3D printed opto-microfluidic autonomous analyzer for photometric applications

Graphical abstract


Hardware in context
Automation and miniaturization are trends for the development of analytical systems. At present, the design of these systems is based on constraints: a high performance of in-situ measurements, high selectivity, multi-sample analysis and devices highly adaptable to different analytical methods. With the miniaturization, one can achieve a significant reduction in the volume of reagents used and waste production during the procedure, and also an increment in the number of analysis samples per day by reducing the analysis time and cost [1,2]. Terms Lab-on-a-chip and lTAS (micro total analysis) [3,4], are considered synonyms for devices that work with fluids and integrate a number of different functionalities on the small scale [5,6], designed for the implementation of analytical process: including sample preparation, transportation, mixing, separation, reaction, detection, signal acquisition and processing. These analytical systems are generally related to the sample volume of the process, not to the structure's size [7]. The method of flow injection analysis (FIA), is one of the most used automated analysis method, due to its simplicity, versatility, low cost and its fast response time, so it can be used in monitoring systems [8,9]. Microfluidics allows the possible to integrate multiple fluid tasks into a chip. The detection cell consists of the optical elements like lenses and waveguides and optoelectronic devices such as light sources, sensors, some ones could be kept off the chip. According to these characteristics, the opto-microfluidic systems provide significant advantages: high portability, efficiency, sensitivity, versatility, and low cost [10].
In biochemical and chemical analysis, usually, the samples are analyzed with two or more solutions or reagents. These solutions must be completely mixed for the chemical reaction to take place. In the opto-microfluidics structure, the mixture is developed into microchannels, reactors, heaters and sensing cells with very small Reynolds numbers. The mixed solutions are carried out mainly by diffusion assuming a laminar behavior. The small distances within the microfluidic channels enable complete mixing in a short time period, in the order of seconds. While in macroscale the mixture is achieved by turbulence.
Colorimetric and UV-VIS spectrophotometry techniques have been used in many analytical methods. They can be used for stand-alone or combined with separation/flow-through systems such as HPLC, FIA, electrophoresis or even portable point-of-care devices [11]. The development of portable and miniaturized devices based on colored reactions is an area still in development due to new materials, manufacturing techniques and an improvement in optoelectronic and electromechanical devices [12], not forgetting the novel computer design tools. Generally, microfluidic systems are manufactured using silicon [13], glass [14,15], polymers [16], or LTCC (low temperature co-fired ceramics) [17] technologies, which require a sophisticated and very expensive infrastructure. For that technologies, the most of the studied cross-sections channels are limited to only rectangular and more recently trapezoidal shapes. Additive manufacturing technology is used as an alternative to those manufacture procedures [18], where one of the most common methods is the Fusion Deposition Modeling (FDM) [19,20], also called ''3D printing".

