sensors Article
A Novel Microfluidic Flow Rate Detection Method Based on Surface Plasmon Resonance Temperature Imaging Shijie Deng, Peng Wang *, Shengnan Liu, Tianze Zhao, Shanzhi Xu, Mingjiang Guo and Xinglong Yu State Key Laboratory of Precision Measurement Technology and Instruments, Tsinghua University, Beijing 100084, China; [email protected] (S.D.); [email protected] (S.L.); [email protected] (T.Z.); [email protected] (S.X.); [email protected] (M.G.); [email protected] (X.Y.) * Correspondence: [email protected]; Tel.: +86-10-6277-2007 Academic Editors: Amine Miled and Jesse Greener Received: 29 April 2016; Accepted: 21 June 2016; Published: 24 June 2016
Abstract: A novel microfluidic flow rate detection method based on surface plasmon resonance (SPR) temperature imaging is proposed. The measurement is performed by space-resolved SPR imaging of the flow induced temperature variations. Theoretical simulations and analysis were performed to demonstrate a proof of concept using this approach. Experiments were implemented and results showed that water flow rates within a wide range of tens to hundreds of µL/min could be detected. The flow rate sensor is resistant to disturbances and can be easily integrated into microfluidic lab-on-chip systems. Keywords: microfluidic flow rate detection; surface plasmon resonance; temperature variation imaging
1. Introduction There is an increasing demand for microfluidic flow rate measurement technologies in various research fields, such as bio-chemical analysis, sample mixing, and environmental monitoring. Various microfluidic flow sensors of different working principles have been proposed in recent years [1–5]; the thermal flow rate sensors working on heat transfer property under the liquid flow, are the most prevalent [6,7]. According to the detection method, thermal flow rate sensors can be classified into three categories: hot element anemometers [8,9], calorimetric flow rate sensors [10,11] and time of flight sensors [12,13]. Calorimetric flow rate sensors which use thermistors that are laid either up- or down-stream of a heater to detect the flow induced temperature distribution, are of the highest sensitivity. The layout of the heater and thermistors is a key design factor that determines the measurement sensitivity and dynamic range [14–16]. Once the distance between the heater and thermistors is determined, the measurable flow is limited to a specific dynamic range inversely proportional to sensibility. To expand the dynamic range and enhance the sensitivity, improvements have been implemented to optimize the arrangement of the heater and thermistors. Several temperature detectors were utilized in the works by N. Sabaté [17] and Seunghyun Kim [18]. The introduction of an array of thermistors facilitated a large expansion of the dynamic range. With the use of fluorescent dyes, Hoera et al. [19] developed a flow sensor based on temperature distribution imaging, that had a wide dynamic range of 10 nL/min to 100 µL/min. Unfortunately, the use of fluorescent dyes impose restrictions on the flow rate sensor in non-labeled detection fields. Surface plasmon resonance (SPR) is a collective charge density oscillation that occurs at a metal-dielectric interface. As SPR is extremely sensitive to the temperature-dependent refractive Sensors 2016, 16, 964; doi:10.3390/s16070964
www.mdpi.com/journal/sensors
Sensors 2016, 16, x
2 of 10
Surface plasmon resonance (SPR) is a collective charge density oscillation that occurs at a 2 of 10 interface. As SPR is extremely sensitive to the temperature-dependent refractive index (RI) just above the metal film, it has extensive applications in temperature monitoring and sensing [20–22]. In work performed et applications al., a variation as small as 0.027 K and was index (RI) just above the metal film, it by has Chiang extensive in temperature monitoring distinguished at 632.8 nm [23]. Furthermore by means of SPR imaging, the refractive index sensing [20–22]. In work performed by Chiang et al., a variation as small as 0.027 K was distinguished distribution on the whole sensing surfaceofcan simultaneously monitored. When liquidon flows at 632.8 nm [23]. Furthermore by means SPRbe imaging, the refractive index distribution the through the sensing unit, the temperature above the sensing surface is redistributed. Because the whole sensing surface can be simultaneously monitored. When liquid flows through the sensing refractive index (RI) is inversely proportional to the temperature, the RI varies along the flow unit, the temperature above the sensing surface is redistributed. Because the refractive index (RI) is channel, proportional leading to ato corresponding variation of thealong SPRthe response. Therefore, the flow rate inversely the temperature, the RI varies flow channel, leading to liquid a corresponding can be resolved from the SPR response variations. variation of the SPR response. Therefore, the liquid flow rate can be resolved from the SPR In this paper, SPR imaging was proposed for the first time to measure the microfluidic flow response variations. rate. InAs the space-resolved able the flow-induced continuous this paper, SPR imagingSPR wasimaging proposedisfor the to firstmonitor time to measure the microfluidic flow temperature variations, rather than several discrete points, this method has more resistance to rate. As the space-resolved SPR imaging is able to monitor the flow-induced continuous temperature disturbances andthan potentially has a wide detection range. Asmore SPRresistance has beentowidely appliedand in variations, rather several discrete points, this method has disturbances bio-chemical analysis, the SPR-based flow rate measurement will be facilely integrated into these potentially has a wide detection range. As SPR has been widely applied in bio-chemical analysis, the analytical systems. SPR-based flow rate measurement will be facilely integrated into these analytical systems. Sensors 2016, 16, 964 metal-dielectric
2. Theory 2. Theoryand andSimulations Simulations 2.1. 2.1. Principle The The schematic schematic of of microfluidic microfluidic flow flow rate rate measurement measurement based based on on SPR SPR temperature temperature imaging imaging is is illustrated 1. The SPRSPR prism coated with awith gold afilm, together with a flow channel, constitutes illustratedininFigure Figure 1. The prism coated gold film, together with a flow channel, the sensing chip. The SPR chip. prism The and the are loaded with high are and low temperature, respectively. constitutes the sensing SPRwater prism and the water loaded with high and low A collimated beam of light projects onto the bottom of the gold film to the initiate SPR.of When the water is temperature, respectively. A collimated beam of light projects onto bottom the gold film to still, the SPR response (intensity of the reflected beam) is uniform. But when the water is flowing, there initiate SPR. When the water is still, the SPR response (intensity of the reflected beam) is uniform. will a temperature Sincegradient the RI isalong inversely proportional the But be when the water isgradient flowing,along therethe willflow be achannel. temperature the flow channel.to Since temperature, the RIproportional decreases along thetemperature, flow channel,the leading to a corresponding decrease of the SPR the RI is inversely to the RI decreases along the flow channel, leading response. The decreasing rateofvaries withresponse. the flow rate. As such, the liquid flow ratethe canflow be resolved to a corresponding decrease the SPR The decreasing rate varies with rate. As from of the such,the thevariation liquid flow rateSPR can response. be resolved from the variation of the SPR response.
Figure 1. 1. Schematic of the the surface surface plasmon plasmon resonance resonance (SPR)-based (SPR)-based microfluidic microfluidic flow flow rate ratedetection. detection. Figure Schematic of
2.2. Flow-Induced Flow-Induced Temperature TemperatureVariations Variations 2.2. The flow-thermal flow-thermal coupling coupling physical physical field field of of the the sensing sensing chip chip was was simulated simulated using using COMSOL. COMSOL. The Since SPR is only able to detect the changes above the upper surface of the prism, the flow channel Since SPR is only able to detect the changes above the upper surface of the prism, the flow channel was placed placed above above the the prism. prism. For by the the top top was For fluid fluid sealing, sealing, the the flow flow channel channel was was pressed pressed down down by Polymethylmethacrylate (PMMA) shell to seal with the bottom prism. Therefore, the flow channel Polymethylmethacrylate (PMMA) shell to seal with the bottom prism. Therefore, the flow channel was sandwiched sandwiched between between aa top top shell shell and and aa bottom bottom SPR SPR prism, prism, as as illustrated illustrated in in Figure Figure 2A. 2A. In In the the was case where the water and prism were at 25 and 35 °C, respectively, the flow (100 μL/min) induced ˝ case where the water and prism were at 25 and 35 C, respectively, the flow (100 µL/min) induced temperaturedistribution distributionof of sensing is shown in 2A. Figure Due to the temperature temperature thethe sensing chipchip is shown in Figure Due2A. to the temperature difference, difference, heat exchange occurred between the liquid and prism. Figure 2B shows the temperature heat exchange occurred between the liquid and prism. Figure 2B shows the temperature variations at different flow rates of 100, 200, 400, 600, 800 and 1000 µL/min. Due to the increase of heating time,
Sensors 2016, 16, x Sensors 2016, 16, x
3 of 10 3 of 10
Sensors 2016, at 16, different 964 3 of of 10 variations flow rates of 100, 200, 400, 600, 800 and 1000 μL/min. Due to the increase variations at different flow rates of 100, 200, 400, 600, 800 and 1000 μL/min. Due to the increase of heating time, the liquid temperature increases approximately linearly along streams, except at the heating time, the liquid temperature increases approximately linearly along streams, except at the inlet and outlet where the flow was unstable. Figure 2C further shows that at slower flow rates, the the liquid temperature increases approximately linearly along streams, the inlet and outlet inlet and outlet where the flow was unstable. further shows except that increase. at at slower flow rates, the liquid maintained a higher temperature and aFigure slower2Crate of temperature Therefore, the where the flow was unstable. Figure 2C further shows that at slower flow rates, the liquid maintained liquid maintained a higher temperature and a slower rate of temperature increase. Therefore, the section of the flow channel at 30–36 mm (between the two dashed lines) was chosen for flow rate asection higheroftemperature and a slower of(between temperature increase. Therefore, thechosen sectionfor of flow the flow the flow channel at 30–36rate mm the two dashed lines) was rate measurement. channel at 30–36 mm (between the two dashed lines) was chosen for flow rate measurement. measurement.
Figure 2. (A) Heat map of the SPR sensing chip. (B,C) the water temperature variations in the flow Figure 2.2. (A) Heat map map of the SPR SPR sensing sensing chip. chip. (B,C) (B,C) the the water water temperature temperature variations variations in in the the flow flow channel at(A) different flowof rates. Figure Heat the channel at different flow rates. channel at different flow rates.
