Multi-parametric polymer-based potentiometric analytical microsystem for future manned space missions

Accepted Manuscript Multi-parametric polymer-based potentiometric analytical microsystem for future manned space missions Antonio Calvo-López, Mar Puyol, Joan Manel Casalta, Julián Alonso-Chamarro PII:

S0003-2670(17)31043-7

DOI:

10.1016/j.aca.2017.08.043

Reference:

ACA 235419

To appear in:

Analytica Chimica Acta

Received Date: 20 April 2017 Revised Date:

1 August 2017

Accepted Date: 28 August 2017

Please cite this article as: A. Calvo-López, M. Puyol, J.M. Casalta, J. Alonso-Chamarro, Multi-parametric polymer-based potentiometric analytical microsystem for future manned space missions, Analytica Chimica Acta (2017), doi: 10.1016/j.aca.2017.08.043. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

SC

RI PT

ACCEPTED MANUSCRIPT

Microanalyzer

Drinking or hygiene water

2 40

TE D

M AN U

Sample from water recycling process

2 20

180

Water monitoring

Potential (mV)

2 00 1 80 1 60 1 40 1 20 1 00 20 00

40 00

T im e (s)

AC C

0

EP

Potential (mV)

2 60

6 00 0

8 00 0

220

60

200

40

160

Potential (mV)

K+

2 80

Space missions

80

240

3 00

Cl-

140 120

20 0

NO3-

-20 -40

100

-60

80 0

2000

4000 T im e (s)

6000

8000

0

2000

4000 T im e (s)

6000

8000

ACCEPTED MANUSCRIPT

Multi-parametric

polymer-based

potentiometric

analytical

microsystem for future manned space missions Antonio Calvo-Lópeza, Mar Puyola, Joan Manel Casaltab and Julián Alonso-

a

Group of Sensors and Biosensors, Department of Chemistry, Autonomous University of

Barcelona, Edifici Cn, 08193 Barcelona, Spain

SENER, c/ Creu Casas i Sicart, 86-88, Parc de l’Alba, 08290 Cerdanyola del Vallès, Spain

SC

b

RI PT

Chamarroa,*

*Corresponding author. E-mail address: [email protected]; Tel: +34935812149

M AN U

Abstract

The construction and evaluation of a Cyclic Olefin Copolymer (COC)-based continuous flow potentiometric microanalyzer to simultaneously monitor potassium, chloride and nitrate ions in samples from an on-board water recycling process expected to be installed in future manned

TE D

space missions is presented. The main goals accomplished in this work address the specific required characteristics for a miniaturized on-line monitoring system to control water quality in such missions. To begin with, the integration of three ion-selective electrodes (ISEs) and a

EP

reference electrode in a compact microfluidic platform that incorporates a simple automatic autocalibration process allows obtaining information about the concentration of the three ions

AC C

with optimal analytical response characteristics, but moreover with low reagents consumption and therefore with few waste generation, which is critical for this specific application. By a simple signal processing (signal removal) the chloride ion interference on the nitrate electrode response can be eliminated. Furthermore, all fluidics management is performed by computercontrolled microvalves and micropumps, so no manual intervention of the crew is necessary. The analytical features provided by the microsystem after the optimization process were a linear range from 6.3 to 630 mg L-1 and a detection limit of 0.51 mg L-1 for the potassium electrode, a linear range from 10 to 1000 mg L-1 and a detection limit of 1.58 mg L-1 for the

1

ACCEPTED MANUSCRIPT chloride electrode and a linear range from 10 to 1000 mg L-1 and a detection limit of 3.37 mg L-1 for the nitrate electrode with a reproducibility (RSD) of 4%, 2% and 3% respectively. Sample throughput was 12 h-1 with a reagent consumptions lower than 2 milliliters per analysis.

Keywords: Lab on a chip, Polymer technology, Miniaturization, Multi-parametric,

RI PT

Potentiometric detection, Space applications.

1. Introduction

In order to overcome the problem of water supply in long term manned space missions [1-3],

SC

water recycling systems are being developed by different space agencies, such as the European Space Agency (ESA), the Russian Federal Space Agency (ROSCOSMOS) and the National

M AN U

Aeronautics and Space Administration (NASA), as it was previously reported [4,5]. Specifically, the water recycling system developed by ESA allows obtaining hygiene water or drinking water from cabin condensate, urine and grey water (waste hygiene water) by means of different processes such as nitrification, nano and ultrafiltrations, reverse osmosis and

TE D

remineralization. There are different analytical parameters according to ESA which must be taken into account in order to verify the proper operation of the different stages of the water recycling process, among them are potassium and nitrate ions [1]. In order to overcome the

EP

restrictions of mass, energy and space related to long term manned space missions, a miniaturized prototype analyzer for the on-line monitoring of those two analytes was presented

AC C

by our investigation group [4] and it was tested to verify that the obtained water meets the requirements of the ESA water quality standards regarding potassium and nitrate concentrations [1]. However, that microsystem had some limitations, such as the interfering effect of chloride ion at high concentration levels of this analyte on the nitrate selective electrode response and the lack of enough miniaturization and automation of the whole continuous flow system because conventional pumping and injection systems were employed, compromising thus their final use in such environments.

2

ACCEPTED MANUSCRIPT Some approaches to address the interference of chloride in the nitrate potentiometric response can be found in the literature. A masking or a precipitation agent [6-8] to eliminate chloride from the solution, such as Ag2SO4, can be added, but in our case, this cannot be performed as the concentration of chloride must be also determined. Other strategies consist in the use of

RI PT

electronic tongues which are based on an array of nonspecific sensors combined with a complex data processing of their signals [9-11]. Finally, there are other reported works describing continuous flow systems with the integration of a chloride selective electrode that allow

correcting the interference on the nitrate electrode by means of a simple signal processing using

SC

the Nikolski-Eisenmann equation [12,13]. Thus, with either of these two types of proposals, a small size triparametric microsystem could be developed matching the features imposed by

M AN U

ESA. As the ESA intends to develop a multi-parametric analyzer to determine all the ions of interest related to water quality [1], this triparametric microanalyzer would be the first prototype approaching such a goal.

