Inorganic salt modified paper substrates utilized in paper based microfluidic sampling for potentiometric determination of heavy metals

Accepted Manuscript Title: Inorganic salt modified paper substrates utilized in paper based microfluidic sampling for potentiometric determination of heavy metals Author: Ruiyu Ding Vida Krikstolaityte Grzegorz Lisak PII: DOI: Reference:

S0925-4005(19)30436-8 https://doi.org/doi:10.1016/j.snb.2019.03.079 SNB 26298

To appear in:

Sensors and Actuators B

Received date: Revised date: Accepted date:

20 November 2018 2 February 2019 17 March 2019

Please cite this article as: R. Ding, V. Krikstolaityte, G. Lisak, Inorganic salt modified paper substrates utilized in paper based microfluidic sampling for potentiometric determination of heavy metals, Sensors and Actuators B: Chemical (2019), https://doi.org/10.1016/j.snb.2019.03.079 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.

Inorganic salt modified paper substrates utilized in paper based microfluidic sampling for potentiometric determination of heavy

ip t

metals

College of Engineering, School of Civil and Environmental Engineering, Nanyang

us

1

cr

Ruiyu Ding1, Vida Krikstolaityte1,2, Grzegorz Lisak1,2*

Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore Nanyang Environment and Water Research Institute, Residues and Resource Reclamation

an

2

Center, 1 Cleantech Loop, CleanTech, Singapore 637141, Singapore

M

*Corresponding author: [email protected]

d

Abstract

te

Inorganic salt modified paper substrates were developed and utilized in microfluidic paper based sampling coupled with potentiometric ion sensors for the determination of heavy

Ac ce p

metals, such as Cd2+ and Pb2+. The application of paper based sampling without any paper substrate pre-treatment were characterized with super Nernstian response, while the paper substrates with inorganic salt pre-treatment were characterized with standard Nernstian response. The application of inorganic salt modified paper substrates was found advantageous when the cation of the salt in the conditioning solution was the same as the primary cations of the ion-selective electrodes (ISEs). A relatively high concentration of inorganic salts during pre-treatment of the paper substrates, namely Cd(NO3)2 or Pb(NO3)2 with subsequent measurement of cadmium(II) and lead(II), by Cd2+- and Pb2+-ISEs, respectively was needed to sustain Nernstian response of sensors. It was also found that paper substrates facilitate sorption of metal ions onto the paper substrates, with stronger binding strength given to Pb2+ over Cd2+. Moreover, the metal-paper interacts suggest concentration dependent sorptiondesorption of metal ion, which can have direct limitations of the use of paper based analytics for the determination of low analyte concentration of heavy metals.

1 Page 1 of 32

Keywords Potentiometric sensors; Paper based sampling; Inorganic salt modification of paper; Metal

ip t

ions sorption on paper

1. Introduction

cr

The increasing contamination of the environment by heavy metals poses a serious ecological and global public health concerns [1]. In terms of toxicity, the most hazardous heavy metals are

us

lead, cadmium, mercury and arsenic [2, 3]. To control the spread of heavy metals, the Environmental Protection Agency (EPA) established the standards for the maximum allowed concentration level of heavy metals in drinking water, e.g. 10 µg L-1 and 5 µg L-1 for lead and

an

cadmium, respectively [4].

To perform a reliable analysis of heavy metals in aqueous environmental samples, the volume

M

of a sample should range from several to several hundred millilitres, depending on the analytical technique used [5]. For aquatic environmental samples such as lake, river, ground or rain water, the sample volume required for the analysis of heavy metals is usually

d

sufficient to perform reliable determination [6-8]. However, samples with small volume of

te

analyte, e.g. animal and plant food products, moist surfaces, spillages or wet soil do not contain a sufficient volume of liquid to be collected for the analysis. To explore a real time

Ac ce p

monitoring of ion activity in small volume samples, Lewenstam et al. developed a micro potentiometric sensor suitable for the ion activity measurements in 10 to 30 µl sample solutions [9]. In the report, however, the risk of electrodes being mechanically damaged by often rough nature of a sampling surfaces was not addressed. To avoid mechanical damage of electrodes, a substrate to be placed between the sapling surface and the electrodes should be introduced. This can be offered by the application of paper substrates, which can simultaneously serve as a sampling substrate as well as a protective layer for ion-selective membranes (ISM) from mechanical damage [10, 11]. Paper is a relatively cheap substrate that has versatile chemical and physical properties, such as chemical composition, mechanical strength, various pore sizes and thicknesses, wettability and the ability to wick the solution. Moreover, paper substrate can easily be modified and customized for specific applications, thus it has been used in a wide spectrum of analytical