2.3. Temperature-Induced SPR Response Variations 2.3. Temperature-Induced SPR Response Variations 2.3. Temperature-Induced SPR Response Due to the dependence of theVariations refractive index on the temperature, the flow-induced Due to the dependence ofvariation the refractive index on the Based temperature, the flow-induced temperature gradient generates a of theon SPR the multi-layer model, Due to the dependence of the refractive index theresponse. temperature, theon flow-induced temperature temperature gradient generates a variation of the SPR response. Based on the multi-layer model, Fresnel were usedoftothe analyze the SPR response. Figure depicts themodel, corresponding gradientequations generates[24] a variation SPR response. Based on the 3multi-layer Fresnel Fresnel equations [24] were used to analyze the SPR response. Figure 3 depicts the corresponding variations of the index the related (RIRF). The RIRF variation shows an inverse equations [24] wererefractive used to analyze SPR factor response. Figure 3 depicts the corresponding variations variations ofwith thethe refractive index related factor As (RIRF). The varies RIRF variation showslinearly an inverse relationship liquid temperature variation. the RIRF approximately and of the refractive index related factor (RIRF). The RIRF variation shows an inverse relationship with relationship with the liquid temperature variation. As the RIRF varies approximately linearly and its rate ΔRIRF/Δl changesAs at the different the ΔRIRF/Δl can beand used quantifyrate the thevariation liquid temperature variation. RIRF flow variesrates, approximately linearly itstovariation its variation rate ΔRIRF/Δl changes at different flow rates, the ΔRIRF/Δl can be used to quantify the flow rate. changes at different flow rates, the ∆RIRF/∆l can be used to quantify the flow rate. ∆RIRF/∆l flow rate.
Figure Figure 3. 3. SPR SPR variations variations in in the the flow flow channel channel at at different different flow flow rates. rates. Figure 3. SPR variations in the flow channel at different flow rates.
Sensors 2016, 16, 964 Sensors 2016, 16,16, x x Sensors 2016,
4 of 10
of 10 4 of410
2.4.2.4. Effect of of thethe Liquid-Prism Temperature Difference 2.4. Effect of the Liquid-Prism Effect Liquid-PrismTemperature TemperatureDifference Difference The temperature difference between the liquid liquid and is prism is a a key key factor affecting The temperature difference between and prism factor affecting thethe The temperature difference between the liquid and prism a keyisfactor affecting the prism-liquid prism-liquid heatexchange, exchange, leading to different different temperature variation ToToresearch thethe effect prism-liquid leading to temperature variation rates. research effect heat exchange,heat leading to different temperature variation rates. To researchrates. the effect of the liquid-prism ˝ of the liquid-prism temperature on the measurement, the water temperature was fixed at 25 °C °C of the liquid-prism temperature on the measurement, the water temperature was fixed at 25 temperature on the measurement, the water temperature was fixed at 25 C while the prism ˝ while the prism temperature was changed to 30, 35 and 40 °C, respectively. The relationship while the prism temperature changed 35 and 40relationship °C, respectively. The temperature was changed to 30,was 35 and 40 C, respectively. The between therelationship ∆RIRF/∆l between theΔRIRF/Δl ΔRIRF/Δl and the logarithm of flow 4. rate is ininFigure 4.4. ItItcan that thethe between the and the flow rate is shown shown Figure canbebeseen seen that and the logarithm of flow rate islogarithm shown in of Figure It can be seen that the detection range is about detection range is about 100–500 μL/min, with the linear range also being approximately detection range is about 100–500 μL/min, with the linear range also being approximately 100–500 µL/min, with the linear range also being approximately 100–500 µL/min. When the fluid is 100–500 μL/min. Whenthe the fluid isis faster faster than 500 μL/min, the variations in in theare 100–500 μL/min. When the fluid than 500 μL/min, the temperature temperature variations the faster than 500 µL/min, temperature variations in the monitoring section of the flow channel monitoring section of the flow channel are hardly distinguishable. Therefore the upper limit of the monitoring section of the flow channel are limit hardly distinguishable. the upper limitlimit of the hardly distinguishable. Therefore the upper of the linear range isTherefore 500 µL/min. The lower of linear range is 500 μL/min. The lower limit of 100 μL/min was chosen due to the performance of the linear range was is 500 μL/min. lower limit of 100ofμL/min was chosen dueand to the of the 100 µL/min chosen dueThe to the performance the pumping system, notperformance due to limitations pumping system, and not due to limitations of the sensor. With larger liquid-prism temperature pumping system, notliquid-prism due to limitations of thedifferences, sensor. With liquid-prism temperature of the sensor. Withand larger temperature thelarger sensitivity of the measurement differences, the sensitivity of the measurement was enhanced. differences, the sensitivity of the measurement was enhanced. was enhanced.
Figure 4. Relationship between the ΔRIRF/Δl and flow rate at different water-prism temperature differences. Figure Relationship between the ΔRIRF/Δl andrate flow rate at different water-prism Figure 4.4.Relationship between the ∆RIRF/∆l and flow at different water-prism temperaturetemperature differences.
differences.
3. Experiments 3. Experiments
3. Experiments 3.1. SPR Imaging System and Sensing Chip 3.1. SPR Imaging System and Sensing Chip The SPR experiments were performed using a home-built SPR differential interferometric 3.1. SPR Imaging System and Sensing Chip The SPR experiments performed using home-built SPR has differential interferometric imaging imaging system shown were in Figure 5A. This SPRa imaging system been discussed in a previous The SPR experiments were performed using a home-built SPR differential interferometric system shown in Figure 5A. This SPR imagingand system has beenbeam discussed in a previous publication publication [25]. Briefly, a p-polarized collimated of light was projected onto [25]. a imaging system shown in Figure 5A. This SPR imaging system has been discussed in a previous Briefly, a p-polarized and collimated beam of light was projected onto a Kretschmann prism-based Kretschmann prism-based sensing chip. The reflected light was split into two orthogonal-polarized publication Briefly, onto a p-polarized and beam ofoflight was projected onto sensing chip. [25]. The light wasCCD split cameras. into twocollimated orthogonal-polarized light and imaged ontoa light beams andreflected imaged two Two interferograms 180° beams phase-difference were ˝ Kretschmann prism-based sensing chip. The reflected light was split into two orthogonal-polarized twosequentially CCD cameras. Two interferograms of 180 phase-difference were sequentially obtained. obtained. light beams and imaged onto two CCD cameras. Two interferograms of 180° phase-difference were sequentially obtained.
Figure 5. Photographs of SPR the SPR imaging system, consisting theoptic SPR system optic system the Figure 5. Photographs of the imaging system, consisting of theofSPR (A), the(A), medium medium injection system (B) and the temperature controller (C). injection system (B) and the temperature controller (C).
Figure 5. Photographs of the SPR imaging system, consisting of the SPR optic system (A), the medium injection system (B) and the temperature controller (C).
Sensors 2016, 16, 964
5 of 10
A Poly (dimethylsiloxane) (PDMS) sheet with a channel of 0.5 ˆ 1 mm2 cross-section was pressed Sensors 2016, 16, x 5 of 10 and sealed with the sensing chip, forming the flow channel. The sensing chip was positioned into an 2 cross-section aluminumAholder, in which a PT100(PDMS) resistance thermometer detector and a polyimidewas heating Poly (dimethylsiloxane) sheet with a channel of 0.5(RTD) × 1 mm pressed and sealed withtemperature the sensing controller chip, forming the flow channel. The sensing film were embedded. A PID (TC200, Thorlabs, Shanghai, China)chip waswas used to positioned into an aluminum holder, in which a PT100 resistance thermometer detector (RTD) and a control the sensing chip temperature. polyimide heating film were embedded. A PID temperature controller (TC200, Thorlabs, Shanghai, China) was used to control the sensing chip temperature. 3.2. Experimental Procedures
A injection system was used to generate different flow rates of water, as shown in 3.2.self-built Experimental Procedures Figure 5B. The system consisted of a peristaltic pump and a pulsation dampener to acquire a stable A self-built injection system was used to generate different flow rates of water, as shown in flow. The flow rate was corrected by a commercial flow rate sensor (SLI-2000, Sensirion AG, Staefa, Figure 5B. The system consisted of a peristaltic pump and a pulsation dampener to acquire a stable Switzerland). Prior towas experiments, prism was preheated overnight achieveAG, theStaefa, expected flow. The flow rate corrected bythe a commercial flow rate sensor (SLI-2000,to Sensirion ˝ ˝ temperature (30, 35, or 40 C). Then deionized ~25 C) was injected into the channel Switzerland). Prior to experiments, the prismwater was (DI, preheated overnight to achieve the flow expected at 50 temperature µL/min. The for deionized 20 min towater achieve stable temperature distribution. Then the (30,injection 35, or 40 lasted °C). Then (DI,a~25 °C) was injected into the flow channel 50 μL/min. The injection lasted30 formin 20 min to achieve stabletemperature temperature distribution distribution. Then theprism. wateratflow was stopped for roughly to recover the ainitial for the water flow was stopped for roughly 30 min to recover the initial temperature distribution for the 800, The measurement of the water flow was repeated three times. Water flows of 100, 200, 400, 600, prism. The measurement of the water flow was repeated three times. Water flows of 100, 200, 400, 1000 µL/min was measured using the above procedure. 600, 800, 1000 μL/min was measured using the above procedure.
3.3. Data Processing
3.3. Data Processing
A time series of RIRF distribution images of the sensing surface was sequentially extracted from A time series of RIRF distribution images of the sensing surface was sequentially extracted from the two interferograms. The temperature distributions in the flow channel were included in the the two interferograms. The temperature distributions in the flow channel were included in the RIRF RIRF images. images.To Toquantify quantifythe theSPR SPRresponse response variation along flow streams, averaged variation along the the flow streams, the the RIRFRIRF was was averaged widthwise and then the average RIRF along the flow channel was plotted using a linear fit widthwise and then the average RIRF along the flow channel was plotted using a linear fit withwith the the distance from the inlet. The RIRF variation rate was used to quantify the flow rates. distance from the inlet. The RIRF variation rate was used to quantify the flow rates.
Figure 6. RIRF distributions across the sensing surface with the water temperature of 30 °C (A) Figure 6. RIRF distributions across the sensing surface with the water temperature of 30 ˝ Cand (A) and 40 °C (B). The RIRF variations in the flow channel (C). 40 ˝ C (B). The RIRF variations in the flow channel (C).