Hardware overview
Instrumental methods for flow analysis have been evolving in order to carry out continuous control of a greater number of samples. The flow injection technique is generally the most used automated method. To carry out this instrumental method, large actuators are integrated, such as a peristaltic pump with an approximate cost of $1,580 and modular valve positioners with an approximate cost of $1,500, and to perform spectrophotometric analysis a Vis spectrophotometer with a cost of $4,437.44 and/or USB4000 spectrophotometer, cost $618. An analytics software application is typically priced at $226. The management of fluids using large volumes of reagents and the waste generated also generate an impact in the cost. Likewise, for determinations of low concentrations of metal ions, colorimetric methods, atomic absorption spectrophotometer ($10,000-$25,000), atomic emission spectrometer ($25,000-$33,000), etc. are integrated into the analytical process. Incorporating this type of equipment in instrumental methods of analysis has the following disadvantages. 1. The analytical process is required to be developed in central laboratories, making it impossible to implement it in-situ or on-line in the measurements. 2. The consumption of reagents as well as the production of waste is high, which is why it is expensive. 3. When performing off-line analysis, the information obtained is not in real time while performing the analysis. 4. They require trained human resources for the operation of this equipment in the analysis.
The proposed hardware of the 3D printed opto-microfluidic autonomous analyzer for photometric applications can be conceptualized in the following four functional modules: 1 Control and data acquisition, 2 Hydrodynamic system (fluid pumping and flow injection), 3 Optical detection system and 4 Wireless communication system. These modules were integrated to an opto-microfluidic structure were the analytic processes take place. Fig. 1 shows the block diagram of the basic elements of the design. The system was designed in modules looking its versatility and all operations should be developed by the control software (virtual interface). This modular design will allow the extension or mod-ification of device without the need for major changes. The opto-microfluidic autonomous analyzer consists of four solution inlets and, one waste outlet and a fluid circuit in an area of 80 mm Â 67 mm with a width of 1.5 mm. While, the detection section consists of a light source, a sample cell with an optical path of 30 mm and a photo-detector array. The handling of fluids in the fluidic circuit and the optical detection system are controlled by a microcontroller with a virtual interface.
The miniaturization offered by this opto-microfluidic analyzer has the following characteristics: Fast response. Low reagent and power consumption for its operation. Autonomous, portable and easy-to-operate by Wireless communication.
Micro-Analyzes in -situ, on-line and in real time. Low cost $2,050.40 USD The integration of micro-devices and opto-microfluidic structures to perform the mixing between solutions and optical detection produces a robust, flexible, portable and low cost analytical microsystem.
The proposed system that is described in this work presents a modular integration and automation of the analytical process not reported in other 3D printed devices developed for similar applications. The main characteristics of operation and design of the 3D printed devices developed were compared with the performance of the 3D Opto-microfluidics autonomous analyzer. This is shown in Table 1.

Build instructions
The developed hardware has four functional modules whose details are described below.

Control and data acquisition
The developed electronic control and data acquisition was implemented in a compact PCB module 50 Â 50 Â 10 mm low cost manufactured by JLCPCB (jlcpcb.com, China). The electronic design consist of a 10-bit ADC with 1 analogic input, and 19 digital outputs, see Fig. 2. The system was designed for control and data acquisition of an automated analytical process. The module was designed with the versatility of performing analytical processes for other photometric measures just by reprogramming the control module without the need for major hardware changes. The analytic process was configured and  operated by the virtual program. The module includes the electronic control for data acquisition, flow sequence, and a wireless communication system. The control was implemented with a PIC18F4550/QFN44 microcontroller unit MCU (Microchip Technology Inc., Chandler, USA). To enable the portability of the system, it was fed with a lithium-ion rechargeable battery VSO-F550H (Amstron PS, Valencia, USA) of 7.2 V-2.2 A/h and a regulated eliminator ELI-2500 from 3 to 12 Vdc, 2.5 A (Steren, USA) as second source. Wireless communication between the device and control electronics consisted in Bluetooth modules LM-058 (LM Technologies, London, UK). A regulated bank was integrated into the control electronics to maintain the independent supply to different devices avoiding fluctuations to obtain a correct measurement.

Power supply
The power supply consists of two power sources: an ELI-2500 regulator and a VSO-F550H battery with a flyback type topology. The voltages delivered by the source are 5 V and 3.3 V referred to a common ground (VSS). The power supply is regulated by 5 solid-state voltage regulators as shown in Fig. 3. U5 regulates the power to the MCU at 3.3 V; U1 feeds a TRIM adjusted (voltage divider) at 5 V as a radiation intensity control of LED to avoid over saturation of the photodetector; U2 provides a regulated 5 V voltage to power the micro-pump drives, the wireless module and the saturation common of the micro-valve power stage transistors. U3 supplies a regulated 5 V to the photo-detector and U4 supplies 5 V to the multispectrum array.
In photometry, the light absorbed by a sample is a function concentration and optical path. Therefore, fluctuations in the light intensity of the beam can affect the measurements. For these reason, the light source should be stabilized in intensity emitting a continuous beam of light at specific wavelength. In general, the light intensity fluctuations can be due to variations in voltage or current, so it is powered independently and with a voltage regulation system. The independence of the regulators U1, U3 y U4 is important to ensure a stable voltage that will produce a stable light intensity.

