Anal Bioanal Chem (2010) 398:1525–1533 DOI 10.1007/s00216-010-4052-6
ORIGINAL PAPER
A compact miniaturized continuous flow system for the determination of urea content in milk Willian Toito Suarez & Osmundo Dantas Pessoa-Neto & Vagner Bezerra dos Santos & Ana Rita de Araujo Nogueira & Ronaldo Censi Faria & Orlando Fatibello-Filho & Mar Puyol & Julián Alonso
Received: 12 May 2010 / Revised: 9 July 2010 / Accepted: 20 July 2010 / Published online: 9 August 2010 # Springer-Verlag 2010
Abstract A multicommutation-based flow system with photometric detection was developed, employing an analytical microsystem constructed with low temperature cofired ceramics (LTCC) technology, a solid-phase reactor containing particles of Canavalia ensiformis DC (urease source) immobilized with glutaraldehyde, and a miniphotometer coupled directly to the microsystem which monolithically integrates a continuous flow cell. The determination of urea in milk was based on the hydrolysis of urea in the solid-phase reactor and the ammonium ions produced were monitored using the Berthelot reaction. The analytical curve was linear in the urea concentration range from 1.0×10−4 to 5.0×10−3 mol L−1 with a limit of detection of 8.0×10−6 mol L−1. The relative standard deviation (RSD) for a 2.0×10−3 mol L−1 urea solution was lower than 0.4% (n=10) and the sample throughput was 13 h−1. To check the reproducibility of the flow system, calibration curves were obtained with freshly prepared solutions on different days and the RSD obtained W. T. Suarez (*) : A. R. de Araujo Nogueira Embrapa Pecuária Sudeste, Caixa Postal 339, 13560-970 São Carlos, SP, Brazil e-mail:
[email protected] O. D. Pessoa-Neto : V. B. dos Santos : R. C. Faria : O. Fatibello-Filho Departamento de Química, Centro de Ciências Exatas e de Tecnologia, Universidade Federal de São Carlos, Caixa Postal 676, 13560-905 São Carlos, SP, Brazil M. Puyol : J. Alonso Sensors and Biosensors Group, Department of Chemistry, Facultat de Ciènces, Edifici Cn, Universitat Autònoma de Barcelona, 08193 Barcelona, Spain
was 4.7% (n=6). Accuracy was assessed by comparing the results of the proposed method with those from the official procedure and the data are in close agreement, at a 95% confidence level. Keywords Urea . Milk . Low temperature co-fired ceramics . Berthelot’s reaction . Multicommutation . Mini-photometer
Introduction Urea has an important role in the metabolism of nitrogencontaining compounds by animals and is the main nitrogencontaining substance in the urine of mammals. It is the non-protein nitrogen compound most often used in cattle feed because of its high nitrogen content [1–3]. Urea measurement allows one to assess the nutritional management of lactating dairy cows because of the high urea content in milk, which may influence milk production [4]. Higher urea levels are related to an increased concentration of non-protein nitrogen in the milk that can affect cheese production, as part of the true protein formed by casein and the serum are replaced by urea [3, 5]. Because of these factors and the growing demand for rapid and easy monitoring of cattle feed, it is necessary to develop versatile miniaturized systems to determine urea in milk, so the user can perform the analysis in situ, in places where large devices would be impractical and consequently real-time analytical information could not otherwise be obtained. The use of analytical microsystems [6, 7] offers many advantages, which include the possibility of large-scale production, which reduces the manufacturing costs, and
1526
low reagent consumption and volume of waste because of the small device size, high sample throughput, and portability. Many technologies have been developed to miniaturize analytical systems. One of them is the use of integrated circuits (ICs) technology based on silicon [7]. However, ICs have some inconveniences, such as long chip development process time and the inherent bidimensionality of the manufactured structures. Polymer technology, using multilayer approach, allows one to obtain tridimensional structures rapidly and at a low cost but has problems related to both chemical compatibility and liquid leakage due to sealing deficiencies during the lamination process. An alternative is to use the green tape or low temperature co-fired ceramics technology (LTCC) [8–10]. Green tape ceramic substrate (not sintered) and its associated technology represent a promising alternative for the development of miniaturized analytical systems. This