4. Results and Discussion
4. Results and Discussion 4.1. SPR Response Distribution across the Sensing Surface
4.1. SPR Response Distribution across the Sensing Surface
It is necessary to test the SPR response distribution across the sensing surface before measuring
It necessaryFigure to test the SPR distribution the sensing surface measuring itsisvariations. 6 shows theresponse RIRF distribution acrossacross the sensing surface when thebefore water was 30 its variations. Figure 6 shows the RIRF distribution across the sensing surface when the water was
Sensors 2016, 16, 964 Sensors 2016, 16, x
6 of 10 6 of 10
30 and 40 ˝ C, respectively. It is obvious that the SPR response was uniform except for some speckles in and channel. 40 °C, respectively. It is the obvious the SPR responsealong was uniform except for some in was the flow To quantify SPR that response variation the flow channel, thespeckles response the flow channel. To quantify the SPR response variation along the flow channel, the response was averaged widthwise and the average response along the flow stream was shown in Figure 6C. It can averaged widthwise and thewas average response along the flow stream wasdisturbance shown in Figure 6C. Itspeckles. can be seen that the SPR response relatively unchanged except for the of some be seen that the SPR response was relatively unchanged except for the disturbance of some speckles. This verified that SPR imaging was able to measure the flow-induced SPR response variations. This verified that SPR imaging was able to measure the flow-induced SPR response variations.
4.2. SPR Response Variations with Temperature
4.2. SPR Response Variations with Temperature
The refractive index (RI)(RI) is sensitive to temperature. on the The refractive index is sensitive to temperature.ToToverify verifythe theinfluence influence of of temperature temperature on SPR response, the flowthe channel was filled water of a temperature varying from 3030toto40 ˝ C the SPR response, flow channel was with filled DI with DI water of a temperature varying from and each change lasted 15lasted min. Figure presents the SPR of water at different 40 °Ctemperature and each temperature change 15 min. 7A Figure 7A presents theresponse SPR response of water at temperatures. It shows thatItthe RIRF was to the water consistent different temperatures. shows that theinversely RIRF was proportional inversely proportional to thetemperature, water temperature, consistent the works of Figure Moreira7B [26]. Figure shows the fitted linear relationship between with the works with of Moreira [26]. shows the7B fitted linear relationship between the RIRF the and the RIRF and the water temperature. It can be fitted linearly to ΔRIRF = 0.0041 × ΔT + 0.34, with the water temperature. It can be fitted linearly to ∆RIRF = 0.0041 ˆ ∆T + 0.34, with the standard deviation standard deviation R2 = 0.993. R2 = 0.993.
Figure 7. The variations versus timeatatdifferent different temperatures and thethe relationship between Figure 7. The RIRFRIRF variations versus time temperatures(A) (A) and relationship between RIRF and temperature (B). RIRF and temperature (B).
4.3. Flow-Induced SPR Response Variations
4.3. Flow-Induced SPR Response Variations
The above mentioned experiments have proven that the SPR response has a linear relationship
The mentioned experiments have proven that the SPR response hasflow a linear relationship withabove the water temperature. The flow-induced temperature variations along the channel were thenwater measured. Water wasThe injected into the flow channel using flow rates of 50,the 100, 200,channel 400, 600,were with the temperature. flow-induced temperature variations along flow 800, 1000 μL/min inducedinto temperature were detected sequentially. then measured. Waterand wasthe injected the flow distributions channel using flow rates of 50, 100, 200, 400, 600, 800, Figure 8 shows the RIRF distribution (inset) of a 50 μL/min water flow rate and the subsequent 1000 µL/min and the induced temperature distributions were detected sequentially. signal processing. It can be seen from the inset that the RIRF decreased slightly along the flow stream. Figure 8 shows the RIRF distribution (inset) of a 50 µL/min water flow rate and the subsequent This occurred because the water flow was heated by the prism and thus the temperature increased signal processing. It can be seen from the inset that the RIRF decreased slightly along the flow stream. gradually at it moved downstream. Due to the negative linear relationship between the temperature This occurred because the water flow was heated by the prism and thus the temperature increased and RIRF, the RIRF decreased along the streams. The widthwise-averaged RIRF variations along the gradually at it moved Dueato the negative linear between the temperature flow streams were downstream. fitted linearly with decreasing rate of nearlyrelationship −7.9 × 10−4 RIRF/mm. and RIRF,The the RIRF RIRFdistributions decreased along the streams. The RIRF variations and processed results ofwidthwise-averaged other flow rates are presented in Figurealong 9. It the 4 RIRF/mm. flow streams were fitted linearly with in a decreasing nearly ´7.9 ˆ 10´rate. shows that higher flow rates result larger RIRF rate and of larger RIRF variation Lower flow rates prolong heating timeand of the water flow, resulting in a flow higher temperature at the in monitoring The RIRFthe distributions processed results of other rates are presented Figure 9.area. It shows As the RIRF is negatively proportional to temperature, RIRF increased with the flow rate. In the that higher flow rates result in larger RIRF and larger RIRF variation rate. Lower flow rates prolong addition, it is possible that the temperature of slower water flow has reached approximately the heating time of the water flow, resulting in a higher temperature at the monitoring area. As the RIRF is equilibrium temperature, while the water of higher flow was still heated and thus generated high negatively proportional to temperature, RIRF increased with the flow rate. In addition, it is possible RIRF variation rate in the SPR monitoring area. This was consistent with the previous COMSOL that the temperature of slower water flow has reached approximately the equilibrium temperature, simulations. while the water of higher flow was still heated and thus generated high RIRF variation rate in the SPR monitoring area. This was consistent with the previous COMSOL simulations.
Sensors 2016, 16, 964 Sensors 2016, 16, x
7 of 10
Sensors 2016, 16, x
7 of 10
7 of 10
Figure 8. The RIRF distribution inflow the channel flow channel the widthwise-averaged RIRF Figure 8. The RIRF distribution in the (inset)(inset) and theand widthwise-averaged RIRF variation variation along streams at 50 μL/min. Figure 8. The RIRF distribution in the flow channel (inset) and the widthwise-averaged RIRF along streams at 50 µL/min. variation along streams at 50 μL/min.
Figure 9. The RIRF distributions(A) (A)and andthe thelinear linear fittings fittings of flow rates. Figure 9. The RIRF distributions of ΔRIRF/Δl ∆RIRF/∆l(B) (B)atatdifferent different flow rates. Figure 9. The RIRF distributions (A) and the linear fittings of ΔRIRF/Δl (B) at different flow rates.
Figure 10 presents the extracted RIRF variation rate, ΔRIRF/Δl, at different flow rates. It shows Figure 10 presents the extracted RIRF variation rate, ∆RIRF/∆l, at different flow rates. It shows that Figure the variation rate increased approximately linearly with the logarithm flowrates. rate It over the 10 presents the extracted RIRF variation rate, ΔRIRF/Δl, at differentofflow shows that thethe variation increased linearly with the logarithm flow rate overover the range that rate increased approximately linearly with the of flow the range of variation 50–800 rate μL/min. Three approximately experiments were performed and thelogarithm resultsoffollowed arate similar trend of range 50–800 Three experiments were performed and the followed a similar trend with ofµL/min. 50–800 μL/min. Three experiments wereThis performed andresults thethe results followed a similar trend with less than 11% relative standard deviation. demonstrates proof of concept of flow rate less thanless 11% relative standard deviation. This demonstrates the proof concept of flow of rate detection with than 11% relative standard deviation. This demonstrates theofproof of concept flow rate detection based on SPR imaging. based on SPR imaging. detection based on SPR imaging.
Figure 10. SPR response variation (ΔRIRF/Δl) vs. flow rate (Q). Figure 10.SPR SPRresponse responsevariation variation (∆RIRF/∆l) (ΔRIRF/Δl) vs. (Q). Figure 10. vs.flow flowrate rate (Q).
Sensors 2016, 16, x Sensors 2016, 16, 964
8 of 10 8 of 10
For thermal flow rate sensors, the detection range and sensitivity are mainly determined by the layout ofthermal the heating and the temperature measurement elements. Increasing the distance For flow element rate sensors, detection range and sensitivity are mainly determined by the between temperature elements expands the detection range, but also layout ofthe theheating heatingand element and the measurement temperature measurement elements. Increasing the distance reduces sensitivity. most existing flow sensors can only measure the temperature of between the the heating andWhile temperature measurement elements expands the detection range, but also several discrete points, SPR imaging continuous space-resolved temperature reduces the sensitivity. While most existingcan flowmeasure sensors can only measure the temperature of several variations. Therefore, it is can possible for continuous a SPR-based flow sensortemperature to achieve variations. an extremely wide discrete points, SPR imaging measure space-resolved Therefore, detection range the same device different temperature sensing regions. Here it it is possible for awith SPR-based flow sensorby to selecting achieve an extremely wide detection range with the same should be mentioned that the temperature lowest limit of 50 μL/min was Here chosen due to the performance of the device by selecting different sensing regions. it should be mentioned that pumping system, and not due limitations sensor. lowest limit of 50 µL/min was to chosen due to of thethe performance of the pumping system, and not due to limitations of the sensor. 4.4 Influence of the Prism-Liquid Temperature Difference 4.4. Influence of the Prism-Liquid Temperature Difference To research the influence of the prism-liquid temperature difference on the measurement, research influence the prism-liquid temperature difference the measurement, water waterToflow rates the with differentofprism-liquid temperature difference wereon detected. Figure 11 shows flowrelationship rates with different prism-liquid temperature were and detected. Figure 11 the the between the SPR response variationdifference rate ΔRIRF/Δl the logarithm of shows flow rate, relationship between the SPR response variation rate ∆RIRF/∆l and the logarithm of flow rate, where where the water was 25 °C and the prism was 30, 35 and 40 °C, respectively. It can be seen that the water was 25 ˝ Cdifferences and the prism 30, 35detection and 40 ˝ C, respectively. It can be seen larger larger temperature lead was to higher sensitivity, consistent with thethat previous temperature differences lead to higher detection sensitivity, consistent with the previous theoretical theoretical analysis. Therefore, a larger temperature difference is desirable to enhance the analysis. Therefore, larger difference is desirable to enhance theas sensitivity. sensitivity. Howevera in the temperature case of temperature sensitive experiments, such biologicalHowever studies, in the case of temperature sensitive experiments, such as studies, temperature suitable temperature differences are necessary to balance thebiological sensitivity and thesuitable object under test. differences are necessary to balance the sensitivity and the object under test. Compared with the conventional thermal flow sensor, which is of high integration based on Compared with thethe conventional flowissensor, which is due of high integration based for on the MEMS technology, SPR-based thermal flow sensor more complex to its optical system the MEMS technology, the SPR-based flowpaper, sensorthe is SPR moresystem complex to itsoptical opticalelements system for space-resolved temperature sensing. In this useddue several space-resolved temperature sensing.its Inperformance, this paper, thefurther SPR system usedthe several opticalbarrier elements for interferometric imaging to improve lowering integration of the interferometric imaging tocomplex improveinterferometric its performance,imaging further lowering the integration barrier the flow flow sensor. However, is unnecessary for flow rate of detection, sensor. However, complex interferometric imaging is unnecessary for flow rate detection, where SPR where SPR intensity imaging which uses less optical elements, is sufficient. Based on the proof of intensityofimaging which uses optical sufficient. Based onpaper, the proof of concept the concept the SPR-based flowless sensor thatelements, has been is demonstrated in this a new compactofflow SPR-based flow that has imaging been demonstrated in this paper, compact flow sensor on sensor based onsensor SPR intensity has been designed and aisnew under development. Thisbased should SPR intensity imaging has been designed and is under development. This should allow for integration allow for integration of the SPR-based flow sensor that will be greatly improved. As SPR has been of the SPR-based sensor that will bethis greatly improved. been widely applied in widely applied inflow bio-chemical analysis, method should As alsoSPR be has easily integrated into these bio-chemical analysis, this method should also be easily integrated into these analytical systems. analytical systems.