On the other hand, in order to improve the miniaturization and automation of the whole

TE D

experimental set-up, solenoid diaphragm micropumps can be used instead of the conventional peristaltic pumps [14]. Since the first reported use in 1996 of such micropumps by Weeks and Johnson, numerous works using computer-controlled miniaturized fluid management elements

EP

appeared using continuous flow techniques like multipumping and multicommutation systems [15-18]. These systems are characterized by the use of only solenoid micropumps or in

AC C

combination with also microvalves. Some of their advantages rely on smaller and lighter equipment, less maintenance, low sample and reagent consumption, low power consumption, easy automation, efficient mixing of the sample and reagents due to the pulsating flow, among others [16,18]. However, drawbacks arise in fact due to the pulsating flow provided by the solenoid micropumps, preventing to obtain in-line dilutions from an analyte stock solution in order to perform an autocalibration procedure with a wide range of concentrations. In this case, conventional peristaltic pumps must be used to achieve the reproducible dilutions needed to perform automated calibration protocols, especially when a high dilution factor is to be obtained

3

ACCEPTED MANUSCRIPT [19-21]. However, they are bulky components and could compromise their future use for space applications. Regarding the substrate material to fabricate the microsystem, polymer technology and specifically the technology involved with Cyclic Olefin Copolymer (COC), offers different

RI PT

important advantages for the fabrication of these microanalyzers. For instance, the easy and cheap manufacture of hermetically sealed three dimensional microfluidic structures, the

possibility to integrate conductive tracks, the chemical inertness to most acids and alkalis, and a

SC

good mechanical resistance [5,22-25].

In this context, the goal of the present work is to develop a new miniaturized, robust and

M AN U

automated COC-based potentiometric microanalyzer prototype to simultaneously monitor potassium, chloride and nitrate ions, which comprises miniaturized fluid management system consisting of micropumps (solenoid and peristaltic) and microvalves to meet the requirements of space applications. The microfluidic platform integrates microfluidics and the detection system monolithically in a single substrate and it is smaller than a credit card. The detection

TE D

system is based on two ion-selective electrodes, for potassium and nitrate ions, and two screenprinted Ag/AgCl electrodes, one acting as a chloride selective electrode and the other as a reference electrode. The prototype has been applied to analyze synthetic samples with a similar

EP

matrix than the expected real samples that will be obtained in a future water treatment process in

AC C

manned space missions.

2. Experimental

2.1. Reagents and materials

The microanalyzer was fabricated using plaques and foils of COC purchased from Topas Advanced Polymers (Florence, KY, USA) in different grades and thicknesses: Topas 5013 plaques of 1 mm thickness and Topas 8007 foils of 25 µm thickness. The conductive support for the polymeric membrane ISEs consists of a graphite-epoxy composite made of a mixture of graphite powder with a particle size of 50 µm (Merck, Germany), epoxy-resin Araldite-M and a hardener HR (both from Ciba-Geigy). The Ag/AgCl electrodes (the chloride selective electrode 4

ACCEPTED MANUSCRIPT and the reference electrode) were prepared using a screen-printed Ag/AgCl C2030812D3 paste (Gwent, Pontypool, UK). All reagents employed in this work were of analytical grade. All solutions were prepared in double distilled water. Stock solutions containing 630 mg L-1 K+, 1000 mg L-1 Cl- and 1000 mg

RI PT

L-1 NO3-, which were used for the in-line automatic process of calibration, and the standard solutions obtained by manual serial dilution, which were employed to verify the proper

functioning of the automatic calibration process, were prepared from a concentrated stock one

SC

of 0.1 M of KNO3 (Merck, Germany) and 0.1 M of NaCl (Sigma Aldrich, Germany). A 0.1 M KCl auxiliary solution (Sigma Aldrich, Germany) was used to maintain constant the signal

M AN U

provided by the reference electrode. A 0.05 M Na2SO4 solution (Panreac, Spain) adjusted to pH 3 with sulfuric acid (Sigma Aldrich, Germany) was used as a pH and ionic strength conditioning solution.

In order to prepare the potassium sensor membrane, valinomycin, Bis(2-ethylhexyl) sebacate (DOS), potassium tetrakis(4-chlorophenyl)borate, polyvinyl chloride (PVC) and tetrahydrofuran

TE D

(THF), obtained from Fluka, were used. For the preparation of the nitrate sensor membrane, tetraoctylammonium nitrate, Tris(2-ethylhexyl) phosphate (TEHP), polyvinyl chloride (PVC) and tetrahydrofuran (THF), obtained from Fluka, were used.

EP

2.2. Fabrication of the microanalyzer

AC C

The fabrication process regarding COC-based microfluidic platforms is described in detail elsewhere [5,26]. This process is based on the lamination of COC layers and foils with different glass transition temperatures (Tg) that once overlapped define a 3D structuring, and integrate as well as the microfluidics as the detection system. Topas 6013 plaques with Tg=130 ºC were used for the mechanization of the microchannels and detection chambers, and Topas 8007 foils with Tg=75 ºC were used as sealant between layers. Four are the main steps of this fabrication process: prototype design, pattern machining, integration of transducing elements, such as electrodes, and final lamination. CAD software was employed for the prototype design. It

5

ACCEPTED MANUSCRIPT consisted on three layers that, once overlapped, provide the three-dimensional structure of the microsystem (Figure 1). The dimensions of the microanalyzer were 40 x 30 x 3 mm with a total weight of 4 g. Microfluidics inside the microsystem included three liquid inlets (Figure 2). Two of them converge in a T-shape confluence point that mixes the carrier solution, the sample or

RI PT

the standard solutions with a highly concentrated buffered solution used to keep constant the pH and the ionic strength. The resulting mixed stream is carried to the detection chambers, where each analyte is determined by the corresponding selective electrode and finally the flow is

directed to the waste through the outlet. In order to keep the potential of the reference electrode

SC

at a constant value, an auxiliary 0.1 M KCl solution is continuously pumped at 100 µl min-1

through the third inlet port and flows through the channel in which the reference electrode is

M AN U

placed [13].