technology allows a much more versatile construction of three-dimensional structures (both microfluidic and micromechanical) by employing a multilayer approach without showing sealing problems, and the device prototyping can be performed rapidly and at a low cost. Additionally, the cost and maintenance of the necessary infrastructure are very low because the use of clean rooms is unnecessary. Several procedures to determine urea in different matrixes have been proposed [11–16]. Flow techniques employing conductometric [17], potentiometric [18], and spectrophotometric detection [19] have also been successfully applied for this purpose. De Faria et al. [17] developed a conductometric flow injection system to determine urea in human serum by using naturally immobilized urease enzyme. In this system, the ammonium ions produced in the enzymatic reaction were converted to NH3 after addition of an alkaline reagent. An increase in the conductance, proportional to the urea concentration, was obtained. A sequential injection system was proposed by Silva et al. [18] for potentiometric determination of urea. In this procedure, a tubular ammonium-selective electrode was employed to assess the ammonium concentration produced by enzymatic hydrolysis of urea from the crude extract of jack bean meal (Canavalia ensiformis DC). In this study, we report the development of an automatic multicommuted flow system with photometric detection (modified Berthelot’s reaction). The system is composed of an analytical microsystem constructed with LTCC technology incorporating a monolithically integrated optical flow cell and a solid-phase reactor with particles of Canavalia ensiformis DC (urease source) immobilized with glutaraldehyde. The detection was carried out with a lab-made mini-photometer, which is a device that has the features based on an ultra-bright light-emitting diode (LED) and a photodiode (PD) as detector.
W.T. Suarez et al.
Materials and methods Reagents and solutions Deionized water (resistivity≥18.2 MΩ cm) was obtained by employing a Milli-Q plus system (Millipore Corp., Bedford, MA, USA). The water was used to prepare all aqueous solutions. All reagents used were of analytical grade. A 0.2 mol L−1 sodium salicylate solution was prepared in 0.1 mol L−1 sodium hydroxide containing 3% (w/w) sodium nitroprusside. A 1.0% (v/v) sodium hypochlorite solution was prepared daily from stock solution. Finally, 1.0×10−2 mol L−1 urea stock solution was prepared daily by dissolving 60 mg of urea in 0.2 mol L−1 sodium chloride in a calibrated flask, until a volume of 100 mL was reached. Standard solutions containing 1.0 × 10−4 to 5.0 × −3 10 mol L−1 of urea were prepared daily by serial dilutions of appropriate volumes of the stock solution in 0.2 mol L−1 sodium chloride. A 0.2 mol L−1 Tris-HCl solution (pH 6.7) was prepared by mixing 50 mL of 0.20 mol L−1 HCl with 50 mL of 0.80 mol L−1 tris-(hydroxymethyl)aminomethane (96.8 g in 1,000 mL of water) and this was then mixed with water to obtain a 200-mL volume. The enzymatic UV test kit used for comparative purposes was purchased from Human do Brasil (registration number 10302240341). Apparatus The flow analyzer set consists of a peristaltic pump (Ismatec, model 7618-50, 12-channel, Zurich, Switzerland) equipped with Tygon® tubes, four three-way solenoid valves (Cole-Parmer, model EW-98302-44 with 12 VDC), and a mini-photometer as detector system, driven by a USB interface. A lab-made mini-photometer was built by using an ultrabright LED (Agilent Technologies, model HLMP-ED-25TW000) as light source emitting radiation at the wavelength of maximum intensity (λmax) of 639 nm, with a bandwidth (λ1/2max) of 17 nm and a photodiode (Texas Advanced Optoeletronic Solutions (TAOS), Plano TX, USA, model TSL250R) as detector. An Intel Pentium V microcomputer (PC) equipped with a USB interface (National Instruments, model NI USB6008) was used to control the system and to perform data acquisition and treatment. The software was developed in Labview® 8.5. A lab-made solenoid valve actuator was built to control the multicommutation of the solenoid valves. An Orion pH meter (model EA 940) equipped with an Analion combined glass electrode (model V620, Analion
Compact miniaturized continuous flow system