Figure vs.vs. flow rate (Q)(Q) with thethe prism temperature of 30, 35 Figure 11. 11. SPR SPRresponse responsevariation variation(ΔRIRF/Δl) (∆RIRF/∆l) flow rate with prism temperature of 30, and 40 °C. 35 and 40 ˝ C.
5. Conclusions 5. Conclusions In this paper, a microfluidic flow rate detection method based on SPR temperature imaging is In this paper, a microfluidic flow rate detection method based on SPR temperature imaging presented. The proof of concept has been demonstrated through theoretical simulations and is presented. The proof of concept has been demonstrated through theoretical simulations and experiments. The novel calorimetric flow rate measurement method allows reliable monitoring of experiments. The novel calorimetric flow rate measurement method allows reliable monitoring of liquid flow rates in a wide range of tens to hundreds of μL/min. As SPR has been commonly used for
Sensors 2016, 16, 964
9 of 10
liquid flow rates in a wide range of tens to hundreds of µL/min. As SPR has been commonly used for bio-chemical analysis, this novel SPR-based flow rate detection method should also be easily integrated into these systems. Other than water-prism temperature difference, some factors such as liquids of different heat capacities and the flow channel dimensions need to be further researched. Furthermore, the performance of the flow sensor needs to be enhanced, including improving its integration and expanding the detection range for biological applications. Acknowledgments: This work was supported by Tsinghua University Initiative Scientific Research Program (20131089190) and the Fund for Joint Project of Beijing. Author Contributions: P. Wang and X. Yu conceived and designed the experiments; S. Liu and T. Zhao performed the theoretical simulations and analysis; S. Deng and M. Guo performed the experiments; S. Deng and S. Xu analyzed the data; S. Deng and S. Liu wrote the paper. Conflicts of Interest: The authors declare no conflict of interest. None of the material in the paper has been published or is under consideration for publication elsewhere.
Abbreviations The following abbreviations are used in this manuscript: SPR RI RIRF CCD PMMA
surface plasmon resonance refractive index refractive index related factor charge coupled device Polymethylmethacrylate
References 1. 2. 3. 4. 5. 6. 7. 8.
9. 10.
11. 12. 13.
Li, Y.; Yan, G.; Zhang, L.; He, S. Microfluidic flowmeter based on micro “hot-wire” sandwiched Fabry-Perot interferometer. Opt. Express 2015, 23, 9483–9493. [CrossRef] [PubMed] Vilares, R.; Hunter, C.; Ugarte, I.; Aranburu, I.; Berganzo, J.; Elizalde, J.; Fernandez, L.J. Fabrication and testing of a SU-8 thermal flow sensor. Sens. Actuators B Chem. 2010, 147, 411–417. [CrossRef] Lien, V.; Vollmer, F. Microfluidic flow rate detection based on integrated optical fiber cantilever. Lab Chip 2007, 7, 1352–1356. [CrossRef] [PubMed] Sengupta, S.; Ziaie, B.; Barocas, V.H. Lag-after-pulsed-separation microfluidic flowmeter for biomacromolecular solutions. Sens. Actuators B Chem. 2004, 99, 25–29. [CrossRef] Collins, J.; Lee, A.P. Microfluidic flow transducer based on the measurement of electrical admittance. Lab Chip 2004, 4, 7–10. [CrossRef] [PubMed] Ahrens, R.; Schlote-Holubek, K. A micro flow sensor from a polymer for gases and liquids. J. Micromech. Microeng. 2009, 19, 74006. [CrossRef] Ernst, H.; Jachimowicz, A.; Urban, G.A. High resolution flow characterization in Bio-MEMS. Sens. Actuators A Phys. 2002, 100, 54–62. [CrossRef] Caldas, P.; Jorge, P.A.; Rego, G.; Frazão, O.; Santos, J.L.; Ferreira, L.A.; Araújo, F. Fiber optic hot-wire flowmeter based on a metallic coated hybrid long period grating/fiber Bragg grating structure. Appl. Opt. 2011, 50, 2738–2743. [CrossRef] [PubMed] Van Oudheusden, B.W. The determination of the effective ambient temperature for thermal flow sensors in a non-isothermal environment. Sens. Actuators A Phys. 1999, 72, 38–45. [CrossRef] Dijkstra, M.; de Boer, M.J.; Berenschot, J.W.; Lammerink, T.; Wiegerink, R.J.; Elwenspoek, M. Miniaturized thermal flow sensor with planar-integrated sensor structures on semicircular surface channels. Sens. Actuators A Phys. 2008, 143, 1–6. [CrossRef] Kim, T.H.; Kim, S.J. Development of a micro-thermal flow sensor with thin-film thermocouples. J. Micromech. Microeng. 2006, 16, 2502. [CrossRef] Berthet, H.; Jundt, J.; Durivault, J.; Mercier, B.; Angelescu, D. Time-of-flight thermal flowrate sensor for lab-on-chip applications. Lab Chip 2011, 11, 215–223. [CrossRef] [PubMed] Chung, J.; Grigoropoulos, C.P.; Greif, R. Infrared thermal velocimetry for nonintrusive flow measurement in silicon microfluidic devices. Rev. Sci. Instrum. 2003, 74, 2911–2917. [CrossRef]
Sensors 2016, 16, 964
14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
24. 25.
26.