Digital output
The control provides 19 digital outputs as shown in Fig. 4. 12 outputs to control the micro-valves (RA0-RA5 and V1-V6); 4 to control the multi-spectrum matrix (CTRL1-CTRL4) and 3 to control the micro-pumps (RB4, RB5, RC4). The digital outputs have a light indicator to show the status of the micro-valves.

Analog input
The control provides 13 analog inputs. Fig. 5, depicts a connector block were we used only the RB3 signal for the acquisition of the analytical signal photodetector through connector J9. VDETECTOR is voltage out coming from a voltage regulator 5 V which supplies the photodetector.

PWM signals
For the acquisition of the signal originating from the multi-spectrum sensor, the FIN input was used as shown in Fig. 4. This discrete signal has a frequency range from 200 Hz to 7.5 KHz, which works with a MCU timer to quantize the operating frequency.
The PWM signal are generated by the MCU timers for the control of the micro-pumps which have an operating frequency of 0 Hz to 60 Hz to generate flow rates up to 7.5 mL/min. Fig. 4 shows the PWM signal RC4, RB4 and RB5, these modulated signals are applied to the MPD-200A micro-pump driver control for its operation.

Wireless communication
For communication a serial port was used. The transmission speed of this port is configurable into the range from 1.2 to 115,200 bauds. Through J12 with a mini-usb interface port, the data are received and sent through the TX and RX control lines as see in Fig. 4. The communication port is based on the RS232 protocol with ''Half Duplex" communication.

Maintenance and programming
The maintenance stage includes the necessary connection for program updates through a programmer Pickit TM 3 (Microship Tech., USA). In the programming of the MCU, the PIC-C software was used. Fig. 4 depicts the signals MCLŔ /VPP, VDD, VSS, PGD (ICSPDAT), PGC (ICSPCLK) and PGM (LVP) for programming.

Connectors description
J1 as shown in Fig. 4, connects the external power supply and the safety voltage output through VCC on pin 1 and VSS on pin 2, both voltajes are supplied by the ELI-2500 source. While Pin 3 with the nomenclature V5VB is the 5 V voltage output to the micro-pumps, in the same way pin 4 JV3.3 is the 3.3 V voltage output that feeds the MCU and pin 5 VSS is the voltage reference or ground.
In Fig. 4 the connector J2 was used to obtain signals from the TCS3200 detector. The CTRL1 and CTRL2 pins select the scaling of the sensor output signal; CTRL3 and CTRL4 select the group of photodiodes from the RGB array. FOUT is the signal processed by the MCU which is a frequency modulated signal with 50% duty cycles. VMATRIX is the 5 V supply voltage to the detector, which is an independent voltage to avoid electronic noise in the acquired analytical signal.
J3 as shown in Fig. 4, connect the signals RB4, RB5 and RC4 from MCU that was used to the control drivers of the MPD-200A. The output voltaje in V5VB is used as power supply of 5 V for these drivers. Also through this connector, power is supplied to the entire system through VCC and VSS.
J4 connector has the function of sending the control signals to the micro-valves through pins V1, V2, V3, V4, V5 and V6. The 5 V power supply uses the V5VB and VSS for the power stage of the injection module. J5 connector controls the intensity of light source (micro-LED) through a TRIM adjustment composed of two potentiometers RV1 and RV2 allowing up to a maximum of 150 mA see in Fig. 6.
J6 connector corresponds to the programming inputs of the MCU. It is a maintenance and firmware update port. The MCU uses a high-speed crystal at 4 MHz to obtain a very high precise and stable frequency of oscillators. An arrangement of 22pF capacitors was implemented to VSS for the management of the timers. The electronic diagram is shown in Fig. 7.
It is recommended to place a decoupling ceramic capacitor at the power supply inputs of the MCU between the 3.3 V and VSS lines.  In the control and data acquisition module, six micro-LEDs were integrated to monitor the states of the micro-valves, to show if the signal that activates the valves was established through wireless communication, which does not interfere with the LEDs, nor with the micro-valves. The micro-LEDs used have a dimension of 0.8 Â 1.2 Â 0.25 mm with a clear water lens, a operating current consumption of 20 mA and a emitting wavelength at 470 nm (Kingbright, USA). Its electronic diagram is shown in Fig. 8.