Inc., Westford, MA, USA), with an external Ag/AgCl (3.0 mol L−1 KCl) reference electrode was used to measure pH. The bean samples were ground in a cutting mill fitted with a 20-mesh screen at the bottom of the cutting chamber (Tecnal mill, model TE 631/1, Piracicaba, SP, Brazil). Components of the flow system based on multicommutation and its functionality A microcomputer, coupled with a USB interface, was used to control and receive the data. For this, software written in Labview 8.5 was used. The interface consisted of 12 analog input ports and 8 digital input/output ports with a resolution of 12 bits. Each digital port defines the address base of the devices coupled to it. Employing this interface, it is possible to switch on the LED, the photodiode, and the solenoid valve actuator by applying a voltage of 5.0 V. For this, a logic level (I/O) of one of the bits is sent by the control lines from the digital ports of the interface. In addition, the interface is used to acquire the signals generated by the photodiode and to plot them on the monitor screen. An electronic filter—composed of two resistors (2.0 kΩ and 27 kΩ), a capacitor (4.8 μF), an operational amplifier 741—and a resistor (potentiometer of 1 kΩ) are used to reduce the noise and to control the luminosity of the LED, respectively. A continuous current of 7.0 mA is supplied to the LED through the interface coupled to it. The TSL250R sensor used (Monolithic Silicon IC) has a wide linear response with good sensitivity and low noise (offset between 0–4 mV) in the region of the electromagnetic spectrum (400–900 nm), with maximum absorption at 740 nm. In the system developed, the LED emission radiation is partially absorbed by the chromophore present in the optical path of the LTCC, before reaching the photosensitive surface of the sensor (1 mm2) generating an electrical current that is converted to potential and it is measured in the output channel (pin 3). The analytical signal is collected by the USB interface by using the input analog port and converted to digital by using the analog–digital converter (ADC) present on the interface. After that, this signal is converted to absorbance by applying the Lambert–Beer equation which was written in software developed in Labview language. A solenoid valve actuator with an integrated circuit (ULN2008) is used to control the solenoid valves. The switching on of each solenoid valve is controlled from a digital signal sent by the computer bit by bit through the software via the interface. These bytes are available in an integrated circuit, generating an electrical signal of 12 V. Nevertheless, in operational conditions, a voltage of 10.5 V and an electric current of
1527
107.5 mA for each solenoid valve (total current of 430 mA) can be used to increase the solenoid valve’s lifetime. Figure 1 shows a block diagram of the components of the compact miniaturized continuous flow system developed to determine urea in milk by employing a multicommutation approach. The PC, through the software written in Labview 8.5 graphic language and the interface (IF), controls the switching action of the LED, PD, and SVA. Performance tests and time response of the mini-photometer used as photometric sensor In order to evaluate the performance of the photometric detector some studies were carried out (Fig. 2). For this purpose, the signal of potential (V) was estimated by measuring the incident radiation on an LED positioned at 180° to the photodetector with the data conveniently amplified by the integrated circuit 741, filtered, and input via the analog port of the USB interface. The USB interface with an analog–digital converter of 12 bits is able to furnish up to 1.221 mV of resolution to the developed photometer. In other words, using this photometric detector it is possible to detect variations of the light intensity greater than 1 mV. As can be seen in Fig. 2, the detector signal started around 5 ms and increased up 37 ms when it reached 4.627 V. After 60 ms, the signal is quite stable at 4.275 V. This slight decrease of potential in the interval from 37 to 60 ms is owing to employment of the low-pass Butterworth filter with third-degree polynomial for which the cutoff frequency is 20 Hz. This virtual filter developed in Labview software caused a little reduction in the signal of the photometer as expected. On the other hand, with this filter the photometer has a more stable signal and reduces its
Fig. 1 Diagram of the components employed to determine urea in milk. The proposed system is composed of a mini-photometer (MP) (basically composed of an LED, LTCC, and photodiode (PD)), solenoid valve actuator (SVA), USB interface (IF), and a personal computer (PC)
1528
W.T. Suarez et al.