10 of 10
Kuo, J.T.; Yu, L.; Meng, E. Micromachined thermal flow sensors—A review. Micromachines 2012, 3, 550–573. [CrossRef] Wu, S.; Lin, Q.; Yuen, Y.; Tai, Y. MEMS flow sensors for nano-fluidic applications. Sens. Actuators A Phys. 2001, 89, 152–158. [CrossRef] Yan, W.; Liu, C.; Li, J.; Ma, L.; Niu, D. Thermal distribution microfluidic sensor based on silicon. Sens. Actuators B Chem. 2005, 108, 943–946. Sabaté, N.; Santander, J.; Fonseca, L.; Gràcia, I.; Cané, C. Multi-range silicon micromachined flow sensor. Sens. Actuators A Phys. 2004, 110, 282–288. [CrossRef] Kim, S.; Nam, T.; Park, S. Measurement of flow direction and velocity using a micromachined flow sensor. Sens. Actuators A Phys. 2004, 114, 312–318. [CrossRef] Hoera, C.; Skadell, M.M.; Pfeiffer, S.A.; Pahl, M.; Shu, Z.; Beckert, E.; Belder, D. A chip-integrated highly variable thermal flow rate sensor. Sens. Actuators B Chem. 2016, 225, 42–49. [CrossRef] Wagner, C.E.; Macedo, L.J.; Opdahl, A. Temperature gradient approach for rapidly assessing sensor binding kinetics and thermodynamics. Anal. Chem. 2015, 87, 7825–7832. [CrossRef] [PubMed] Srivastava, S.K.; Gupta, B.D. Simulation of a localized surface-plasmon-resonance-based fiber optic temperature sensor. JOSA A 2010, 27, 1743–1749. [CrossRef] [PubMed] Turhan-Sayan, G. Temperature effects on surface plasmon resonance: Design considerations for an optical temperature sensor. J. Lightwave Technol. 2003, 21, 805. Chiang, H.; Yeh, H.; Chen, C.; Wu, J.; Su, S.; Chang, R.; Wu, Y.; Tsai, D.P.; Jen, S.U.; Leung, P.T. Surface plasmon resonance monitoring of temperature via phase measurement. Opt. Commun. 2004, 241, 409–418. [CrossRef] Ekgasit, S.; Thammacharoen, C.; Yu, F.; Knoll, W. Evanescent field in surface plasmon resonance and surface plasmon field-enhanced fluorescence spectroscopies. Anal. Chem. 2004, 76, 2210–2219. [CrossRef] [PubMed] Wang, D.; Ding, L.; Zhang, W.; Luo, Z.; Ou, H.; Zhang, E.; Yu, X. A high-throughput surface plasmon resonance biosensor based on differential interferometric imaging. Meas. Sci. Technol. 2012, 23, 65701. [CrossRef] Moreira, C.S.; Lima, A.; Neff, H.; Thirstrup, C. Temperature-dependent sensitivity of surface plasmon resonance sensors at the gold-water interface. Sens. Actuators B Chem. 2008, 134, 854–862. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
A Novel Microfluidic Flow Rate Detection Method Based on Surface Plasmon Resonance Temperature Imaging Shijie Deng, Peng Wang *, Shengnan Liu, Tianze Zhao, Shanzhi Xu, Mingjiang Guo and Xinglong Yu State Key Laboratory of Precision Measurement Technology and Instruments, Tsinghua University, Beijing 100084, China; [email protected] (S.D.); [email protected] (S.L.); [email protected] (T.Z.); [email protected] (S.X.); [email protected] (M.G.); [email protected] (X.Y.) * Correspondence: [email protected]; Tel.: +86-10-6277-2007 Academic Editors: Amine Miled and Jesse Greener Received: 29 April 2016; Accepted: 21 June 2016; Published: 24 June 2016
Abstract: A novel microfluidic flow rate detection method based on surface plasmon resonance (SPR) temperature imaging is proposed. The measurement is performed by space-resolved SPR imaging of the flow induced temperature variations. Theoretical simulations and analysis were performed to demonstrate a proof of concept using this approach. Experiments were implemented and results showed that water flow rates within a wide range of tens to hundreds of µL/min could be detected. The flow rate sensor is resistant to disturbances and can be easily integrated into microfluidic lab-on-chip systems. Keywords: microfluidic flow rate detection; surface plasmon resonance; temperature variation imaging
1. Introduction There is an increasing demand for microfluidic flow rate measurement technologies in various research fields, such as bio-chemical analysis, sample mixing, and environmental monitoring. Various microfluidic flow sensors of different working principles have been proposed in recent years [1–5]; the thermal flow rate sensors working on heat transfer property under the liquid flow, are the most prevalent [6,7]. According to the detection method, thermal flow rate sensors can be classified into three categories: hot element anemometers [8,9], calorimetric flow rate sensors [10,11] and time of flight sensors [12,13]. Calorimetric flow rate sensors which use thermistors that are laid either up- or down-stream of a heater to detect the flow induced temperature distribution, are of the highest sensitivity. The layout of the heater and thermistors is a key design factor that determines the measurement sensitivity and dynamic range [14–16]. Once the distance between the heater and thermistors is determined, the measurable flow is limited to a specific dynamic range inversely proportional to sensibility. To expand the dynamic range and enhance the sensitivity, improvements have been implemented to optimize the arrangement of the heater and thermistors. Several temperature detectors were utilized in the works by N. Sabaté [17] and Seunghyun Kim [18]. The introduction of an array of thermistors facilitated a large expansion of the dynamic range. With the use of fluorescent dyes, Hoera et al. [19] developed a flow sensor based on temperature distribution imaging, that had a wide dynamic range of 10 nL/min to 100 µL/min. Unfortunately, the use of fluorescent dyes impose restrictions on the flow rate sensor in non-labeled detection fields. Surface plasmon resonance (SPR) is a collective charge density oscillation that occurs at a metal-dielectric interface. As SPR is extremely sensitive to the temperature-dependent refractive Sensors 2016, 16, 964; doi:10.3390/s16070964
www.mdpi.com/journal/sensors
Sensors 2016, 16, x
2 of 10
Surface plasmon resonance (SPR) is a collective charge density oscillation that occurs at a 2 of 10 interface. As SPR is extremely sensitive to the temperature-dependent refractive index (RI) just above the metal film, it has extensive applications in temperature monitoring and sensing [20–22]. In work performed et applications al., a variation as small as 0.027 K and was index (RI) just above the metal film, it by has Chiang extensive in temperature monitoring distinguished at 632.8 nm [23]. Furthermore by means of SPR imaging, the refractive index sensing [20–22]. In work performed by Chiang et al., a variation as small as 0.027 K was distinguished distribution on the whole sensing surfaceofcan simultaneously monitored. When liquidon flows at 632.8 nm [23]. Furthermore by means SPRbe imaging, the refractive index distribution the through the sensing unit, the temperature above the sensing surface is redistributed. Because the whole sensing surface can be simultaneously monitored. When liquid flows through the sensing refractive index (RI) is inversely proportional to the temperature, the RI varies along the flow unit, the temperature above the sensing surface is redistributed. Because the refractive index (RI) is channel, proportional leading to ato corresponding variation of thealong SPRthe response. Therefore, the flow rate inversely the temperature, the RI varies flow channel, leading to liquid a corresponding can be resolved from the SPR response variations. variation of the SPR response. Therefore, the liquid flow rate can be resolved from the SPR In this paper, SPR imaging was proposed for the first time to measure the microfluidic flow response variations. rate. InAs the space-resolved able the flow-induced continuous this paper, SPR imagingSPR wasimaging proposedisfor the to firstmonitor time to measure the microfluidic flow temperature variations, rather than several discrete points, this method has more resistance to rate. As the space-resolved SPR imaging is able to monitor the flow-induced continuous temperature disturbances andthan potentially has a wide detection range. Asmore SPRresistance has beentowidely appliedand in variations, rather several discrete points, this method has disturbances bio-chemical analysis, the SPR-based flow rate measurement will be facilely integrated into these potentially has a wide detection range. As SPR has been widely applied in bio-chemical analysis, the analytical systems. SPR-based flow rate measurement will be facilely integrated into these analytical systems. Sensors 2016, 16, 964 metal-dielectric
2. Theory 2. Theoryand andSimulations Simulations 2.1. 2.1. Principle The The schematic schematic of of microfluidic microfluidic flow flow rate rate measurement measurement based based on on SPR SPR temperature temperature imaging imaging is is illustrated 1. The SPRSPR prism coated with awith gold afilm, together with a flow channel, constitutes illustratedininFigure Figure 1. The prism coated gold film, together with a flow channel, the sensing chip. The SPR chip. prism The and the are loaded with high are and low temperature, respectively. constitutes the sensing SPRwater prism and the water loaded with high and low A collimated beam of light projects onto the bottom of the gold film to the initiate SPR.of When the water is temperature, respectively. A collimated beam of light projects onto bottom the gold film to still, the SPR response (intensity of the reflected beam) is uniform. But when the water is flowing, there initiate SPR. When the water is still, the SPR response (intensity of the reflected beam) is uniform. will a temperature Sincegradient the RI isalong inversely proportional the But be when the water isgradient flowing,along therethe willflow be achannel. temperature the flow channel.to Since temperature, the RIproportional decreases along thetemperature, flow channel,the leading to a corresponding decrease of the SPR the RI is inversely to the RI decreases along the flow channel, leading response. The decreasing rateofvaries withresponse. the flow rate. As such, the liquid flow ratethe canflow be resolved to a corresponding decrease the SPR The decreasing rate varies with rate. As from of the such,the thevariation liquid flow rateSPR can response. be resolved from the variation of the SPR response.
Figure 1. 1. Schematic of the the surface surface plasmon plasmon resonance resonance (SPR)-based (SPR)-based microfluidic microfluidic flow flow rate ratedetection. detection. Figure Schematic of
2.2. Flow-Induced Flow-Induced Temperature TemperatureVariations Variations 2.2. The flow-thermal flow-thermal coupling coupling physical physical field field of of the the sensing sensing chip chip was was simulated simulated using using COMSOL. COMSOL. The Since SPR is only able to detect the changes above the upper surface of the prism, the flow channel Since SPR is only able to detect the changes above the upper surface of the prism, the flow channel was placed placed above above the the prism. prism. For by the the top top was For fluid fluid sealing, sealing, the the flow flow channel channel was was pressed pressed down down by Polymethylmethacrylate (PMMA) shell to seal with the bottom prism. Therefore, the flow channel Polymethylmethacrylate (PMMA) shell to seal with the bottom prism. Therefore, the flow channel was sandwiched sandwiched between between aa top top shell shell and and aa bottom bottom SPR SPR prism, prism, as as illustrated illustrated in in Figure Figure 2A. 2A. In In the the was case where the water and prism were at 25 and 35 °C, respectively, the flow (100 μL/min) induced ˝ case where the water and prism were at 25 and 35 C, respectively, the flow (100 µL/min) induced temperaturedistribution distributionof of sensing is shown in 2A. Figure Due to the temperature temperature thethe sensing chipchip is shown in Figure Due2A. to the temperature difference, difference, heat exchange occurred between the liquid and prism. Figure 2B shows the temperature heat exchange occurred between the liquid and prism. Figure 2B shows the temperature variations at different flow rates of 100, 200, 400, 600, 800 and 1000 µL/min. Due to the increase of heating time,
Sensors 2016, 16, x Sensors 2016, 16, x
3 of 10 3 of 10
Sensors 2016, at 16, different 964 3 of of 10 variations flow rates of 100, 200, 400, 600, 800 and 1000 μL/min. Due to the increase variations at different flow rates of 100, 200, 400, 600, 800 and 1000 μL/min. Due to the increase of heating time, the liquid temperature increases approximately linearly along streams, except at the heating time, the liquid temperature increases approximately linearly along streams, except at the inlet and outlet where the flow was unstable. Figure 2C further shows that at slower flow rates, the the liquid temperature increases approximately linearly along streams, the inlet and outlet inlet and outlet where the flow was unstable. further shows except that increase. at at slower flow rates, the liquid maintained a higher temperature and aFigure slower2Crate of temperature Therefore, the where the flow was unstable. Figure 2C further shows that at slower flow rates, the liquid maintained liquid maintained a higher temperature and a slower rate of temperature increase. Therefore, the section of the flow channel at 30–36 mm (between the two dashed lines) was chosen for flow rate asection higheroftemperature and a slower of(between temperature increase. Therefore, thechosen sectionfor of flow the flow the flow channel at 30–36rate mm the two dashed lines) was rate measurement. channel at 30–36 mm (between the two dashed lines) was chosen for flow rate measurement. measurement.
Figure 2. (A) Heat map of the SPR sensing chip. (B,C) the water temperature variations in the flow Figure 2.2. (A) Heat map map of the SPR SPR sensing sensing chip. chip. (B,C) (B,C) the the water water temperature temperature variations variations in in the the flow flow channel at(A) different flowof rates. Figure Heat the channel at different flow rates. channel at different flow rates.