2 Page 2 of 32

diagnostics [12, 13], e.g. platform for electrode printing, coating and integration [14-19], and electrochemical sensing devices [20-28]. Recently, paper-based microfluidic sampling has been explored as alternative option for sampling and sample handling of low volume aqueous samples. The microfluidic paper based

ip t

sampling was coupled with potentiometric sensors, where the sample solution was wicked to the paper matrix and delivered to the sensor surface, where the electrodes were pressed onto

cr

the paper substrate. This allowed to close electric circuit and successfully run the potentiometric measurements of ions wicked by the paper substrate [10, 11, 29]. It was found

us

that measurements of clinically relevant ions, e.g. K+, Na+ or Cl– was not affected by the paper based sampling. To the contrary, measurements of pH, Cd2+ and Pb2+ were heavily interfered by the paper based sampling. It was observed that paper based substrate alters pH

chemical bonding to the cellulose material [11].

an

of weakly buffered samples, while heavy metals undergo either physical adsorption or

M

In this work, crystalline membrane solid-state Cd2+- and a solid-contact Pb2+-ISEs were used to investigate the possibility to facilitate the use paper based microfluidic sampling coupled

d

with potentiometric sensors for the determination of heavy metals. Paper pre-treatment was

Ac ce p

and the paper substrates.

te

applied to reduce the metal ions interaction between the standards and the sample solutions

2. Experimental

2.1. Chemicals, materials and electrodes. Cadmium nitrate, Cd(NO3)2 (purity ≥ 98%), lead nitrate, Pb(NO3)2 (purity ≥ 99%), 70% HNO3 (purity ≥ 99%, trace metal grade), sodium polystyrene sulfonate (NaPSS) (purity ≥ 99%), 3,4-Ethylenedioxythiophene (EDOT) (purity ≥ 99%), lead ionophore IV (purity ≥ 99%), potassium tetrakis(4–chlorophenyl) borate (KTpClPB) (purity ≥ 99%), 2-Nitrophenyl octylether (o-NPOE) (purity ≥ 99%), poly(vinyl chloride), (PVC) (purity ≥ 99%), tetrahydrofuran (THF) (purity ≥ 97%) were obtained from Sigma–Aldrich (Germany). Potassium chloride (KCl) (purity ≥ 99%) was purchased from Merck KGaA (Germany). Sodium chloride (NaCl) (purity ≥ 99%), potassium hexacyanoferrate (III) (purity ≥ 99%) were obtained from VWR International (USA). The quality control standard 21 was obtained from PerkinElmer (USA). Five filter paper of different pore size grades, namely: 388 3 Page 3 of 32

(particle retention, 12-15 µm), further named as PS1; 389 (particle retention, 8-12 µm), further named as PS2; 390 (particle retention, 3-5 µm), further named as PS3; 391(particle retention, 2-3 µm), further named as PS4 and 392 (particle retention, 5-8 µm), further named as PS5 were used as paper based substrates. All filter papers were made out of 100% cotton

ip t

linters with a α-cellulose content of > 98 % and were acid washed to make the papers ashless and achieve its high purity. The filter papers were obtained from Sartorius (Germany). InLab Surface Pro ISM pH electrode (flat glass membrane electrode) coupled with Seven Compact

cr

pH/ion S220 meter was purchased from Mettler-Toledo (Switzerland). A crystalline solidstate membrane cadmium(II)-selective electrode and single junction silver/silver chloride (3

us

M KCl) reference electrode were obtained from Thermo Fisher (USA). Glassy carbon (GC) disk electrodes were purchased from Bioanalytical Systems (USA). All aqueous solutions

an

were prepared using ultra-pure water (18 MΩ cm) obtained using Milli-Q Integral Water

M

Purification System (USA).

