have higher surface roughness (R rms = 5.2 ± 1.3 nm) compared to silicon wafers. We projected that the higher surface roughness of OTS-modified glass slides will result in higher contact angle hysteresis, [36][37][38] which in turn may magnify the differences in droplet sliding angles across dieselkerosene blends (see Supporting Information, Section S5), thereby allowing reliable detection of adulteration. Furthermore, compared to silicon wafers, glass slides are inexpensive, which makes our device cost-effective and improves its scalability. Indeed, our OTS-modified glass slides displayed a wider range in the sliding angles across the diesel-kerosene blends (see Figure 3b and Table S2) at all droplet volumes compared to OTSmodified silicon wafers (see Figure 3a and Table S2). For example, 5 μL droplets of diesel and diesel + 10% kerosene displayed sliding angles of 25 and 18°, respectively, on OTSmodified glass slides. This indicates a significantly larger difference in the droplet sliding angles on rougher OTSmodified glass substrates compared to smoother OTS-modified silicon substrates. Such larger differences in sliding angles across diesel-kerosene blends will produce more noticeable changes in droplet mobility. In turn, this reduces the error and allows for reliable detection of diesel adulteration by kerosene.
Based on the droplet mobility results of diesel-kerosene blends on OTS-modified glass slides, we fabricated a simple, inexpensive, and field-deployable device using 3D printing (see Supporting Information, Section S6 and Figure S3) to demonstrate rapid fuel adulteration detection. Our fuel adulteration detection device consists of multiple wedges that can accommodate OTS-modified substrates at a wide variety of tilt angles α (see Figures 3c and S3). When liquid droplets of an unknown composition (diesel or diesel-kerosene blends) and same volume are placed on wedges with different tilt angles α, the droplets would remain adhered when sliding angle ω > α and the droplets would slide when the sliding angle ω < α. As demonstrated previously (see Figure 3a,b), for diesel-kerosene blends, the sliding angle ω decreases with increasing kerosene concentration (i.e., droplet mobility increases with increasing adulteration). So, by placing droplets of a diesel-kerosene blend with unknown composition and fixed volume on wedges with different tilt angles and identifying the mobility of the droplet (i.e., the tilt angle at which the droplet slides) and comparing it with the sliding angles of diesel-kerosene blends on OTS substrates (see Figure 3b), we can estimate the composition diesel-kerosene blend. To demonstrate this (see Figure 3d-f), we used 5 μL droplets of different diesel-kerosene blends on our device with wedges having 15, 20, and 25°tilt angles. We chose 5 μL droplets of diesel-kerosene blends because they not
Article only displayed a sufficiently large difference in sliding angles (i.e., 17°difference between diesel and kerosene) but also displayed sufficiently large sliding velocities. Unlike 5 μL droplets, 20 μL droplets displayed low differences in sliding angles (i.e., 5°d ifference between diesel and kerosene), making it difficult to differentiate between different diesel-kerosene blends based on droplet sliding. Even though 2 μL droplets displayed much larger differences in sliding angles (i.e., 26°difference between diesel and kerosene) compared to 5 μL droplets, their lower sliding velocities led to longer detection times, making them impractical for rapid detection. Furthermore, smaller droplet volumes display higher evaporation rates (due to higher surfaceto-volume ratio), [39][40][41] which may produce inconsistent results due to changes in droplet volume, especially with volatile fuels like diesel and kerosene.
To demonstrate rapid fuel adulteration detection, we sequentially placed 5 μL droplets of diesel-kerosene blends with 5, 10, and 25% kerosene concentrations on our device with OTS-modified glass slides on wedges having 15, 20, and 25°tilt angles. The droplets of diesel + 5% kerosene remained adhered on wedges with 15 and 20°tilt angles, while they slid down the wedge with 25°tilt angle. This is because the sliding angle of diesel + 5% kerosene droplets (ω = 21°) on OTS-modified glass slides is higher than 15 and 20°(i.e., ω > α) but lower than 25°( i.e., ω < α; see Figure 3d). The droplets of diesel + 10% kerosene (γ lv = 23.9 mN m -1 ) remained adhered on the wedge with 15°tilt angle, while they slid down the wedges with 20 and 25°tilt angles. This is because the sliding angle of diesel + 10% kerosene droplets (ω = 18°) on OTS-modified glass slides is higher than 15°(see Figure 3e). The droplets of diesel + 25% kerosene (γ lv = 22.3 mN m -1 ) slid down all the three wedges because their tilt angles are higher than sliding angle of diesel + 25% kerosene blend (ω = 12°) on OTS-modified glass slides (see Figure 3f and Movie S1). Thus, leveraging our knowledge of diesel-kerosene blend droplet mobility on OTS-modified glass slides, we can easily detect and approximately estimate diesel adulteration even at kerosene concentrations as low as 5%. To put this into context, diesel adulteration is typically as high as 35-50% kerosene. 7,8 In a similar manner, this strategy can be extended to other liquid systems for developing inexpensive and field-deployable devices that can rapidly detect compositional changes. Furthermore, we investigated the durability (i.e., retention of slipperiness) of OTS-modified surfaces by measuring sliding angles of 20 μL of water and diesel droplets on OTS-modified glass slides after exposing the surfaces to sunlight for 1 day, to steam (at 100 °C and 1 atm) for 1 day, and sliding thousands of water and diesel droplets past the surface (see Supporting Information, Section S7). Our results indicate that OTSmodified surfaces retained their slipperiness even after 1 day exposure to sunlight (see Figure S4a) and 1 day exposure to steam (see Figure S4b), and even after 100,000 water and diesel droplets slid past the surface (see Figure S4c). While these results demonstrate the reusability of our OTS-modified surfaces, a more comprehensive study is necessary to exhaustively quantify the durability of these surfaces for their practical use in detecting fuel adulteration.