Fig. 2 Time response of the mini-photometer used as photometric detector
level of noise to ca. 5 mV, which is compatible with the offset voltage of the photodiode presented in the manufacturer’s datasheet, typically between 0 and 4 mV. The stability of the signal was monitored over 5 h and a potential of 4.227± 0.001 V with a relative standard deviation of 0.023% (n=60) was obtained. This good performance of the lab-made mini-photometer is very important for acquiring satisfactory results using the proposed analytical method. Preparation of the solid-phase reactor containing fragments of Canavalia ensiformis DC The solid-phase reactor containing the natural source of the urease enzyme (Canavalia ensiforms DC) was prepared by using a 60-mm-long and 4-mm-internal-diameter Tygon tube. It was filled with 200 mg of milled beans (particle size 100–350 μm) immobilized in 12.5% (v/v) glutaraldehyde in 0.2 mol L−1 Tris-HCl buffer solution (pH=6.7). The particles were placed in the Tygon tube with the aid of a syringe and then pieces of glass wool were placed at both the ends of the tube to prevent fragments from entering the transmission lines. Construction of the analytical microsystem using LTCC technology The LTCC device was constructed using green ceramic tapes (951-AX DuPont Microcircuit material) with 254-μm thickness. The microsystem was designed using the AutoCAD (Autodesk) program by employing the multilayer technique. The design of the ceramic tapes was made taking into account that their final overlap should result in the threedimensional geometry desired and, mainly, that ceramic
layers will suffer ca. 15% shrinkage after the sintering process; because of this fact, the dimensions were incremented 15% in the CAD design. The green tapes were processed by an LPKF Protolaser 100 machine (LPKF, Germany) to obtain microchannels after sintering with 0.85-mm width and 0.17-mm depth and cavities with diameters of 1.5, 3.0, and 10 mm. Figure 3 shows each of the layers used to construct the analytical microsystem used in this study. Layer A is the top and layer E is the bottom of the device. The number of layers that can compose an analytical microsystem using LTCC is variable. In general between six and twelve layers are used. More can be prejudicial because during the lamination step the internal microfluidic structures may be damaged [8]. The lamination of the layers to construct one block was performed by using a hydraulic press (Talleres Francisco Camps SA, Granollers, Spain); ca. 3,000-psi pressure was applied at 100 °C for 1 min. The sintering was then carried out using a programmed furnace (Carbolite CBCWF11/ 23P16, Afora, Spain). The temperature program was selected and implemented as per the green tape manufacturer’s data sheet with the final temperature being 850 °C. Seven sheets of green tape were used to construct the analytical microsystem, one green tape for layers A, B, D, and E and three for layer C. The flow cell with a 0.5-mm optical path was constructed by placing two Pyrex glass sheets in the top and bottom layers. Microfluidic channel in layer C was constructed with a meander configuration to enlarge the reactor length, thereby enhancing the mixing process between the reactants and the analyte and increasing residence time. Layer A defines the inlets (a, b, and c) and outlet (d) of the solutions. In this layer, brass connectors (fixed with epoxy resin) are coupled to allow the solutions to enter and
Compact miniaturized continuous flow system
1529
reactor containing fragments of beans (Canavalia ensiformis DC, urease source) and determination of the ammonium ions produced through Berthelot’s reaction [20]. Figure 4 illustrates the analysis module employed to measure urea content in milk. The measuring process involves three steps, depending on the characteristics of the chemical reaction: (1) urea hydrolysis and ammonium generation; (2) formation of indophenol blue, and (3) system cleaning. All the steps were realized through the software written in Labview 8.5 and the USB interface. In the first step, valve V1 is switched on for 4 s after inserting 200 μL of the sample in the system. After that, valve V2 is switched on for 30 s, carrying the sample zone to until point x, where hydrolysis of the urea occurs, generating ammonium ions by the action of the urease enzyme immobilized in the reactor. In the second step, valves V2, V3, and V4 are simultaneously switched on for 12 s. The ammonium ions produced in the first step react with the salicylate and then with the hypochlorite at the confluence point z. The Berthelot reaction is based on the reaction between ammonia in alkaline medium, hypochlorite, and salicylate (instead of the original reagent phenol) catalyzed by the nitroprusside, which results in the formation of indophenol blue, which was photometrically monitored at 639 nm in the optical flow cell path (k) integrated in the microsystem. Because of the slow reaction rate, 3 min were necessary to acquire the analytical signal. Finally, the microchannels of the microsystems are cleaned by switching on valve V2 for 40 s, and afterward a new sample can be inserted to start a new analysis cycle.