Holes and microchannels were machined onto the COC plaques by means of a computer numerically controlled (CNC) micromilling machine (Protomat C100/HF, LPKF, Spain). The dimensions of the channels were 0.4 mm wide and 0.25 mm height. The diameter of each

TE D

detection chamber was 3.5 mm. The dead volume of the detection chambers was of 15 µL and the total microsystem dead volume was 70 µL.

The conductive epoxy resin used as a solid inner contact for potassium and nitrate selective

EP

electrodes was prepared by mixing Araldite-M and the hardener HR in a 1:0.4 weight ratio. Then, this mixture was blended with graphite powder in a 1:1 weight ratio. The resulting

AC C

conductive composite was placed in the corresponding bas-reliefs (see Figure 1) machined in the COC plaque and was cured at 40 °C for 24 h. Finally, in order to allow the successful lamination of the plaque, the electrodes were polished until a smooth surface at the same level as the rest of the platform was obtained. The Ag/AgCl electrodes were fabricated depositing the Ag/AgCl paste by the screen-printing technique, using a screen-printer machine (DEK 248, DEK, Spain). Alignment of the different COC layers and final lamination was performed in a thermocompression press (Francisco Camps, Granollers, Spain) at 105 ºC under 6 atm of pressure 6

ACCEPTED MANUSCRIPT using an aluminum support with 4 alignment pins. Thus, the microsystem was completely sealed and became a hermetic monolithic substrate. Finally, the fluidic connectors were fixed onto the polymer inlet/outlet ports with a holder and screws. In order to integrate the electrodes, the potassium selective polymeric membrane was prepared

RI PT

following the previously optimized composition [4,27] by weighing out and mixing 1% valinomycin, 65.5% DOS, 0.5% Potassium tetrakis(4-chlorophenyl)borate, 33% PVC and 3 mL THF. Likewise, the nitrate selective polymeric membrane was prepared [4,28] by weighing out

SC

and mixing 6% tetraoctylammonium nitrate, 65% TEHP, 29% PVC and 3 mL THF. Both membrane cocktails were deposited dropwise inside their corresponding cavity, which is

M AN U

defined over the epoxy-graphite composite and below the microchannels by following the next optimized protocol: 2 µL of membrane cocktail were added and let evaporate for five minutes. This was repeated until the cavities of the membranes were filled, avoiding trapping of bubbles during the PVC membrane formation due to the THF evaporation. Finally, detection chambers were sealed with an adhesive film.

TE D

2.3. Experimental setup

Figure 2 shows the continuous flow system setup. It consists of two solenoid micro-pumps one

EP

of 4 µL per pulse (Figure 2, SP1) (P/N 030SP124-4TV, BiochemValve Inc., Montluçon Cedex, France) and another one of 10 µL per pulse (Figure 2, SP2) (P/N 120SP1210-5TP,

AC C

BiochemValve Inc., Montluçon Cedex, France), one peristaltic micro-pump (Kamoer KPS10DHC0, Shanghai, China) with a customized rpm controller (TMI, Barcelona, Spain) using Tygon® tubing (Ismatec, Wertheim, Germany) with 0.19 mm internal diameter, and three threeway solenoid valves (161T031, NResearch, Switzerland). Teflon tubing (Scharlab, S.L., Cambridge, England) of 0.8 mm internal diameter was used to connect the different flow elements to the microsystem. A controller of the fluid management devices (Flowtest™, Biotray, France) with its corresponding CosDesigner™ software was used in order to program the multicommutated calibration protocol and to automate the whole analytical procedure without user interaction. Signal acquisition and data processing was performed by means of a 7

ACCEPTED MANUSCRIPT customized potentiometer (6016 4-electrodes, TMI, Barcelona, Spain). As well the fluids management set-up as the signal processing from the different ISEs were controlled by a personal computer.

3.1. Design and optimization of the device

RI PT

3. Results and discussion

The main goal of this work was the development of an automated, small sized, simple, selective and robust microsystem for on-line monitoring of potassium, nitrate and chloride in water

SC

samples, which fulfills the restrictions imposed by the proposed aerospace application in terms of volume, mass and waste generation, and solves the well-known interference of chloride ion

M AN U

into the response of the nitrate ISE. Our approach, through the integration of that third chloride selective electrode, was to correct the interfering effect on the signal of the nitrate electrode at the same time as chloride ion was determined. In addition, a miniaturized computer-controlled fluid management set-up was implemented in order to reduce the whole procedure and limit the

TE D

required volume of sample and reagents.

The composition of the polymeric membrane selective electrodes (for potassium and nitrate ion) and the configuration of the detection chamber were optimized in a previous work [4] in order

EP

to obtain the best performance in terms of sensitivity, baseline stability and detection limits. In this work, smarter microfluidics have been designed and optimized to minimize sample and

AC C

reagent consumption, reducing the total dead volume by more than 50% compared to previously developed microanalyzer [4]. Moreover, a drastic reduction of 67% of the weight of the microsystem was achieved using COC instead of ceramics, with the advantage of being also less mechanically fragile. All these features address the ESA requirements for the proposed microsystems. Taking advantage of the acquired experience regarding the compatibility between Ag/AgCl inks and the COC substrate that were previously reported [5], and the subsequent optimization of its integration as a reference electrode, a Ag/AgCl electrode was integrated, evaluated and

8

ACCEPTED MANUSCRIPT optimized for the construction of the chloride selective electrode. In this sense, different calibration studies were performed in batch conditions with chloride selective electrodes constructed by screen-printing this conductive paste onto COC substrates and the suitability of the Ag/AgCl paste as a chloride selective electrode was demonstrated.

RI PT

Once all the ion selective electrodes were integrated in the microfluidic platform, the influence of chemical and hydrodynamic variables was also evaluated. Sodium sulfate was chosen as

conditioning solution because it provides better response features in terms of sensitivity, peak

SC

height and baseline stability of all the ISEs, as it was expected [4,29,30].

Chemical and hydrodynamic parameters were evaluated using a univariate optimization

M AN U

procedure in order to achieve a compromise between a proper linear working range, expected sensitivity, baseline signal stability, minimum reagent and sample consumption and an adequate sample throughput.