In this work, we developed a simple, inexpensive, and fielddeployable device to rapidly detect fuel adulteration. The working principle of our device is identifying the changes in mobility of fuel droplets on non-textured and non-polar solid surfaces. We elucidated the working principle of our device by investigating droplet mobility of multiple liquid systems (i.e., water-SDS mixtures, water-ethanol mixtures, and homologous series of alkanes), including diesel-kerosene blends on nontextured and non-polar solid surfaces with different chemical (i.e., surface chemistry) and physical (i.e., surface roughness) properties. We demonstrated a rapid detection of fuel (i.e., diesel) adulteration even at subsidized fuel (i.e., kerosene) concentrations, an order of magnitude below typical adulteration concentrations. We envision that our inexpensive, easy-touse, and field-deployable device as well as the design strategy will pave the way for novel fuel quality sensors.
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.langmuir.3c00578. Materials and methods; characterization of surface chemistry; sliding angles of different liquid systems on OTS-and FDTS-modified silicon wafers; sliding angles of diesel-kerosene blends on OTS-modified substrates; contact angles and contact angle hysteresis of dieselkerosene blends; fabrication of 3D-printed fuel adulteration detection devices; and durability of OTS-modified surfaces (PDF)
Mobility of 5 μL diesel-kerosene blend droplets on OTSmodified glass slides tilted at 15, 20, and 25°on our fuel adulteration detection device (MP4)
Fabrication of OTS-modified and FDTS-modified surfaces: Silicon wafers (<1 0 0> orientation; University Wafers) and glass slides (Fisher Scientific) were cleaned by rinsing thoroughly with acetone (Fisher Scientific) and DI water, and then dried with nitrogen. The cleaned substrates were exposed to oxygen plasma (Plasma Etch PE-25) for 15 min for hydroxylation. To prepare OTSmodified surfaces, the hydroxylated substrates were immersed in a solution consisting of 15 mL toluene (Fisher Scientific), 12 L of hydrochloric acid (Fisher Scientific), and 16 L of octadecyltrichlorosilane (OTS; Gelest) for 2 hours at room temperature (see Figure S1). To prepare FDTS-modified surfaces, the hydroxylated substrates were immersed in a solution consisting of 15 mL toluene (Fisher Scientific), 12 L of hydrochloric acid (Fisher Scientific), and 16 L of heptadecafluoro-1,1,2,2-tetrahydrodecyl trichlorosilane (FDTS; Gelest) for 2 hours at room temperature (see Figure S1). Finally, the silanized samples were cleaned by rinsing thoroughly with anhydrous toluene, DI water and ethanol (Fisher) sequentially, and then dried with nitrogen. Contact angle, sliding angle and surface tension measurements: The contact angles and sliding angles were measured with sessile droplets and liquid surface tension was measured with pendant droplets using a contact angle goniometer/tensiometer (Ramé-Hart 200-F1). At least three Ellipsometry measurements: The thickness of OTS and FDTS brush layers on our surfaces were determined using a variable angle spectroscopic ellipsometer (VASE-VB-250). A spectral scan of the surface was collected between 500 nm and 900 nm for an incident angle between 55° and 75° with an increment of 5°. The thickness of OTS/FDTS brush layers (refractive index j 1.46) were determined using a three-layer planar model (air/brush/silica). For all our ellipsometry data, at least three measurements were conducted on different samples. Our ellipsometry results indicate that the average thickness of both OTS and FDTS-modified surfaces is j 1 nm.