Fig. 3 Multilayer technique employing green tapes to fabricate the microsystem
exit the microsystem. The cavities a, b, c, and d of the top layer have a diameter larger than the cavities of layer B, as these cavities are adjusted to allow the solutions to enter layer C, which determines the shape of the microchannels. The analysis module The method employed to measure the urea content in milk is based on the hydrolysis of the urea in a solid-phase
Fig. 4 Schematic diagram of the flow system. C carrier solution; S sample; R1 hypochlorite; R2 solution of salicylate and of the nitroprusside in alkaline medium; a, b, and c inlet of solutions in the microsystem; d outlet of solutions of the microsystem; SPR solidphase reactor; z confluence points; V1, V2, V3, and V4 three-way solenoid valves; k optical path of 0.5 mm; W waste
1530
The computer is used to select the steps involved in the analytical procedure: measurement of the blank signal, measurement of the analytical signal, and cleaning of the system, besides the control variables—timing of opening and closing the valves, number of replicates, and time interval to acquire the analytical signal. Once these data are obtained, the software takes over the analytical process, performing the sequence of opening and closing the valves according to the flowchart shown in Fig. 5.
Results and discussion Optimization of the method for urea measurement in milk Most of the available spectrophotometric procedures employ Berthelot’s reaction (phenol–hypochlorite) to measure ammonia and urea in various samples [20]. In this study, phenol was substituted by salicylate, which is a less toxic reagent. The proposed method is based on the hydrolysis of urea in the solid-phase reactor containing bean fragments (Canavalia ensiformis DC), producing ammonium ions that react with salicylate and hypochlorite in alkaline pH, generating indophenol blue, whose color intensity is proportional to the urea level in the sample analyzed. To optimize the chemical and flow system variables, the univariate method was employed in all the experiments performed in this study to identify the best compromise between the magnitude and repeatability of the analytical signals, stability of the baseline, and sample throughput. All experiments were carried out at a constant temperature of 25±1 °C. There are several enzymatic sources that contain urease in their substrates. However, for an analytical method to be Fig. 5 Software flowchart employed to control the analysis module
W.T. Suarez et al.
feasible it is necessary to have an accessible source and be able to achieve a high rate of analyte conversion, or else to convert urea into ammonium efficiently. Thus, the first experiments were performed using the Cajanus cajans L. Millsp. bean as a urease source. The results obtained through this urease source had very low absorbance values, even when using 1.0×10−2 mol L−1 of urea. On the other hand, the initial experiments using Canavalia ensiformis DC showed excellent results, with significantly higher absorbance values than those obtained with Cajanus cajans L. Millsp. Therefore, Canavalia ensiformis DC was chosen as the urease source. One of the disadvantages of a microsystem constructed with LTCC is its small optical path. It is therefore necessary to search for alternatives to increase the sensitivity of the method. An alternative to improve the sensitivity of the flow system employing Berthelot’s reaction was proposed by Luca and Reis [21]. These authors used a solution of 0.25 mol L−1 NaCl in an enzymatic flow injection system with spectrophotometric detection to measure urea in animal blood plasma. This significantly increased the magnitude of the analytical signal, probably due to the hydration of the enzymatic source when the solution passed through the reactor, because in saline medium the cells absorb more water to establish osmotic equilibrium. We also used a 0.25 mol L−1 NaCl concentration, which provided a fivefold increase in the analytical signal compared to the experiments conducted without use of this salt. To avoid high pressure due to clogging inside the reactor, the milled beans were immobilized in 12.5% (v/v) glutaraldehyde in 0.2 mol L−1 Tris-HCl buffer solution (pH=6.7), as described by Rover-Júnior et al. [22]. The ground bean particles were placed in Tygon tubes with an internal diameter of 4.0 mm.
Compact miniaturized continuous flow system
Particle size and length of the reactor—the main factors related to the solid-phase reactor that affect the indophenol blue formation—were evaluated. The particle size is an important factor for efficient conversion of urea into ammonium, because when large particles (>500 μm) are present in the reactor, the magnitude of the analytical signals is affected due to the small contact surface area between the sample zone and the immobilized enzyme. On the other hand, small particle size (