Thus, the flow rate of the carrier and the conditioning solutions was tested from 200 to 1000 µl min-1 (with a fixed 100 µl min-1 flow rate of the 0.1 M KCl auxiliary solution), the sample

TE D

injection volume was varied from 25 to 600 µL and the conditioning solution was evaluated at concentrations ranging from 0.005 to 0.1 M. The optimal results were obtained using a flow rate of 350 µl min-1 for both, the carrier and the conditioning solutions, a sample injection volume of

EP

117 µL and a conditioning solution of Na2SO4 0.05 M set to pH 3. However it is worth mentioning that all these parameters can be modified to address different requirements.

AC C

Regarding the potentiometric selectivity coefficients (Ki,jpot) of potassium and nitrate electrodes, the fixed interference method was performed using a 0.01 M concentration background of the interfering compounds, except for NH4+, which concentration was fixed to 0.001 M [31]. The potentiometric selectivity coefficients of the chloride electrode have not been evaluated because no significant presence of its common interfering ions (I-, Br- and CN-) [32] is expected in the water samples to be analyzed. The obtained results for potassium and nitrate are presented in table 1.

9

ACCEPTED MANUSCRIPT As it can be seen, chloride ion was the main interfering compound for the whole microanalyzer. It can be found at much higher concentration levels than NH4+ in such recycled water samples and it cannot be masked or eliminated from the matrix, such as HCO3- [4]. One can find works, where its elimination by means of precipitation with silver has been proposed [6-8]. But this is

RI PT

not feasible because of two main reasons. On the one hand, chloride is also one of the analytes to be determined in this application and on the other hand, the precipitation could block the

small channels of the microfluidic system. In order to solve this problem, the following linear mathematical expression derived from the Nikolskii-Eisenman equation was applied, taking

SC

profit of the information provided by both selective electrodes, the nitrate and the chloride ones

M AN U

[12,13]:

E = a + b log([NO3-]+ Ki,jpot [Cl-])

(1)

As it can be seen, the expression includes the nitrate concentration and the contribution of the interfering effect of chloride ions on the potential measured by the nitrate selective electrode. In this sense the term Ki,jpot corresponds to the potentiometric selectivity coefficient. Thus, it is

TE D

possible to obtain the corrected nitrate ion concentration when the chloride concentration is also known. The potentiometric selectivity coefficient of chloride over the nitrate selective electrode was determined experimentally as 0.026 ±0.003, so that, the final mathematical expression used

EP

to obtain the nitrate ion concentration was (see Analytical performance section): (2)

AC C

E = a + b log ([NO3-]+ 0.026 ±0.003 [Cl-])

3.2. Flow system

In order to reduce reagent and sample consumption and to automate the whole experimental setup, different pumping elements (solenoid micropumps and a peristaltic micropump) and microvalves were used. All were managed with a controller of fluid manage devices (FlowtestTM). A multicommutation approach was successfully implemented to prepare the standard solutions required for the on-line calibration processes by a sequence of dilutions from a single concentrated stock solution that contained the analytes. A wide range of concentrations

10

ACCEPTED MANUSCRIPT were prepared with good precision and repeatability. Initially a multipumping system, based on solenoid micropumps, was selected due to its simplicity. They can be easily switched on and off by the intermittent application of a voltage. However, the lack of dilution capability, which is limited by the volume of each pulse of the micropumps, and the lack of repeatability of the

RI PT

dilutions prevented their use with reliability. The stock concentrated solution contained the three analytes to be determined to simplify the

process of calibration. It consisted of a mixture of KNO3 and NaCl, in the correct proportions

SC

taking into account that sodium and chloride ions are interfering compounds of potassium and nitrate electrodes respectively, and that the highest chloride concentration to be determined

M AN U

(according to ESA requirements) is 1000 mg L-1. With the aim to have a ratio analyte/interfering compound of 1:1 and avoid any interference (according to Ki,jpot), a mixed stock solution of 630 mg L-1 K+, 657 mg L-1 Na+, 1000 mg L-1 Cl- and 1000 mg L-1 NO3- was prepared. In order to verify that the use of this mixed stock solution does not present a significant interfering effect over the response of the potassium and nitrate electrodes, some tests were carried out. In this

TE D

sense, two calibration experiments were performed using standard solutions containing only KNO3 or standard solutions obtained from the mixed stock solution. Similar peak heights were obtained in both cases, thus demonstrating that the interference effect at this ion concentration

EP

ratio was not significant. With this strategy, the upper limit of the tested ions concentration range was established. Regarding the lower limit, it was dependent on technical parameters such

AC C

as the injection time and the least possible time of actuation of the microvalves. Injection time was 20 seconds, as a result of the optimization of flow rate (350 µL min-1), injection volume (117 µL) and to make the dilution sequence programming easier. The highest commutation speed of the solenoid microvalve was of 100 ms [20]. Thus, the sequence of dilutions that can be made by the multicommutation system was from no dilution (by the injection of the stock solution during the 20 s) to the highest dilution of 3.2 mg L-1 K+, 5 mg L-1 Cl- and 5 mg L-1 NO3- (by the injection of the stock solution during only 0.1s and the remaining 19.99 s of water). The working range determined by the analytical response

11

ACCEPTED MANUSCRIPT of each electrode and repeatability studies of the dilutions are discussed in the next section (Analytical performance). Finally, in order to verify the proper operation of the automated multicommutation approach, hand-made standard solutions and automatically diluted standard solutions were analyzed and

RI PT

compared. The obtained results show that there were no significant differences between both methods.

Figure 3 shows the schematic sequence of each fluid control element in order to fill the whole

SC

system with liquid, inject two standard solutions and inject a sample. As an example (see also Figure 2 for better understanding): step 1 is the filling of stock solution up to valve 1, step 2 the

M AN U

filling of H2O up to valve 2, step 3 the filling of sample up to valve 2 and step 4 the filling of H2O up to valve 3. To fill the microsystem with conditioning solution, H2O and KCl auxiliary solution to stabilize the baseline, step 5 is performed. For instance, to make a standard solution 100 times more diluted than the stock one (6.3 mg L-1 K+, 10 mg L-1 Cl- and 10 mg L-1 NO3-) the following multicommutation sequence is done in step 6 (two cycles of 0.1 s of stock solution

TE D

and 9.9 s of water in the 20 s of injection volume). However, to obtain the highest ion concentration standards, the stock solution is injected during the 20 s without dilution following step 7. Finally, a sample is injected in step 8.