2.2. Preparation of electrodes and paper substrates.

d

Before the use, the solid-state Cd2+-ISE was gently polished using 0.05 µm alumina slurry (Al2O3) on a polishing soft pad and then vigorously rinsed with ultra-pure water. The Cd2+-

te

ISE was stored in 10-3 M Cd(NO3)2 solution between the measurements. The solid-contact Pb2+-ISEs were prepared by polishing glassy carbon (GC) disk electrodes using 0.05 µm

Ac ce p

alumina slurry (Al2O3) on a polishing soft pad and vigorously rinsing with ultra-pure water. Then electopolymerisation of PEDOT (PSS) on GC electrodes (working electrodes) was conducted in a three-electrode cell, where a Pt mesh and an Ag/AgCl (3 M KCl) served as counter and reference electrodes, respectively. The electropolymerisation was performed using CHI760E electrochemical workstation, CH Instruments, Inc. (USA) by applying a constant current of 0.014 mA (0.2 mA cm-2) for 714s [30, 31]. The electropolymerisation was performed in the electrolyte containing 0.01 M EDOT and 0.1 M NaPSS. After the electropolymerisation, electrodes were rinsed with ultra-pure water and dried overnight in open air at room temperature (23-25℃). Then the Pb2+- ion selective membranes were dropcasted on dried PEDOT(PSS) layer from a Pb2+ selective membrane cocktail consisting of (w/w %): 1% lead ionophore (IV), 0.5% KTpClPB, 65.2% o-NPOE and 33.3% PVC dissolved in 2 mL THF. The membrane casting was performed in three consecutive, 20 µL Pb2+ selective membrane cocktail portions, with one hour interval between each portion 4 Page 4 of 32

applied. The electrodes were left overnight for evaporation of the residual THF. The prepared Pb2+-ISEs were conditioned in 10-3 M Pb(NO3)2 for 12 h before the use. The Pb2+-ISEs were stored in 10-3 M Pb(NO3)2 solution between measurements. The paper substrates of different grade were cut into pieces of 3 × 3 cm2 and immersed in

ip t

ultra-pure water for 30 min while gently stirring in order to remove any water-soluble components. Then, wet paper substrates were oven-dried at 70°C for 30 min before their

2.3. Physicochemical characterization of paper substrates.

cr

subsequent use in physiochemical characterization and in microfluidic paper-based sampling.

us

To evaluate the liquid absorption capacity, different paper substrates were cut into pieces of 3 × 3 cm2. The mass change before and after soaking the paper substrates in ultra-pure water

an

for 5 s was used to determine the liquid absorbed in each paper substrate. The evaporation rate of solution in each paper substrate was measured by registering the mass change immediately after soaking the paper substrate in ultra-pure water for 5 s and after 1 min being

M

in the open air (23-25 ). The pH measurements were performed using flat-surface pH glass electrode. The paper substrates were placed on a clean and dried Teflon plate. Then ultra-pure

d

water was pipetted on the paper substrate to reach each of the papers full liquid retention capacity. Subsequently, the flat surface pH glass electrode was then placed on top of the wet

te

paper substrate and pH measurement was performed until the stabilization of the pH readout

Ac ce p

was attained (approx. within one minute).

To determine, if the force used to press the ISEs onto the paper substrates has a detrimental effect on the ion-selective membrane, thus, the potentiometric response, various pressures, e.g. 7.0, 13.8, 27.7, 41.5, and 69.2 N cm-2 were applied between the ISEs and paper substrates. In details, the electrode stand was used to fix and press the lead(II) solid-contact electrode and the reference electrode towards the paper substrates after pipetting a standard solution of lead(II) nitrate. Subsequently, to evaluate the mechanical resistance of the ISM to the pressure applied towards the paper substrates, JSM-7600F Field Emission Scanning Electron Microscopy (Japan) was used before and after application of various pressures between ISEs and the paper substrate. Before that, each paper substrate and Pb2+-ISM were cut into two pieces where one piece was characterized as before (prepared dry paper substrates and conditioned Pb2+-ISEs), whereas another one after the pressure test was completed. 2.4. The paper-based potentiometric measurements with Cd2+- and Pb2+-ISEs. 5 Page 5 of 32

The potentiometric measurements were carried out using an EMF16 Interface potentiometer, Lawson Labs Inc. assisted with L-EMF DAQ 3.0 software (USA). Potentiometric response in each standard solution was registered for 60 s at room temperature (23-25 ). The measurements were done with the approximately tenfold change of primary ion activity, from

ip t

10-7.0 to 10-1.4 M and from 10-7.0 to 10-1.3 M for Cd2+- and Pb2+-ISEs, respectively, in the order of increasing ion activity in the analyte. Standard deviation was calculated using three consecutives measurements of the same type. Before paper-based sampling assisted

cr

potentiometric measurements, the Cd2+- and Pb2+-ISEs were tested for their Nernstian response in a conventional beaker-based potentiometric cell. The electrodes that exhibited

us

Nernstian response in beaker-based measurements were further used and coupled with paperbased microfluidic sampling. In the paper-based assisted potentiometric measurements, the

an

potentiometric cell was pressed by gravitational force (if not stated otherwise) against paper substrates that contained either Cd(NO3)2 or Pb(NO3)2 standard solutions. Once electrodes were positioned, the potentiometric response was then recorded for 60 s in each standard