Fluid pumping module
For the continuous flow system piezoelectric micro-pumps (Takasago Electric, Inc., Nagoya, Japan) were used, with dimensions 25 mm Â 25 mm Â 4.8 mm. All external connecting tubes were silicone tubing (NResearch, N.J., USA) with a 1.5 mm internal diameter. The micro-pumps were controlled by MPD-200A controller (Takasago Electric, Inc., Nagoya, Japan). Each controller can operate 2 micro-pumps.   Micro-pumps can be controlled by frequency or voltage. For this device, frequency control was selected. The MPD-200A controller can generate a range of frequencies from 1 Hz to 60 Hz with a maximum flow rate of 9 mL min À1 and minimum flow rate of 650 lL min À1 . Three MPD-200A controllers with dimensions of 30 mm Â 30 mm Â 18 mm are connected as well as the VSO-F550H battery. Through J3 connector there is communication interface with the control and acquisition module. The control is done through the RC4, RB4 and RB5 pins of the MCU which send frequency modulated signals to the MPD-200A.
In Fig. 9, one seen that U4, U5 and U6 correspond to 3 MPD-200A controllers, while J1 connector corresponds to the communication with the control and acquisition module through its connector J3. J4 connector corresponds to the connection of the pumps to their respective controllers and J7 connects the VSO-F550H battery.
The pumping module is shown in Fig. 10. The micro pumps were fixed with epoxy glue on an acrylic sheet with a dimension of 14.5 Â 3.5 cm.

Flow injection module
An injection system was designed by multi-commutation using three-way solenoid micro-valves (161T010, NResearch, NJ, USA). In order to develop a robust and flexible hydrodynamic system, a collector with 6 incorporated micro-valves was implemented for future analysis in case that a greater number of reagents will be involved. Each micro-valve can be Fig. 10. Flow injection module with micro-valves and micro-pumps assembly. interconnected with a micro-pump, so this hydrodynamic system has the capacity to manage 6 micro-valves for 12 solution outlet ports where each port can be configured to a different injection volume. Fig. 10 depicts the flow injected module. The micro-valves are powered by 5 V, in multi-conmutation, considering an on time (open) of 20 ms and an off time (closed) of 30 ms. The power circuit was implemented with 6 NPN Bipolar Transistors -BCP56-16/SOT223/SC-73 (Nexperia, Guangdong) allowing a current collector-emitter, Ice = 500 mA. Each transistor energizes the micro-valves, it switched from VCC = 5 V and to VSS ground, voltage obtained from the control and acquisition module as shown in Fig. 11. J7 connector receives the control signals from the control and acquisition module. From J1 to J6, the power supply is connected to the micro-valves V1, V2, V3, V4, V5 and V6 and diodes D1, D2, D3, D4, D5 and D6.
The function of the injection module is to control the flow (volume) of reagents and samples to the microanalyzer where the analytical process takes place.  Optical detection module The optical radiation source consists of an RGB LEDs module, size of 6 Â 6 mm 2 (Mod. AAA-505103, Kingbright Co., CA, USA). The packaged dimensions are 4 Â 4 mm 2 for lens and the 3 micro-LEDs distribution into de radiation area of the RGB LEDs is <2 mm 2 with 120°of viewing angle [25]. The emission wavelengths are: for the blue LED k = 450 nm with a spectral linewidth of 20 nm and optical power of 0.6 W; for the red LED k = 624 nm, with a line width of 30 nm and an optical power of 0.336 W and for the green LED k = 525 nm with a line width of 30 nm and an optical power of 0.6 W. In addition, a second LED module was integrated, with a size of 1.6 mm Â 0.8 mm Â 0.6 mm (KG DELLS1.22-JGKH-24, ams OSRAM, Regensburg, Germ.), whose emission wavelength is k = 570 nm with a spectral bandwidth of 22 nm and optical power of 0.5 W. These modules can be replaced with different radiation sources according to the wavelength required for other colorimetric analysis, demonstrating its versatility. The light coupled into the sample cell is well homogenized and verified it with a spectrophotometer in the output window of the sample cell.
The detection module consists of a TAOS TCS3200 programmable light-to-frequency converter, (Texas Advanced Optoelectronic Solution TAOS, Texas, USA.) and dimension of 5 Â 6.5 mm. This detector integrates an array of 64 photodiodes of which 16 photodiodes correspond to each color filter: red, green, blue (RGB) and another 16 photodetectors without filter (clear) for a greater detection spectral range. All four types of photodiodes are interdigitated to minimize non-uniformity of incident light radiation and are interspersed within the photoactive area. It is possible to select the degree of scaling of the obtained signal through an internal frequency divider that achieves frequency values of 100%, 20%, 2% and off. The detector window has a diameter of 2.8 mm where the photodiode array with a photoactive area of 1 Â 1 mm is located. The light source and detection module, was placed and integrated modularly on the microfluidic device as shown in Fig. 12.
The output of the TAOS TCS3200 sensor is a 50% duty cycle square signal with a frequency proportional to light intensity. The full scale output frequency can be scaled by one of three present values through two control inputs. Digital inputs and digital outputs allow direct interface to the microcontroller or any other logic circuit. The output is put into a high impedance state to share this line to the microcontroller. The S2 and S3 inputs are used to select which group of photodiodes (blue, green, red, and white) will be actívated. While, S0 and S1 are used to select the scaling of the output signal, as well as to turn off the device. This type of logical operation makes it possible to implement it through control electronics. In Fig. 13 see a schematic diagram. Fig. 13. Functional block diagram of detector [26].