EP

3.3. Analytical performance

AC C

Analytical characterization of the microsystem was carried out by different calibration experiments obtained using an automated dilution sequence from the stock solution. Figure 4 shows the obtained recorded signal for one calibration of each electrode. The obtained Nernst equations (as the average of 3 calibration data in consecutive days with triplicates for each concentration; 95% confidence) were E = 275 (±2) + 56 (± 1) log [K+] with r2=0.9991, E = -227 (±2) - 56 (± 1) log [Cl-] with r2=0.9990 and E = -251.0 (±3) - 59 (± 1) log ([NO3-] + 0.026 (±0.003) [Cl-]) with r2=0.9990. The corresponding working ranges were 6.3-630 mg L-1, 101000 mg L-1 and 10-1000 mg L-1 for the potassium, chloride and the nitrate selective electrodes respectively. Note that these working ranges were limited by the established lower and higher 12

ACCEPTED MANUSCRIPT values of the dilution sequence and do not correspond to the linear response of each electrode. This corresponds to 1-5000 mg L-1, 6-2300 mg L-1 and 10-4000 mg L-1 for the potassium, chloride and the nitrate selective electrodes respectively. Regarding detection limits (n = 3, 95% confidence) and according to IUPAC [33], they were 0.5 ± 0.1 mg L-1 for potassium ion, 2.2 ±

RI PT

0.4 mg L-1 for chloride ion and 3.0 ± 0.7 mg L-1 for nitrate ion. All limits of quantification lay under the minimum concentration allowed by the ESA quality standards in recycled water, which are 12, 200 and 25 mg L-1 respectively [1].

SC

Repeatability studies were performed by successive injections (n=10) of different diluted

standard solutions in order to find the maximum dilution that it is possible to do with a relative

M AN U

error between peak heights lower than 5%. In this sense, standard solutions 50 times more diluted than the stock (12.6 mg L-1 K+, 20 mg L-1 Cl- and 20 mg L-1 NO3-), 100 times (6.3 mg L-1 K+, 10 mg L-1 Cl- and 10 mg L-1 NO3-) and 200 times (3.2 mg L-1 K+, 5 mg L-1 Cl- and 5 mg L-1 NO3-) were automatically prepared by dilution and analyzed. The relative standard deviations of the signals were 6.0% (at 3.2 mg L-1), 2.5% (at 6.3 mg L-1) and 1.2% (at 12.6 mg L-1) for the

TE D

potassium selective electrode, 13.8% (at 5 mg L-1), 3.8% (at 10 mg L-1) and 1.5% (at 20 mg L-1) for the chloride electrode and 3.4% (at 5 mg L-1), 2.5% (at 10 mg L-1) and 1.4% (at 20 mg L-1) for the nitrate selective electrode. As it can be seen, standard solutions diluted 200 times than

EP

the stock show worse repeatability and therefore, the lowest concentration that can be included in the working range is 100 times more diluted than the concentrated stock solution.

AC C

Reproducibility was also tested from five calibration experiments performed along 15 days. Mean slopes of 58, -56 and -61 with RSD values of 4%, 2% and 3% were achieved for potassium, chloride and nitrate electrodes respectively, thereby demonstrating the reproducibility of the whole system. Finally, with the presented microsystem configuration and the stated chemical and hydrodynamic parameters, a sampling rate of 12 samples h-1 was obtained with a total consumption of reagent and sample of 1.8 ml for each analysis. Furthermore, the microsystem lifetime was at least 9 months, obtaining means slopes of -59, 54 and -55 with RSD of 4 %, 9 % 13

ACCEPTED MANUSCRIPT and 2 % for NO3-, K+ and Cl- electrodes respectively, through 15 calibrations along this period of time. These results showed the potentiality and the robustness of the whole experimental setup for the purposed application. Nevertheless, some improvements can be performed in order to reduce the sample and reagent

RI PT

consumption volume. For instance by suppressing the flow of the KCl solution employed to maintain constant the potential of the reference electrode and by decreasing the size of each

reduction of up to 50% the volume of consumables.

3.4. Samples mimicking the final application

SC

detection chamber from 15 µL to 8 µL. These modifications have been tested confirming a

M AN U

Synthetic samples with a similar matrix than the expected real samples, which can be found in the water recycling unit managed by ESA and placed in the Antarctic Concordia test station [4], were analyzed using the developed microsystem for the simultaneous determination of potassium, chloride and nitrate ions. They contain different concentration levels of potassium,

TE D

chloride and nitrate ions besides 0.7 mg L-1 ammonium, 0.8 mg L-1 sodium and 7 mg L-1 sulfate, with the aim to explore the limits and the working range of the microfluidic system, especially when the concentration of chloride is higher than the nitrate ion. Results obtained are shown in

EP

Table 2.

As it can be seen, the results show good agreement between spiked and found amounts of each

AC C

analyte and give recoveries ranging from 96.0% to 103.4%. Moreover, four concentrations are under the working range of their respective electrodes (one sample for potassium, one sample for the chloride and two samples for the nitrate electrode) but are also under the quality standards allowed by the ESA. Therefore, this confirms that the developed analytical microsystem is suitable for the simultaneous determination of potassium, chloride and nitrate ions in water from the recycled system developed by ESA, even when the concentration of chloride ion is higher than the nitrate ion.

Conclusions

14

ACCEPTED MANUSCRIPT In this paper, a prototype of a microanalyzer for the simultaneous determination of potassium, chloride and nitrate ions, based on polymer technology and potentiometric measurements, has been designed, developed, characterized and applied to the analysis of synthetic samples. It allows the measurement of chloride ion and therefore, the possibility to correct its interfering

RI PT

effect on the nitrate electrode by means of a mathematical correction. The analytical features are comparable to those of commercial analytical systems. Furthermore, an instrumentation

compaction has been done, implementing a miniaturized and automated computer-controlled fluid management system, which incorporates micropumps and microvalves. In addition, the

SC

utility of the Ag/AgCl paste, compatible with COC, as substrate of a chloride selective electrode

M AN U

has been demonstrated, besides of as reference electrode, which was previously demonstrated. All these features and the optimization of the microsystem design have allowed the development of a small and light device with a compact and small fluid control instrumentation with low sample and reagent consumption and completely automated, fulfilling the

Acknowledgments

TE D

requirements of ESA for manned space applications.