M

solution by registering EMF every 0.2s. For each paper based sampling, a new paper substrate was used.

d

The paper-based sampling were performed in two various protocols of the samplings, which further on are denoted as protocols 1 and 2 that differ in the time that the paper substrate was

te

in contact with the standard solutions. Thus protocol 1 sampling was performed when

Ac ce p

pipetting a standard solution of cadmium/lead(II) nitrate on the clean oven-dried paper substrates, while in protocol 2 the clean oven-dried paper substrates were soaked in standard solutions for 30 min. After that paper substrates were taken out, the excess of the standard solution was removed through gentle shaking off the paper substrate, placed on the clean Teflon plate with the subsequent potentiometric cell placement, followed by immediate potentiometric measurement.

Subsequently, for paper based sampling two more paper substrates pre-treatments protocols were investigated, further one denoted as protocols 3 and 4. In both cases, the paper substrates was pre-treated in the inorganic salt solutions (Cd(NO3)2 or Pb(NO3)2) at concentrations 10-6 M, 10-5 M, 10-4 M and 10-3 M for 30 min while gently stirring. Then, the soaked paper substrates were vigorously washed with ultra-pure water and oven-dried at 70°C for 30 min. Such paper substrates were used in protocols 3 and 4. In protocol 3, the ion used for pre-treatment of the paper based substrates was same as the one of primary ion in the standard solutions used in potentiometric measurement, e.g. paper based substrates were pre6 Page 6 of 32

treated with different concentrated solutions of Pb(NO3)2 or Cd(NO3)2, with subsequent measurement of lead(II) and cadmium(II), respectively. In protocol 4, the ion (interfering ion) used for pre-treatment of the paper based substrates was different as the one of primary ion in the standard solutions used in potentiometric measurement, e.g. paper based substrates were

ip t

pre-treated with different concentrated solutions of Pb(NO3)2 or Cd(NO3)2, with subsequent measurement of cadmium(II) and lead(II), respectively. Then a standard solution of cadmium/lead(II) nitrate was pipetted on the paper substrates, after that the excess of the

cr

standard solution was removed through gentle shaking off the paper substrate, placed on the clean Teflon plate with the subsequent potentiometric cell placement, followed by immediate

us

potentiometric measurement.

2.5. Determination of Pb2+ in artificial samples containing background sample matrix of an

an

environmental surface water.

The protocol 3 was chosen to showcase the performance of Pb2+-ISE coupled with

M

microfluidic paper based sampling in the determination of lead(II) in artificial sample prepared with background matrix of an environmental surface water. For the sample preparation, the surface water was collected from residual (locally drained) lake located

d

within NTU campus. The separated clear water was filtered using Acrodisc 25 mm Syringe

te

Filter with 0.45 µm supor membrane (Pall, USA). The pH of the lake water was found to be 7.2 thus an appropriate aliquot of 70% HNO3 was added to the sample to obtain pH of 3.5, to

Ac ce p

assure full ionization of metals in the sample solution. Then the elemental concentrations (Pb, Cd, As, Ca, Co, Cu, Fe, Mg, Mn, Ni, Sb, Sr, Zn, Li, Cr) in the surface water was measured by

PerkinElmer Optima 8300 inductively coupled plasma optical emission spectrometry (ICPOES). As a result, low concentration of lead in such solution was found (1.3 µM Pb), thus the surface water sample was used to prepare a stock solution of 0.1 M Pb(NO3)2, from which through dilution with prepared surface water three sample solutions with various lead(II) concentrations (in the range of 10-5.0-10-2.2 M Pb2+) were prepared. The pH of all three samples remain the same (3.5 ± 0.2). Prior the determination of lead(II) in artificial samples, a calibration curve for Pb2+-ISE was prepared in the range of 10-5.0-10-2.2 M Pb(NO3)2 pure standard solutions utilizing microfluidic paper based sampling protocol 3 when PS1 was pretreated with 10-3 M Pb(NO3)2. The PS1 pre-treated with 10-3 M Pb(NO3)2 and the protocol for Pb2+-ISE calibration were as described in section 2.4. After the calibration was completed and was deemed to be Nernstian (29.0 mV/dec) the determinations of lead(II) in artificial samples were performed. The potential of Pb2+-ISE in each artificial sample was measured 7 Page 7 of 32

three times (n= 3), for each measurement the fresh sample solutions and the fresh paper substrates were used. The results of Pb2+ concentrations determined by Pb2+-ISE were compared to those obtained by ICP-OES by the determination of total lead concentration in sample solutions. The determinations of lead concentration by both analytical methods was

ip t

performed at room temperature (23-25 ). 2.6. Cd2+ and Pb2+ ions retention in the paper substrates.