Wireless communication
The transmission and reception of data between the hardware of the opto-microfluidic analytical system and the computer that processes the information was carried out by wireless communication using two LM-058 modules with Bluetooth protocol, with a communication range of up to 100 m, than that provided by wiring (maximum 12 m).
The wireless communication is focused on the automation of analysis through a graphical interface developed in a virtual instrument software. The main objective for the interface consists in the remote operation of the microanalyzer and the visualization of the analytical data. The link was point to point in an ad-oc network, where the PC is configured as master and the microanalyzer as slave. In Fig. 14, the electronic diagram shows the serial communications transceiver implemented in both modules: the control and acquisition, which establishes the communication between the devices, it converts the TTL voltages of the MCU to the RS232 protocol voltages. This transceiver is based on a MAX232ACWE + T (Maxim Integrate, Cal. USA). PIN 11 and 12 are the Tx and Rx signals of the MCU.

Opto-microfluidic structure
The opto-microfluidic was manufactured used the fusion deposition modeling (FDM) technology also called ''3D printing" and Polypropylene PP (TecSol3D, Monterrey, Mex.) was used as material. It is resistant to chemical solvents, bases and acids. The geometry of the opto-microfluidic device was designed in Solidworks Ò , which consisted of: a micromixer with circular 3D serpentine chaotic advection, a sample cell with a volume of 95 lL, four solution inlets and one waste outlet. The connectors for inlet and outlet fluids were made of titanium (MAFSA, Torreón, Mex.) with an internal diameter of 1 mm and an external diameter of 1.5 mm, them were fixed with epoxy glue applied on the surface of the top. The detection cell was integrated in the microfluidic device with an optical path length of 30 mm, and the windows were covered with a 0.12 mm thick borosilicate glass (Corning Ò cover glasses) (Sigma Aldrich, Naucalpan, Mex.) allowing light beam to pass through the cell. The borosilicate glass allows 90% transmittance whenever working at a wavelength between 0.4 and 2.0 lm, and it will not affect the absorbance measurements. The microfluidic device is shown in Fig. 15.

Operation instructions
Once all its parts have been connected, according to Fig. 4, the following analytical considerations and set up operations are carried out. Select the colorimetric analytical method for the measurement of the analyte to be determined. Simplify and miniaturize the analytical method. Develop the timing diagram for injection sequence. Set point: Set the hydrodynamic values for the flow rate and injection volume according to requirements of the process. Set the working wavelength and calibrate the intensity of the radiation source by adjusting TRIM. Select the microvalves to operate according to the injection-timing diagram and feed the hydraulic ports with solutions and reagents. Run the virtual program for the automation process. Configuration. Automation.
For this application case, fluoride was determined.