The authors would like to thank the European Space Agency (ESA) and NTE-SENER for its

EP

financial support through project ‘Preliminary definition of on-line Chemical Water Quality Analysis equipment (Contract No. 4000104044/10/NL/NA)’. This work has been also supported

AC C

by the Spanish Ministry of Science and Innovation (MICINN) through projects CTQ200912128 and CTQ2012-36165, co-funded by FEDER (Fondo Europeo de Desarrollo Regional).

References

[1] C. Tamponnet, C.J. Savage, P. Amblard, J.C. Lasserre, J.C. Personne, J.C. Germain, Water Recovery in Space, ESA Bull.-Eur. Space 97 (1999) 56-60. [2] A.I. Grigoriev, Y.E. Sinyak, N.M. Samsonov, L.S. Bobe, N.N. Protasov, P.O. Andreychuk, Regeneration of water at space stations, Acta Astronaut. 68 (2011) 1567-1573.

15

ACCEPTED MANUSCRIPT [3] H.W. Jones, M.H. Kliss, Exploration life support technology challenges for the crew exploration vehicle and future human missions, Adv. Space Res. 45 (2010) 917-928. [4] A. Calvo-Lopez, E. Arasa-Puig, M. Puyol, J.M. Casalta, J. Alonso-Chamarro, Biparametric potentiometric analytical microsystem for nitrate and potassium monitoring in water recycling

RI PT

processes for manned space missions, Anal. Chim. Acta 804 (2013) 190-196. [5] A. Calvo-Lopez, O.Ymbern, M. Puyol, J.M. Casalta, J. Alonso-Chamarro, Potentiometric

analytical microsystem based on the integration of a gas-diffusion step for on-line ammonium

SC

determination in water recycling processes in manned space missions, Anal. Chim. Acta 874 (2015) 26-32.

M AN U

[6] J. Alonso-Chamarro, J. Bartrolí, S. Jun, J. L. F. C. Lima, M. C. B. S. M. Montenegro, Sequential determination of calcium and nitrate ions in waters by potentiometric flow injection , Analyst 118 (1993) 1527-1532.

[7] E.W. Rice, R.B. Baird, A. D. Eaton, L. S. Clesceri, in Standard Methods for the

Washington DC, 2012.

TE D

Examination of Water and Wastewater, 22nd ed., American Public Health Association,

[8] D. Midgley, K. Torrance, Potentiometric Water Analysis, Wiley, Chichester 1990.

EP

[9] M. Gutiérrez, S. Alegret, R. Cáceres, J. Casadesús, O. Marfà, M. del Valle, Nutrient solution monitoring in greenhouse cultivation employing a potentiometric electronic tongue, J.

AC C

Agric. Food Chem., 56 (2008) 1810-1817. [10] L. T. Dimitrakopoulos, T. Dimitrakopoulos, Evaluation of a four sensor array used in a wall-jet configured flow cell for flow injection potentiometry, Electroanal., 13 (2001) 161-163. [11] M.L. Rodriguez-Mendez, Electronic Nose and Tongue in Food Science, Academic Press, 2016, 316 pp. [12] J. Janata, Principles of Chemical Sensors, 2nd Edition, Springer, 2009, 340 pp.

16

ACCEPTED MANUSCRIPT [13] N. Ibáñez-García, M. Baeza, M. Puyol, R. Gómez, M. Batlle, J. Alonso-Chamarro, Biparametric Potentiometric Analytical Microsystem Based on the Green Tape Technology, Electroanal. 22 (2010) 2376-2382. [14] D.A. Weeks, K.S. Johnson, Solenoid Pumps for Flow Injection Analysis, Anal. Chem. 68

RI PT

(1996) 2717-2719. [15] D.L. Rocha, F.R.P. Rocha, A flow-based procedure with solenoid micro-pumps for the spectrophotometric determination of uric acid in urine, Microchem. J., 94 (2010) 53-59.

SC

[16] C.Henríquez, B. Horstkotte, V. Cerdà, A highly reproducible solenoid micropump system for the analysis of total inorganic carbon and ammonium using gas-diffusion with

M AN U

conductimetric detection, Talanta 118 (2014) 186-194.

[17] G.C.S. de Souza, P.A.B. da Silva, D.M.S. Leotério, A.P.S. Paim, A.F. Lavorante, A multicommuted flow system for fast screening/sequential spectrophotometric determination of dichromate, salicylic acid, hydrogen peroxide and starch in milk samples, Food Control 46

TE D

(2014) 127-135.

[18] M.F. de Andrade, S.G.F. de Assis, A.P.S. Paim, B.F. dos Reis, Multicommuted flow analysis procedure for total polyphenols determination in wines employing chemiluminescence

EP

detection, Food Anal. Method., 7 (2014) 967-976. [19] J.F. Ventura-Gayete, S. Armenta, S. Garrigues, A. Morales-Rubio, M. de la Guardia,

AC C

Multicommutation-NIR determination of Hexythiazox in pesticide formulations, Talanta 68 (2006) 1700-1706.

[20] Z.M. da Rocha, C.S. Martinez-Cisneros, A.C. Seabra, F. Valdés, M.R. Gongora-Rubio, J. Alonso-Chamarro, Compact and autonomous multiwavelength microanalyzer for in-line and in situ colorimetric determinations, Lab Chip 12 (2012) 109-117. [21] O.D. Pessoa-Neto, V.B. dos Santos, F.C. Vicentini, W.T. Suarez, J. Alonso-Chamarro, O. Fatibello-Filho, R.C. Faria, A low-cost automated flow analyzer based on low temperature co-

17

ACCEPTED MANUSCRIPT fired ceramic and LED photometer for ascorbic acid determination, Cent. Eur. J. Chem., 12 (2014) 341-347. [22] H. Becker, C. Gärtner, Polymer microfabrication technologies for microfluidic systems, Anal. Bioanal. Chem. 390 (2008) 89-111.