2.3. Temperature-Induced SPR Response Variations 2.3. Temperature-Induced SPR Response Variations 2.3. Temperature-Induced SPR Response Due to the dependence of theVariations refractive index on the temperature, the flow-induced Due to the dependence ofvariation the refractive index on the Based temperature, the flow-induced temperature gradient generates a of theon SPR the multi-layer model, Due to the dependence of the refractive index theresponse. temperature, theon flow-induced temperature temperature gradient generates a variation of the SPR response. Based on the multi-layer model, Fresnel were usedoftothe analyze the SPR response. Figure depicts themodel, corresponding gradientequations generates[24] a variation SPR response. Based on the 3multi-layer Fresnel Fresnel equations [24] were used to analyze the SPR response. Figure 3 depicts the corresponding variations of the index the related (RIRF). The RIRF variation shows an inverse equations [24] wererefractive used to analyze SPR factor response. Figure 3 depicts the corresponding variations variations ofwith thethe refractive index related factor As (RIRF). The varies RIRF variation showslinearly an inverse relationship liquid temperature variation. the RIRF approximately and of the refractive index related factor (RIRF). The RIRF variation shows an inverse relationship with relationship with the liquid temperature variation. As the RIRF varies approximately linearly and its rate ΔRIRF/Δl changesAs at the different the ΔRIRF/Δl can beand used quantifyrate the thevariation liquid temperature variation. RIRF flow variesrates, approximately linearly itstovariation its variation rate ΔRIRF/Δl changes at different flow rates, the ΔRIRF/Δl can be used to quantify the flow rate. changes at different flow rates, the ∆RIRF/∆l can be used to quantify the flow rate. ∆RIRF/∆l flow rate.
Figure Figure 3. 3. SPR SPR variations variations in in the the flow flow channel channel at at different different flow flow rates. rates. Figure 3. SPR variations in the flow channel at different flow rates.
Sensors 2016, 16, 964 Sensors 2016, 16,16, x x Sensors 2016,
4 of 10
of 10 4 of410
2.4.2.4. Effect of of thethe Liquid-Prism Temperature Difference 2.4. Effect of the Liquid-Prism Effect Liquid-PrismTemperature TemperatureDifference Difference The temperature difference between the liquid liquid and is prism is a a key key factor affecting The temperature difference between and prism factor affecting thethe The temperature difference between the liquid and prism a keyisfactor affecting the prism-liquid prism-liquid heatexchange, exchange, leading to different different temperature variation ToToresearch thethe effect prism-liquid leading to temperature variation rates. research effect heat exchange,heat leading to different temperature variation rates. To researchrates. the effect of the liquid-prism ˝ of the liquid-prism temperature on the measurement, the water temperature was fixed at 25 °C °C of the liquid-prism temperature on the measurement, the water temperature was fixed at 25 temperature on the measurement, the water temperature was fixed at 25 C while the prism ˝ while the prism temperature was changed to 30, 35 and 40 °C, respectively. The relationship while the prism temperature changed 35 and 40relationship °C, respectively. The temperature was changed to 30,was 35 and 40 C, respectively. The between therelationship ∆RIRF/∆l between theΔRIRF/Δl ΔRIRF/Δl and the logarithm of flow 4. rate is ininFigure 4.4. ItItcan that thethe between the and the flow rate is shown shown Figure canbebeseen seen that and the logarithm of flow rate islogarithm shown in of Figure It can be seen that the detection range is about detection range is about 100–500 μL/min, with the linear range also being approximately detection range is about 100–500 μL/min, with the linear range also being approximately 100–500 µL/min, with the linear range also being approximately 100–500 µL/min. When the fluid is 100–500 μL/min. Whenthe the fluid isis faster faster than 500 μL/min, the variations in in theare 100–500 μL/min. When the fluid than 500 μL/min, the temperature temperature variations the faster than 500 µL/min, temperature variations in the monitoring section of the flow channel monitoring section of the flow channel are hardly distinguishable. Therefore the upper limit of the monitoring section of the flow channel are limit hardly distinguishable. the upper limitlimit of the hardly distinguishable. Therefore the upper of the linear range isTherefore 500 µL/min. The lower of linear range is 500 μL/min. The lower limit of 100 μL/min was chosen due to the performance of the linear range was is 500 μL/min. lower limit of 100ofμL/min was chosen dueand to the of the 100 µL/min chosen dueThe to the performance the pumping system, notperformance due to limitations pumping system, and not due to limitations of the sensor. With larger liquid-prism temperature pumping system, notliquid-prism due to limitations of thedifferences, sensor. With liquid-prism temperature of the sensor. Withand larger temperature thelarger sensitivity of the measurement differences, the sensitivity of the measurement was enhanced. differences, the sensitivity of the measurement was enhanced. was enhanced.
Figure 4. Relationship between the ΔRIRF/Δl and flow rate at different water-prism temperature differences. Figure Relationship between the ΔRIRF/Δl andrate flow rate at different water-prism Figure 4.4.Relationship between the ∆RIRF/∆l and flow at different water-prism temperaturetemperature differences.
differences.
3. Experiments 3. Experiments
3. Experiments 3.1. SPR Imaging System and Sensing Chip 3.1. SPR Imaging System and Sensing Chip The SPR experiments were performed using a home-built SPR differential interferometric 3.1. SPR Imaging System and Sensing Chip The SPR experiments performed using home-built SPR has differential interferometric imaging imaging system shown were in Figure 5A. This SPRa imaging system been discussed in a previous The SPR experiments were performed using a home-built SPR differential interferometric system shown in Figure 5A. This SPR imagingand system has beenbeam discussed in a previous publication publication [25]. Briefly, a p-polarized collimated of light was projected onto [25]. a imaging system shown in Figure 5A. This SPR imaging system has been discussed in a previous Briefly, a p-polarized and collimated beam of light was projected onto a Kretschmann prism-based Kretschmann prism-based sensing chip. The reflected light was split into two orthogonal-polarized publication Briefly, onto a p-polarized and beam ofoflight was projected onto sensing chip. [25]. The light wasCCD split cameras. into twocollimated orthogonal-polarized light and imaged ontoa light beams andreflected imaged two Two interferograms 180° beams phase-difference were ˝ Kretschmann prism-based sensing chip. The reflected light was split into two orthogonal-polarized twosequentially CCD cameras. Two interferograms of 180 phase-difference were sequentially obtained. obtained. light beams and imaged onto two CCD cameras. Two interferograms of 180° phase-difference were sequentially obtained.
Figure 5. Photographs of SPR the SPR imaging system, consisting theoptic SPR system optic system the Figure 5. Photographs of the imaging system, consisting of theofSPR (A), the(A), medium medium injection system (B) and the temperature controller (C). injection system (B) and the temperature controller (C).
Figure 5. Photographs of the SPR imaging system, consisting of the SPR optic system (A), the medium injection system (B) and the temperature controller (C).
Sensors 2016, 16, 964
5 of 10
A Poly (dimethylsiloxane) (PDMS) sheet with a channel of 0.5 ˆ 1 mm2 cross-section was pressed Sensors 2016, 16, x 5 of 10 and sealed with the sensing chip, forming the flow channel. The sensing chip was positioned into an 2 cross-section aluminumAholder, in which a PT100(PDMS) resistance thermometer detector and a polyimidewas heating Poly (dimethylsiloxane) sheet with a channel of 0.5(RTD) × 1 mm pressed and sealed withtemperature the sensing controller chip, forming the flow channel. The sensing film were embedded. A PID (TC200, Thorlabs, Shanghai, China)chip waswas used to positioned into an aluminum holder, in which a PT100 resistance thermometer detector (RTD) and a control the sensing chip temperature. polyimide heating film were embedded. A PID temperature controller (TC200, Thorlabs, Shanghai, China) was used to control the sensing chip temperature. 3.2. Experimental Procedures
A injection system was used to generate different flow rates of water, as shown in 3.2.self-built Experimental Procedures Figure 5B. The system consisted of a peristaltic pump and a pulsation dampener to acquire a stable A self-built injection system was used to generate different flow rates of water, as shown in flow. The flow rate was corrected by a commercial flow rate sensor (SLI-2000, Sensirion AG, Staefa, Figure 5B. The system consisted of a peristaltic pump and a pulsation dampener to acquire a stable Switzerland). Prior towas experiments, prism was preheated overnight achieveAG, theStaefa, expected flow. The flow rate corrected bythe a commercial flow rate sensor (SLI-2000,to Sensirion ˝ ˝ temperature (30, 35, or 40 C). Then deionized ~25 C) was injected into the channel Switzerland). Prior to experiments, the prismwater was (DI, preheated overnight to achieve the flow expected at 50 temperature µL/min. The for deionized 20 min towater achieve stable temperature distribution. Then the (30,injection 35, or 40 lasted °C). Then (DI,a~25 °C) was injected into the flow channel 50 μL/min. The injection lasted30 formin 20 min to achieve stabletemperature temperature distribution distribution. Then theprism. wateratflow was stopped for roughly to recover the ainitial for the water flow was stopped for roughly 30 min to recover the initial temperature distribution for the 800, The measurement of the water flow was repeated three times. Water flows of 100, 200, 400, 600, prism. The measurement of the water flow was repeated three times. Water flows of 100, 200, 400, 1000 µL/min was measured using the above procedure. 600, 800, 1000 μL/min was measured using the above procedure.
3.3. Data Processing
3.3. Data Processing
A time series of RIRF distribution images of the sensing surface was sequentially extracted from A time series of RIRF distribution images of the sensing surface was sequentially extracted from the two interferograms. The temperature distributions in the flow channel were included in the the two interferograms. The temperature distributions in the flow channel were included in the RIRF RIRF images. images.To Toquantify quantifythe theSPR SPRresponse response variation along flow streams, averaged variation along the the flow streams, the the RIRFRIRF was was averaged widthwise and then the average RIRF along the flow channel was plotted using a linear fit widthwise and then the average RIRF along the flow channel was plotted using a linear fit withwith the the distance from the inlet. The RIRF variation rate was used to quantify the flow rates. distance from the inlet. The RIRF variation rate was used to quantify the flow rates.
Figure 6. RIRF distributions across the sensing surface with the water temperature of 30 °C (A) Figure 6. RIRF distributions across the sensing surface with the water temperature of 30 ˝ Cand (A) and 40 °C (B). The RIRF variations in the flow channel (C). 40 ˝ C (B). The RIRF variations in the flow channel (C).