Analytical method
According to the Mexican standard NMX-AA-077-SCFI-2001 for the measurement of fluorides, the SPANDS method was used based on the reaction between fluoride ions and the colorful complex of Zirconyl-SPADNS. As the fluoride concentration increases, the intensity of the color decreases. Therefore, the absorbance is inversely proportional to the concentration of fluorides. The reaction was carried out in an acid medium. The stock solution was based on NaF 100 mg F -L -1 , as reagent ZrCl 2 Á8H 2 O (Zirconium chloride octahydrate) with concentrated HCl, and SPANDS solution. Fluoride standard solutions were prepared with concentrations of 0.5, 1.0, 1.5, 2.0 and 2.5 mg L -1 . The colored solution from the reaction between fluoride ion and zirconyl acid-SPANDS reagent was measured at 570 nm. The single light module was selected because of the wavelength.
Analytical measurements were evaluated in terms of absorbance, which can be described as [27] A where S a is the analytical signal for the sample, S r is the reference signal (bi-distilled water) and S d is the noise or dark current signal.

Set point
Before carrying out an analysis, it is necessary to perform a characterization of the modules that consists of setting operating parameters in order not to allow some variables to change that could alter the analytical measurement. a) Optical detection module: For the detection system, the intensity of the light source (LED) was adjusted to 75 mW by a resistive adjustment TRIM system, obtaining an optimal value at 30 X and a maximum operating current of 50 mA. b) Development of hydrodynamic variables: To simplify the operation, the micropumps are set at the same flow rate at Q = 3.9 mL min À1 , (F = 13 Hz) in order to control the injection volume. If the flow rate is known, which represents volume per unit of time (speed) in the microchannel area, it is possible to define a desired injection volume. The injec- tion operation of reagents and standard solutions was based the on-off time operation. Therefore, a time sequence for the valve operations was programed and then it was implemented in time function. The multi-commutation time for the injection was set at 100 ms.

Auto-calibration processes
The auto-calibration process develops a series of operations to establish the relationship between the response of the microanalyzer and fluoride concentration (F), based on standard solutions, which leads us to define the operating range in which the microanalyzer will be able to perform measurements to determine the concentration of fluoride in a real sample.
The analysis involves the injection of two reagents, the standard solution and bidistilled water. According to the proposed method (miniaturized and simplified), the ratio of sample to reagents was 5:1.
The procedure for calibration consisted in the injection of a fluoride standard solution + bi-distilled water, selecting the microactuators P1, P2, V1. This process was executed by multi-commutation of V1, developed 0.5, 1.0, 1.5, 2.0 and 2.5 mg L -1 concentrations of standard solutions.
The SPANDS solution and the Zirconium + HCl reagent were injected selecting the microactuators P3, P4, V2. Then, the sample was injected at a 5:1 ratio between sample/SPAND + reactive in acid medium, generating volumes in the order of 500 lL/100 lL respectively. The timing sequence for the automated micro-analytical procedure is shown in Fig. 16 were t 0 represents the time to generate the reference line or base line in V2 and t m is the sampling time in V1.
The switching times of the microvalves for the auto-calibration and sample injection process are fixed in a virtual interface as shown in Fig. 17.
For this microanalyzer, the opto-microfluidic has a total volume of V = 1.63 mL, with a total length to the detection cell of It is important to consider that the volume of the sample and the flow rate must be adjusted to obtain the best signal from the ''FIAgram", see in Fig. 18. The optimal value for the best FIAgram is shown in Fig. 18 C).
This program, could also allow the use of other microfluidic structures with different geometry and dimensions for greater robustness and scaling to the microanalyzer.

Automated analysis
A control panel was designed for the automated processes based on a virtual instrument program. The design of this virtual instrument is made up of two control windows, where all the necessary operations for the development of the auto-  calibration of the microanalyzer were considered. This design was based on a sequential state or event machine architecture, correlating each of them through structures.