RI PT

[23] C.G. Koh, W. Tan, M. Zhao, A.J. Ricco, Z.H. Fan, Integrating polymerase chain reaction, valving, and electrophoresis in a plastic device for bacterial detection, Anal. Chem. 75 (2003) 4591-4598.

monitoring systems, Talanta 56 (2002) 355-363.

SC

[24] M. Sequeira, M. Bowden, E. Minogue, D. Diamond, Towards autonomous environmental

M AN U

[25] D. Lee, H. Yang, K. Chung, H. Pyo, Wafer-scale fabrication of polymer-based microdevices via injection molding and photolithographic micropatterning protocols, Anal. Chem. 77 (2005) 5414-5420.

[26] O. Ymberm, N. Sández, A. Calvo-López, M. Puyol, J. Alonso, Gas diffusion as a new

TE D

fluidic unit operation for centrifugal microfluidic platforms, Lab Chip 14 (2014) 1014-122. [27] S.G. Lemos, A.R.A. Nogueira, A. Torre-Neto, A. Parra, J. Artigas, J. Alonso, In-soil potassium sensor system, J. Agr. Food Chem. 52 (2004) 5810-5815.

EP

[28] A. Parra, Thesis dissertation, Universitat Autònoma de Barcelona, 2006.

AC C

[29] C.A.B. Garcia, L.R. Júnior, G.O. Neto, Determination of potassium ions in pharmaceutical samples by FIA using a potentiometric electrode based on ionophore nonactin occluded in EVA membrane, J. Pharmaceut. Biomed. 31 (2003) 11-18. [30] N. Zárate, E. Irazu, A.N. Araújo, M.C.B.S.M. Montenegro, R. Pérez-Olmos, Simultaneous determination of potassium and nitrate ions in mouthwashes using sequential injection analysis with potentiometric detection, Anal. Sci. 24 (2008) 803-807. [31] Y. Umezawa, K. Umezawa, H. Sato, Selectivity coefficients for ion-selective electrodes: Recommended methods for reporting KA,Bpot values, Pure Appl. Chem. 67 (1995) 507-518.

18

ACCEPTED MANUSCRIPT [32] M. Jin, J. Xu, L. Jiang, Y. Xu, H. Chu, Investigation on the performance characteristics of chloride selective electrode in concrete, Ionics 21 (2015) 2981-2992. [33] E. Lindner, Y. Umezawa, Performance evaluation criteria for preparation and measurement of macro- and microfabricated ion-selective electrodes (IUPAC Technical Report), Pure Appl.

M AN U

SC

RI PT

Chem. 80 (2008) 85-104.

Figure 1: Picture of the developed device; a) Fluidic connections; b) Microfluidics; c) Detection chambers with the ion selective electrodes; d) Reference electrode; e) Electrical

AC C

EP

TE D

connectors.

Figure 2: Scheme of the experimental setup, where: a) KCl 0.1M auxiliary solution; b) 0.05 M Na2SO4 conditioning solution at pH 3; c) Sample; d) H2O; e) Stock solution; W: Waste outlet; Vx: three-way solenoid microvalve; PP: peristaltic micropump; SPx: solenoid micropump; C: Controller for fluid control devices; 1) Potassium selective electrode; 2) Chloride selective electrode; 3) Nitrate selective electrode; 4) Reference electrode. Red arrows show the flow direction.

19

ACCEPTED MANUSCRIPT

V2 V3 SP1 SP2 PP

on off on off on off on off

on off on off

step 2

step 3

step 4

step 5

step 6

step 7

step 8

SC

step 1

RI PT

V1

Figure 3. Schematic sequence of each fluid control elements where: on) 12 V are applied to

M AN U

switch each one on; off) 0 V are applied to switch each one off; Vx: three-way solenoid microvalve; SPx: solenoid micropump; PP: peristaltic micropump. Note that the duration of

AC C

EP

TE D

each step is not necessarily the same.

20

300 280

Potential (mV)

260 240 220 200

200 180 160 140 120 100 80 60 40 20 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 d log[K+]

h g

f e

c

180

b a

160 140 120 100 0

2000

4000

6000

Time (s)

220

a b

M AN U

c

180

d

160

Peak Height (mV)

Potential (mV)

200

140 120 100 80

0 -20 -40 -60 -80 -100 -120 -140 -160 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 log[Cl-]

40 20

f

g

4000

h

6000

8000

Time (s)

b

c

Peak Height (mV)

0

-20

AC C

e

a

EP

Potential (mV)

60

2000

TE D

0

C 80

8000

SC

B 240

RI PT

A

Peak Height (mV)

ACCEPTED MANUSCRIPT

-40 -60

0

d 0 -20 -40 -60 -80 -100 -120 -140 -160 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 log[NO3-]

2000

e

4000

f g h

6000

8000

Time (s)

Figure 4. Signal recordings and obtained calibration curves for the microanalyzer calibration using a multicommutation dilution program from a standard solution of 630 mg L-1 K+, 1000 mg L-1 Cl- and 1000 mg L-1 NO3-. (A) Potassium electrode, K+ dilutions of 3.15 mg L-1 (a), 6.3 mg L-1 (b), 12.6 mg L-1 (c), 31.5 mg L-1 (d), 63 mg L-1 (e), 126 mg L-1 (f), 315 mg L-1 (g) and 630 mg L-1 (h). (B) Chloride electrode, Cl- dilutions of 5 mg L-1 (a), 10 mg L-1 (b), 20 mg L-1 (c), 50 21

ACCEPTED MANUSCRIPT mg L-1 (d), 100 mg L-1 (e), 200 mg L-1 (f), 500 mg L-1 (g) and 1000 mg L-1 (h). (C) Nitrate electrode NO3-, dilutions of 5 mg L-1 (a), 10 mg L-1 (b), 20 mg L-1 (c), 50 mg L-1 (d), 100 mg L-1 (e), 200 mg L-1 (f), 500 mg L-1 (g) and 1000 mg L-1 (h). Table 1: Potentiometric selectivity coefficients, logKi,jpot, for K+ and NO3- selective electrodes in

RI PT

treated water samples using the fixed interference method.

logKi,jpot

Main interfering compound (j) for NO3- selective electrode

logKi,jpot

Na+

-3.50

HCO3-

-2.22

NH 4+

-1.24

Cl-

Ca2+

-3.95

Mg2+

-3.92

-1.59

M AN U

SC

Main interfering compound (j) for K+ selective electrode

Table 2. Concentration mean values in mg L-1 (n=3, 95%) from the analysis of water samples

Potassium

TE D

using the proposed microsystem.