The root mean square roughness Rrms of the silicon wafers and glass slides was measured using an optical profilometer (Zygo Zescope). At least three measurements were performed on each surface at spatially distinct locations. The standard deviation is reported as error with all measurements. Our optical profilometry results indicate that the Rrms of our surfaces did not change significantly upon surface modification with OTS or FDTS.
The surface chemistry of substrates was characterized using X-ray photo-electron spectroscopy (XPS; Physical Electronics PHI-5800 spectrometer). XPS was conducted using a monochromatic Al Ka X-ray source operated at 15 kV, and photoelectrons were collected at a takeoff angle of 45° relative to the sample surface. XPS data was acquired from at least 5 spatially different locations on the surface and the spectral analysis was conducted using PHI Multipak software.
The sliding angles of water-SDS mixtures, water-ethanol mixtures, and homologous series of alkanes with different droplet volumes on OTS and FDTS-modified silicon wafers are reported in Table S1. For all sliding angles, the measurements were conducted on at least three different samples and the standard deviation is reported as the error with all measurements.
Table S1. Sliding angles of 20 L, 10 L, and 2 L droplets of water-SDS mixtures, water-ethanol mixtures, and homologous series of alkanes on OTS and FDTS-modified silicon wafers Liquid Droplet volume ( L) Sliding angle (°) OTS FDTS Water 20 10±0.5 16±0.8 10 15±0.5 25±0.8 2 85±2.5 No sliding 1 mM SDS 20 10±0.5 13±0.8 10 15±0.5 23±0.8 2 69±2 81±2.5 2 mM SDS 20 10±0.5 11±0.8 10 14±0.5 19±0.8 2 54±2 69±2 3 mM SDS 20 9±0.5 10±0.8 10 14±0.5 17±0.8 2 48±1.5 53±2 4 mM SDS 20 9±0.5 9±0.5 Table S1 (contd.). Sliding angles of 20 L, 10 L, and 2 L droplets of water-SDS mixtures, water-ethanol mixtures, and homologous series of alkanes on OTS and FDTS-modified silicon wafers Liquid Droplet volume ( L) Sliding angle (°) OTS FDTS 4 mM SDS 10 13±0.5 17±0.5 2 35±1.5 46±2 5 mM SDS 20 8±0.5 8±0.5 10 13±0.5 15±0.8 2 28±1.5 33±1.5 6 mM SDS 20 6±0.5 8±0.5 10 11±0.5 13±0.8 2 24±1.5 30±1.5 Water+20% Ethanol 20 9±0.5 14±0.8 10 13±0.8 23±0.8 2 47±2 85±2.5 Water+40% Ethanol 20 8±0.5 14±0.5 10 11±0.8 22±0.8 2 26±2 57±2 Water+50% Ethanol 20 6±0.5 13±0.8 10 8±0.5 22±0.8 2 15±1.5 47±1.5 Water+60% Ethanol 20 4.5±0.5 10±0.8 Table S1 (contd.). Sliding angles of 20 L, 10 L, and 2 L droplets of water-SDS mixtures, water-ethanol mixtures, and homologous series of alkanes on OTS and FDTS-modified silicon wafers Liquid Droplet volume ( L) Sliding angle (°) OTS FDTS Water+60% Ethanol 10 6±0.47 21±0.8 2 10±0.8 36±1.5 Water+80% Ethanol 20 3.5±0.5 8.5±0.5 10 5±0.5 19±0.8 2 8±0.5 29±1.5 Ethanol 20 3±0.5 7±0.5 10 4±0.5 15±0.8 2 6±0.5 20±1.5 Hexadecane 20 2±0.5 13±0.8 10 4±0.5 23±0.8 2 10±0.8 No sliding Tetradecane 20 1.5±0.5 8±0.5 10 3±0.5 17±0.8 2 6±0.8 79±2.5 Dodecane 20 1±0.5 7±0.5 10 2±0.5 15±0.8 2 5±0.8 56±2 Decane 20 0.5±0.5 7±0.5 Table S1 (contd.). Sliding angles of 20 L, 10 L, and 2 L droplets of water-SDS mixtures, water-ethanol mixtures, and homologous series of alkanes on OTS and FDTS-modified silicon wafers Liquid Droplet volume ( L) Sliding angle (°) OTS FDTS Decane 10 1.5±0.5 14±0.8 2 4±0.5 48±1.5 Octane 20 0.5±0.5 5±0.5 10 0.5±0.5 11±0.8 2 4±0.5 41±1.5
The sliding angles of diesel-kerosene blends with different droplet volumes on OTS-modified silicon wafers and OTS-modified glass slides are reported in Table S2. For all sliding angles, the measurements were conducted on at least three different samples and the standard deviation is reported as the error with all measurements.
Table S2.
https://doi.org/10.1021/acs.langmuir.3c00578Langmuir 2023, 39, 9044-9050