1-Set up configuration window
The CONFIGURATION window is where the actions for the configuration of the hydrodynamic system, the detection system and the serial communication port are defined. The section VALVES, allows the user to turn the valves on and off and activate only those that intervene in the process. The section PUMPS, with the virtual knob FREQUENCY, allows the user to select the micro-pumps flow rate. The section DETECTOR, allows the user to select the filter to use and the scaling of the signal of the detection module.
In the section called Port, the user selects the serial communication port to activate the wireless module and the PC is linking with the electronic control module.  By pressing the Set button once the configuration is done, the user fixes these configuration values in the memory. Pressing the ACQUISITION button, the user can see the acquisition data in a waveform graph and the TextBox in a voltaje/time function, according to the parameters selected. If the user no longer wishes to acquire this signal, the ACQUISITION button will have to be pressed again. This process must be done before the microanalysis is carried out.
This configuration window is shown in Fig. 19.

2-Automation analytical processes window
In the frontal panel (AUTOMATION window) there are controls to program or adjust each one of the variables that are necessary to carry out the calibration autonomously, see Fig. 20. These controls are located in a specific order to be activated sequentially one after the other.
The START button, allows the user to start the automation of the analytical process by activating the valves involved in the development of the baseline or transport flow and start the acquisition of the signal generated by the detection system (Sr) from ec. 1. It also resets the injection time COUNTER to zero. With the ENABLE INJECTION button, the user activates the sample injection process, switching the valves involved. With the TIME INJECTION control, the user programs the injection time in seconds, according to the miniaturized and simplified method proposed. Finally, the user executes the sample injection process with the boolean control INJECTION during the previously defined time and this is displayed in a numeric indicator COUNTER and acquired the analytical signal, (Sa). Two waveform graphs are included in the frontal panel, where the user can observe the analytical signal generated in real time (graph above) and the filtered one (graph below). It is also possible to save the data in a text file by activating the SAVE W button for later processing. The user can turn off the valves with the SHUT DOWN VALVES control after performing the calibration.

Validation and characterization
In auto-calibration processes, standard solutions with fluoride concentration ranging from 0.5 to 2.5 mg L À1 were developed. A linear response was obsereved. The equation that describes this linear behavior is A = À0.0473 [F] + 0.2052; R 2 = 0.9946, a sensitivity of 0.00473 mg L -1 a wavelength of 570 nm. For a concentration of 1.5 mg L -1 , the relative standard deviation RSD calculated was 0.07% (n = 4; 95% confidence). In this application, the detection limit was estimated as 0.5 mg L -1 (RSD (%) = 0.07, (n = 4; 95% confidence)). The result was compared with those obtained with a spectrophotometer. Fig. 21 depicts both results, which were plotted against each other. A linear behavior can be observed with a slope close to the unit (0.9997). This confirms traceability between both measurements.  Three groundwater samples with fluoride concentration were automatically analyzed. The samples were taken in deep wells for drinking water in the Lagunera region, which were provided by federal depended National Water Comission, Mexico (CNA). The absorbance measured by the microanalyzer was compared with measurements made previously using a    Table 2. The system response for fluoride in groundwater samples show high accuracy and reproducibility comparable to that obtained with a commercial instrument. By integrating a hydrodynamic system based on miniaturized actuators for fluid management, the consumption of reagents is significantly reduced, as is the generation of waste. The automation of the injection process controls the amounts of reagents by injecting only the necessary volume, this is because the delivered volume depends on the on-off time of the valve, which can be controlled and measured. The auto-calibration process reduced the consumption of reagents by 50 times compared to the consumption using a peristaltic pump and a modular valve positioner, as shown in Table 3.
According to Table 4, the microanalyzer has a consumption of 183.5 mA in standby and 750 mA in operation. Sampling time is 30 s, the battery can perform 330 samples with a frequency of 2 samples per hour. The hydrodynamic system can operate for 7 continuous days.

Discussion
The miniaturization and simplification of the analytical method must be carried out as an analysis prior to the operation of the microanalyzer. The values of simple volume and injection time are obtained from the dimensions of the opto-microfluidic structure to determine the mass transport and mixing time through the micro-channels towards the detection cell.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.