4. Results and Discussion
4. Results and Discussion 4.1. SPR Response Distribution across the Sensing Surface
4.1. SPR Response Distribution across the Sensing Surface
It is necessary to test the SPR response distribution across the sensing surface before measuring
It necessaryFigure to test the SPR distribution the sensing surface measuring itsisvariations. 6 shows theresponse RIRF distribution acrossacross the sensing surface when thebefore water was 30 its variations. Figure 6 shows the RIRF distribution across the sensing surface when the water was
Sensors 2016, 16, 964 Sensors 2016, 16, x
6 of 10 6 of 10
30 and 40 ˝ C, respectively. It is obvious that the SPR response was uniform except for some speckles in and channel. 40 °C, respectively. It is the obvious the SPR responsealong was uniform except for some in was the flow To quantify SPR that response variation the flow channel, thespeckles response the flow channel. To quantify the SPR response variation along the flow channel, the response was averaged widthwise and the average response along the flow stream was shown in Figure 6C. It can averaged widthwise and thewas average response along the flow stream wasdisturbance shown in Figure 6C. Itspeckles. can be seen that the SPR response relatively unchanged except for the of some be seen that the SPR response was relatively unchanged except for the disturbance of some speckles. This verified that SPR imaging was able to measure the flow-induced SPR response variations. This verified that SPR imaging was able to measure the flow-induced SPR response variations.
4.2. SPR Response Variations with Temperature
4.2. SPR Response Variations with Temperature
The refractive index (RI)(RI) is sensitive to temperature. on the The refractive index is sensitive to temperature.ToToverify verifythe theinfluence influence of of temperature temperature on SPR response, the flowthe channel was filled water of a temperature varying from 3030toto40 ˝ C the SPR response, flow channel was with filled DI with DI water of a temperature varying from and each change lasted 15lasted min. Figure presents the SPR of water at different 40 °Ctemperature and each temperature change 15 min. 7A Figure 7A presents theresponse SPR response of water at temperatures. It shows thatItthe RIRF was to the water consistent different temperatures. shows that theinversely RIRF was proportional inversely proportional to thetemperature, water temperature, consistent the works of Figure Moreira7B [26]. Figure shows the fitted linear relationship between with the works with of Moreira [26]. shows the7B fitted linear relationship between the RIRF the and the RIRF and the water temperature. It can be fitted linearly to ΔRIRF = 0.0041 × ΔT + 0.34, with the water temperature. It can be fitted linearly to ∆RIRF = 0.0041 ˆ ∆T + 0.34, with the standard deviation standard deviation R2 = 0.993. R2 = 0.993.
Figure 7. The variations versus timeatatdifferent different temperatures and thethe relationship between Figure 7. The RIRFRIRF variations versus time temperatures(A) (A) and relationship between RIRF and temperature (B). RIRF and temperature (B).
4.3. Flow-Induced SPR Response Variations
4.3. Flow-Induced SPR Response Variations
The above mentioned experiments have proven that the SPR response has a linear relationship
The mentioned experiments have proven that the SPR response hasflow a linear relationship withabove the water temperature. The flow-induced temperature variations along the channel were thenwater measured. Water wasThe injected into the flow channel using flow rates of 50,the 100, 200,channel 400, 600,were with the temperature. flow-induced temperature variations along flow 800, 1000 μL/min inducedinto temperature were detected sequentially. then measured. Waterand wasthe injected the flow distributions channel using flow rates of 50, 100, 200, 400, 600, 800, Figure 8 shows the RIRF distribution (inset) of a 50 μL/min water flow rate and the subsequent 1000 µL/min and the induced temperature distributions were detected sequentially. signal processing. It can be seen from the inset that the RIRF decreased slightly along the flow stream. Figure 8 shows the RIRF distribution (inset) of a 50 µL/min water flow rate and the subsequent This occurred because the water flow was heated by the prism and thus the temperature increased signal processing. It can be seen from the inset that the RIRF decreased slightly along the flow stream. gradually at it moved downstream. Due to the negative linear relationship between the temperature This occurred because the water flow was heated by the prism and thus the temperature increased and RIRF, the RIRF decreased along the streams. The widthwise-averaged RIRF variations along the gradually at it moved Dueato the negative linear between the temperature flow streams were downstream. fitted linearly with decreasing rate of nearlyrelationship −7.9 × 10−4 RIRF/mm. and RIRF,The the RIRF RIRFdistributions decreased along the streams. The RIRF variations and processed results ofwidthwise-averaged other flow rates are presented in Figurealong 9. It the 4 RIRF/mm. flow streams were fitted linearly with in a decreasing nearly ´7.9 ˆ 10´rate. shows that higher flow rates result larger RIRF rate and of larger RIRF variation Lower flow rates prolong heating timeand of the water flow, resulting in a flow higher temperature at the in monitoring The RIRFthe distributions processed results of other rates are presented Figure 9.area. It shows As the RIRF is negatively proportional to temperature, RIRF increased with the flow rate. In the that higher flow rates result in larger RIRF and larger RIRF variation rate. Lower flow rates prolong addition, it is possible that the temperature of slower water flow has reached approximately the heating time of the water flow, resulting in a higher temperature at the monitoring area. As the RIRF is equilibrium temperature, while the water of higher flow was still heated and thus generated high negatively proportional to temperature, RIRF increased with the flow rate. In addition, it is possible RIRF variation rate in the SPR monitoring area. This was consistent with the previous COMSOL that the temperature of slower water flow has reached approximately the equilibrium temperature, simulations. while the water of higher flow was still heated and thus generated high RIRF variation rate in the SPR monitoring area. This was consistent with the previous COMSOL simulations.
Sensors 2016, 16, 964 Sensors 2016, 16, x
7 of 10
Sensors 2016, 16, x
7 of 10
7 of 10
Figure 8. The RIRF distribution inflow the channel flow channel the widthwise-averaged RIRF Figure 8. The RIRF distribution in the (inset)(inset) and theand widthwise-averaged RIRF variation variation along streams at 50 μL/min. Figure 8. The RIRF distribution in the flow channel (inset) and the widthwise-averaged RIRF along streams at 50 µL/min. variation along streams at 50 μL/min.
Figure 9. The RIRF distributions(A) (A)and andthe thelinear linear fittings fittings of flow rates. Figure 9. The RIRF distributions of ΔRIRF/Δl ∆RIRF/∆l(B) (B)atatdifferent different flow rates. Figure 9. The RIRF distributions (A) and the linear fittings of ΔRIRF/Δl (B) at different flow rates.
Figure 10 presents the extracted RIRF variation rate, ΔRIRF/Δl, at different flow rates. It shows Figure 10 presents the extracted RIRF variation rate, ∆RIRF/∆l, at different flow rates. It shows that Figure the variation rate increased approximately linearly with the logarithm flowrates. rate It over the 10 presents the extracted RIRF variation rate, ΔRIRF/Δl, at differentofflow shows that thethe variation increased linearly with the logarithm flow rate overover the range that rate increased approximately linearly with the of flow the range of variation 50–800 rate μL/min. Three approximately experiments were performed and thelogarithm resultsoffollowed arate similar trend of range 50–800 Three experiments were performed and the followed a similar trend with ofµL/min. 50–800 μL/min. Three experiments wereThis performed andresults thethe results followed a similar trend with less than 11% relative standard deviation. demonstrates proof of concept of flow rate less thanless 11% relative standard deviation. This demonstrates the proof concept of flow of rate detection with than 11% relative standard deviation. This demonstrates theofproof of concept flow rate detection based on SPR imaging. based on SPR imaging. detection based on SPR imaging.
Figure 10. SPR response variation (ΔRIRF/Δl) vs. flow rate (Q). Figure 10.SPR SPRresponse responsevariation variation (∆RIRF/∆l) (ΔRIRF/Δl) vs. (Q). Figure 10. vs.flow flowrate rate (Q).
Sensors 2016, 16, x Sensors 2016, 16, 964
8 of 10 8 of 10
For thermal flow rate sensors, the detection range and sensitivity are mainly determined by the layout ofthermal the heating and the temperature measurement elements. Increasing the distance For flow element rate sensors, detection range and sensitivity are mainly determined by the between temperature elements expands the detection range, but also layout ofthe theheating heatingand element and the measurement temperature measurement elements. Increasing the distance reduces sensitivity. most existing flow sensors can only measure the temperature of between the the heating andWhile temperature measurement elements expands the detection range, but also several discrete points, SPR imaging continuous space-resolved temperature reduces the sensitivity. While most existingcan flowmeasure sensors can only measure the temperature of several variations. Therefore, it is can possible for continuous a SPR-based flow sensortemperature to achieve variations. an extremely wide discrete points, SPR imaging measure space-resolved Therefore, detection range the same device different temperature sensing regions. Here it it is possible for awith SPR-based flow sensorby to selecting achieve an extremely wide detection range with the same should be mentioned that the temperature lowest limit of 50 μL/min was Here chosen due to the performance of the device by selecting different sensing regions. it should be mentioned that pumping system, and not due limitations sensor. lowest limit of 50 µL/min was to chosen due to of thethe performance of the pumping system, and not due to limitations of the sensor. 4.4 Influence of the Prism-Liquid Temperature Difference 4.4. Influence of the Prism-Liquid Temperature Difference To research the influence of the prism-liquid temperature difference on the measurement, research influence the prism-liquid temperature difference the measurement, water waterToflow rates the with differentofprism-liquid temperature difference wereon detected. Figure 11 shows flowrelationship rates with different prism-liquid temperature were and detected. Figure 11 the the between the SPR response variationdifference rate ΔRIRF/Δl the logarithm of shows flow rate, relationship between the SPR response variation rate ∆RIRF/∆l and the logarithm of flow rate, where where the water was 25 °C and the prism was 30, 35 and 40 °C, respectively. It can be seen that the water was 25 ˝ Cdifferences and the prism 30, 35detection and 40 ˝ C, respectively. It can be seen larger larger temperature lead was to higher sensitivity, consistent with thethat previous temperature differences lead to higher detection sensitivity, consistent with the previous theoretical theoretical analysis. Therefore, a larger temperature difference is desirable to enhance the analysis. Therefore, larger difference is desirable to enhance theas sensitivity. sensitivity. Howevera in the temperature case of temperature sensitive experiments, such biologicalHowever studies, in the case of temperature sensitive experiments, such as studies, temperature suitable temperature differences are necessary to balance thebiological sensitivity and thesuitable object under test. differences are necessary to balance the sensitivity and the object under test. Compared with the conventional thermal flow sensor, which is of high integration based on Compared with thethe conventional flowissensor, which is due of high integration based for on the MEMS technology, SPR-based thermal flow sensor more complex to its optical system the MEMS technology, the SPR-based flowpaper, sensorthe is SPR moresystem complex to itsoptical opticalelements system for space-resolved temperature sensing. In this useddue several space-resolved temperature sensing.its Inperformance, this paper, thefurther SPR system usedthe several opticalbarrier elements for interferometric imaging to improve lowering integration of the interferometric imaging tocomplex improveinterferometric its performance,imaging further lowering the integration barrier the flow flow sensor. However, is unnecessary for flow rate of detection, sensor. However, complex interferometric imaging is unnecessary for flow rate detection, where SPR where SPR intensity imaging which uses less optical elements, is sufficient. Based on the proof of intensityofimaging which uses optical sufficient. Based onpaper, the proof of concept the concept the SPR-based flowless sensor thatelements, has been is demonstrated in this a new compactofflow SPR-based flow that has imaging been demonstrated in this paper, compact flow sensor on sensor based onsensor SPR intensity has been designed and aisnew under development. Thisbased should SPR intensity imaging has been designed and is under development. This should allow for integration allow for integration of the SPR-based flow sensor that will be greatly improved. As SPR has been of the SPR-based sensor that will bethis greatly improved. been widely applied in widely applied inflow bio-chemical analysis, method should As alsoSPR be has easily integrated into these bio-chemical analysis, this method should also be easily integrated into these analytical systems. analytical systems.