Chloride

Sample

Nitrate

% Rec1

Added

Found

% Rec

Added

Found

% Rec

100.1

14.3

14.5 ± 0.2

101.3

25

25.8 ± 0.9

103.1

102.8

0*

-

-

10

10.3 ± 0.5

103.4

32.3 ± 2.6

100.9

264

262.6 ± 2.9

99.4

25

24.4 ± 1.3

97.5

63

62.5 ± 1.1

99.2

28.6

28.5 ± 1.2

99.7

50

49.8 ± 2.3

99.6

10.2

10.3 ± 0.2

100.8

18.2

18.3 ± 0.5

100.7

13.3

13.5 ± 1.2

101.3

4*

-

-

18.2

17.9 ± 0.4

98.1

2.9*

-

-

7

6.9 ± 0.2

98.9

18.2

18.1 ± 0.8

99.5

7.8*

-

-

8

9.6

9.5 ± 0.7

98.7

188.2

185.7 ± 3.1

98.7

11.8

11.5 ± 0.9

97.1

9

15.8

15.6 ± 0.5

98.8

100

97.5 ± 5.2

97.5

25

24.2 ± 1.2

96.9

10

12.6

12.1 ± 0.7

96.0

200

196.2 ± 6.6

98.1

20

19.6 ± 1.5

97.8

630

640.5 ± 11.4

101.7

1000

1012.0 ± 17.4

101.2

1000

1010.2 ± 18.4

101.0

Found

1

32

32.0 ± 1.0

2

6.3

6.5 ± 0.4

3

32

4

AC C

EP

Added

5 6 7

11 1

% Recovery.*Values below the working range of the ISE, under the fixed optimized

operational conditions. 22

ACCEPTED MANUSCRIPT Table 1: Potentiometric selectivity coefficients, logKi,jpot, for K+ and NO3- selective electrodes in treated water samples using the fixed interference method.

logKi,jpot

Main interfering compound (j) for NO3- selective electrode

logKi,jpot

Na+

-3.50

HCO3-

-2.22

NH 4+

-1.24

Cl-

Ca2+

-3.95

Mg2+

-3.92

RI PT

Main interfering compound (j) for K+ selective electrode

SC

-1.59

M AN U

Table 2

Concentration mean values in mg L-1 (n=3, 95%) from the analysis of water samples using the proposed microsystem. Potassium

Chloride

Sample

Nitrate

Found

% Rec1

Added

Found

% Rec

Added

Found

% Rec

1

32

32.0 ± 1.0

100.1

14.3

14.5 ± 0.2

101.3

25

25.8 ± 0.9

103.1

2

6.3

6.5 ± 0.4

102.8

0*

-

-

10

10.3 ± 0.5

103.4

3

32

32.3 ± 2.6

100.9

264

262.6 ± 2.9

99.4

25

24.4 ± 1.3

97.5

4

63

62.5 ± 1.1

99.2

28.6

28.5 ± 1.2

99.7

50

49.8 ± 2.3

99.6

5

10.2

10.3 ± 0.2

100.8

18.2

18.3 ± 0.5

100.7

13.3

13.5 ± 1.2

101.3

6

4*

-

-

18.2

17.9 ± 0.4

98.1

2.9*

-

-

EP

AC C

7

TE D

Added

8 9 10 11 1

7

6.9 ± 0.2

98.9

18.2

18.1 ± 0.8

99.5

7.8*

-

-

9.6

9.5 ± 0.7

98.7

188.2

185.7 ± 3.1

98.7

11.8

11.5 ± 0.9

97.1

15.8

15.6 ± 0.5

98.8

100

97.5 ± 5.2

97.5

25

24.2 ± 1.2

96.9

12.6

12.1 ± 0.7

96.0

200

196.2 ± 6.6

98.1

20

19.6 ± 1.5

97.8

630

640.5 ± 11.4

101.7

1000

1012.0 ± 17.4

101.2

1000

1010.2 ± 18.4

101.0

% Recovery.*Values below the working range of the ISE, under the fixed optimized

operational conditions.

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

a

b

e

d c e

SP1

b

SP2 W

V2

c

PP

V3

TE D

C

EP

e

V1

AC C

d

W

M AN U

a

SC

RI PT

ACCEPTED MANUSCRIPT

4

2 1

3

SP2 PP

SC

off

M AN U

SP1

on

on off on off

TE D

V3

off

on off on

EP

V2

on

off

step 1

AC C

V1

RI PT

ACCEPTED MANUSCRIPT

step 2

step 3

step 4

step 5

step 6

step 7

step 8

300 280

Potential (mV)

260 240 220 200

200 180 160 140 120 100 80 60 40 20 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 d log[K+]

h g

f e

c

180

RI PT

A

Peak Height (mV)

ACCEPTED MANUSCRIPT

b a

160 140 120

0

2000

SC

100 4000

6000

B 240 220

a b c

180

d

120 100 80

0

C 80

f

2000

g

h

4000

6000

8000

Time (s)

AC C

a

60

b

40

c

20 0

Peak Height (mV)

Potential (mV)

e

TE D

140

0 -20 -40 -60 -80 -100 -120 -140 -160 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 log[Cl-]

EP

160

Peak Height (mV)

Potential (mV)

200

M AN U

Time (s)

8000

-20 -40 -60 0

d 0 -20 -40 -60 -80 -100 -120 -140 -160 -4.5 -4.0 -3.5 -3.0 -2.5 -2.0 -1.5 log[NO3-]

2000

e

4000 Time (s)

f g h

6000

8000

ACCEPTED MANUSCRIPT Prototype of a life support system for human spaceflight missions. Autonomous and automatic chemical sensing in water recycling processes. Microfluidic platform based on polymer technology for simultaneous potentiometric

AC C

EP

TE D

M AN U

SC

Synthetic samples were successfully analyzed.

RI PT

determination of chloride, nitrate and potassium ions.