Figure vs.vs. flow rate (Q)(Q) with thethe prism temperature of 30, 35 Figure 11. 11. SPR SPRresponse responsevariation variation(ΔRIRF/Δl) (∆RIRF/∆l) flow rate with prism temperature of 30, and 40 °C. 35 and 40 ˝ C.
5. Conclusions 5. Conclusions In this paper, a microfluidic flow rate detection method based on SPR temperature imaging is In this paper, a microfluidic flow rate detection method based on SPR temperature imaging presented. The proof of concept has been demonstrated through theoretical simulations and is presented. The proof of concept has been demonstrated through theoretical simulations and experiments. The novel calorimetric flow rate measurement method allows reliable monitoring of experiments. The novel calorimetric flow rate measurement method allows reliable monitoring of liquid flow rates in a wide range of tens to hundreds of μL/min. As SPR has been commonly used for
Sensors 2016, 16, 964
9 of 10
liquid flow rates in a wide range of tens to hundreds of µL/min. As SPR has been commonly used for bio-chemical analysis, this novel SPR-based flow rate detection method should also be easily integrated into these systems. Other than water-prism temperature difference, some factors such as liquids of different heat capacities and the flow channel dimensions need to be further researched. Furthermore, the performance of the flow sensor needs to be enhanced, including improving its integration and expanding the detection range for biological applications. Acknowledgments: This work was supported by Tsinghua University Initiative Scientific Research Program (20131089190) and the Fund for Joint Project of Beijing. Author Contributions: P. Wang and X. Yu conceived and designed the experiments; S. Liu and T. Zhao performed the theoretical simulations and analysis; S. Deng and M. Guo performed the experiments; S. Deng and S. Xu analyzed the data; S. Deng and S. Liu wrote the paper. Conflicts of Interest: The authors declare no conflict of interest. None of the material in the paper has been published or is under consideration for publication elsewhere.
Abbreviations The following abbreviations are used in this manuscript: SPR RI RIRF CCD PMMA
surface plasmon resonance refractive index refractive index related factor charge coupled device Polymethylmethacrylate
References 1. 2. 3. 4. 5. 6. 7. 8.
9. 10.
11. 12. 13.
Li, Y.; Yan, G.; Zhang, L.; He, S. Microfluidic flowmeter based on micro “hot-wire” sandwiched Fabry-Perot interferometer. Opt. Express 2015, 23, 9483–9493. [CrossRef] [PubMed] Vilares, R.; Hunter, C.; Ugarte, I.; Aranburu, I.; Berganzo, J.; Elizalde, J.; Fernandez, L.J. Fabrication and testing of a SU-8 thermal flow sensor. Sens. Actuators B Chem. 2010, 147, 411–417. [CrossRef] Lien, V.; Vollmer, F. Microfluidic flow rate detection based on integrated optical fiber cantilever. Lab Chip 2007, 7, 1352–1356. [CrossRef] [PubMed] Sengupta, S.; Ziaie, B.; Barocas, V.H. Lag-after-pulsed-separation microfluidic flowmeter for biomacromolecular solutions. Sens. Actuators B Chem. 2004, 99, 25–29. [CrossRef] Collins, J.; Lee, A.P. Microfluidic flow transducer based on the measurement of electrical admittance. Lab Chip 2004, 4, 7–10. [CrossRef] [PubMed] Ahrens, R.; Schlote-Holubek, K. A micro flow sensor from a polymer for gases and liquids. J. Micromech. Microeng. 2009, 19, 74006. [CrossRef] Ernst, H.; Jachimowicz, A.; Urban, G.A. High resolution flow characterization in Bio-MEMS. Sens. Actuators A Phys. 2002, 100, 54–62. [CrossRef] Caldas, P.; Jorge, P.A.; Rego, G.; Frazão, O.; Santos, J.L.; Ferreira, L.A.; Araújo, F. Fiber optic hot-wire flowmeter based on a metallic coated hybrid long period grating/fiber Bragg grating structure. Appl. Opt. 2011, 50, 2738–2743. [CrossRef] [PubMed] Van Oudheusden, B.W. The determination of the effective ambient temperature for thermal flow sensors in a non-isothermal environment. Sens. Actuators A Phys. 1999, 72, 38–45. [CrossRef] Dijkstra, M.; de Boer, M.J.; Berenschot, J.W.; Lammerink, T.; Wiegerink, R.J.; Elwenspoek, M. Miniaturized thermal flow sensor with planar-integrated sensor structures on semicircular surface channels. Sens. Actuators A Phys. 2008, 143, 1–6. [CrossRef] Kim, T.H.; Kim, S.J. Development of a micro-thermal flow sensor with thin-film thermocouples. J. Micromech. Microeng. 2006, 16, 2502. [CrossRef] Berthet, H.; Jundt, J.; Durivault, J.; Mercier, B.; Angelescu, D. Time-of-flight thermal flowrate sensor for lab-on-chip applications. Lab Chip 2011, 11, 215–223. [CrossRef] [PubMed] Chung, J.; Grigoropoulos, C.P.; Greif, R. Infrared thermal velocimetry for nonintrusive flow measurement in silicon microfluidic devices. Rev. Sci. Instrum. 2003, 74, 2911–2917. [CrossRef]
Sensors 2016, 16, 964
14. 15. 16. 17. 18. 19. 20. 21. 22. 23.
24. 25.
26.
10 of 10
Kuo, J.T.; Yu, L.; Meng, E. Micromachined thermal flow sensors—A review. Micromachines 2012, 3, 550–573. [CrossRef] Wu, S.; Lin, Q.; Yuen, Y.; Tai, Y. MEMS flow sensors for nano-fluidic applications. Sens. Actuators A Phys. 2001, 89, 152–158. [CrossRef] Yan, W.; Liu, C.; Li, J.; Ma, L.; Niu, D. Thermal distribution microfluidic sensor based on silicon. Sens. Actuators B Chem. 2005, 108, 943–946. Sabaté, N.; Santander, J.; Fonseca, L.; Gràcia, I.; Cané, C. Multi-range silicon micromachined flow sensor. Sens. Actuators A Phys. 2004, 110, 282–288. [CrossRef] Kim, S.; Nam, T.; Park, S. Measurement of flow direction and velocity using a micromachined flow sensor. Sens. Actuators A Phys. 2004, 114, 312–318. [CrossRef] Hoera, C.; Skadell, M.M.; Pfeiffer, S.A.; Pahl, M.; Shu, Z.; Beckert, E.; Belder, D. A chip-integrated highly variable thermal flow rate sensor. Sens. Actuators B Chem. 2016, 225, 42–49. [CrossRef] Wagner, C.E.; Macedo, L.J.; Opdahl, A. Temperature gradient approach for rapidly assessing sensor binding kinetics and thermodynamics. Anal. Chem. 2015, 87, 7825–7832. [CrossRef] [PubMed] Srivastava, S.K.; Gupta, B.D. Simulation of a localized surface-plasmon-resonance-based fiber optic temperature sensor. JOSA A 2010, 27, 1743–1749. [CrossRef] [PubMed] Turhan-Sayan, G. Temperature effects on surface plasmon resonance: Design considerations for an optical temperature sensor. J. Lightwave Technol. 2003, 21, 805. Chiang, H.; Yeh, H.; Chen, C.; Wu, J.; Su, S.; Chang, R.; Wu, Y.; Tsai, D.P.; Jen, S.U.; Leung, P.T. Surface plasmon resonance monitoring of temperature via phase measurement. Opt. Commun. 2004, 241, 409–418. [CrossRef] Ekgasit, S.; Thammacharoen, C.; Yu, F.; Knoll, W. Evanescent field in surface plasmon resonance and surface plasmon field-enhanced fluorescence spectroscopies. Anal. Chem. 2004, 76, 2210–2219. [CrossRef] [PubMed] Wang, D.; Ding, L.; Zhang, W.; Luo, Z.; Ou, H.; Zhang, E.; Yu, X. A high-throughput surface plasmon resonance biosensor based on differential interferometric imaging. Meas. Sci. Technol. 2012, 23, 65701. [CrossRef] Moreira, C.S.; Lima, A.; Neff, H.; Thirstrup, C. Temperature-dependent sensitivity of surface plasmon resonance sensors at the gold-water interface. Sens. Actuators B Chem. 2008, 134, 854–862. [CrossRef] © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).