cr

The retention of Cd2+ or Pb2+ within the paper substrates was evaluated by ELAN® DRC-e inductively coupled plasma mass spectrometry (ICP-MS) (USA) after paper substrates (3 × 3

us

cm2) pre-treatment with various concentrations of Cd(NO3)2 or Pb(NO3)2 at concentrations 10-6 M, 10-5 M, 10-4 M and 10-3 M for 30 min while gently stirring. Then, the soaked paper substrates were vigorously washed with ultra-pure water and oven-dried at 70°C for 30 min.

an

After that, paper substrates were weighted and cut into smaller pieces (approx. 1 × 1 cm2) and transferred into a 100 mL Teflon vessel to run microwave digestion in a solution of 7 mL

M

70% trace metal grade HNO3. The microwave digestion was performed using the Titan Microwave Sample Preparation System (MPS) (USA), where the sample was kept for 5 min in 120℃ and 30 bar, then while keeping the pressure constant the temperature was further

d

increased to 170℃ and kept for another 5 min, finally the temperature was decreased to 50℃

te

and kept for 10 min. After the digestion the sample was left to cool down at a room temperature (23-25℃). Then digested sample was transferred into a 25 mL volumetric flask

Ac ce p

and topped up with ultra-pure water. Acrodisc 25 mm Syringe Filter with 0.45 µm supor membrane (Pall, USA) was used to filter out the possible impurities before the sample analysis by ELAN® DRC-e ICP-MS (USA).

3. Results and discussion

3.1. Physiochemical parameters of paper substrates and paper based microfluidic sampling coupled with ISEs. Paper is a versatile material with numerous possibilities of modifications and utilization, especially in sensing technology [12]. To understand the full capability of using paper substrates to sample and measure heavy metals in various samples, in this study, unmodified and modified cellulose paper substrates were utilized for paper based microfluidic sampling. The physiochemical properties of paper substrates and paper based microfluidic sampling 8 Page 8 of 32

were summarized in Table 1. In general, PS1, PS2 and PS5 had a similar liquid absorption capacity, which was higher than the one observed for PS3 and PS4. The PS3 and PS4 paper substrates had the smallest pore size, which directly influenced their ability to absorb the liquid compared to the other paper substrates. The paper substrates were also tested for their

ip t

ability to retain liquid in paper matrix over short period of time. This was investigated through looking at evaporation rate within the designated potentiometric measurement time (60 s). It could be observed that the evaporation rate for all paper substrates was approx. 2 to

cr

3 µL min-1. The analyte concentration change caused by the evaporation after 60 s (applied measurement time) for PS1, PS2, PS3, PS4 and PS5 were 1.3, 1.2, 2.4, 2.5 and 1.8%,

us

respectively. As a result, evaporation of water ought not to have a significant influence on overall potentiometric measurements performed within 60 s single measurement time. The

an

pH of all paper substrates were slightly acidic, approx. 5 to 5.7. The acidity of papers may be a result of the presence of hydroxyl groups [32] or simply paper substrates manufacturing process, where all substrates are acid washed to reach ashless and a high purity of the

M

material [33]. Nonetheless, such acidity of the paper substrates ought not to have a detrimental effect on the speciation of Cd2+ and Pb2+ in the standard solutions, at which ions stay in nearly 100% ionized form [34, 35]. The influence of force applied between the solid-

d

contact Pb2+-ISEs in the potentiometric cell and the paper substrates was investigated in order

te

to verify the mechanical resistance of ISM to mechanical stress and the influence of such mechanical stress on the potentiometric response of the liquid polymeric membrane ISEs.

Ac ce p

Since the solid-state crystalline membrane Cd2+-ISE had a robust membrane, only solidcontact Pb2+-ISEs with relatively fragile PVC based ISM were investigated in this study. Fig 1 (A) presents the measurements with Pb2+-ISEs performed utilizing PS5 with the microfluidic sampling protocol 2 to obtain calibration curves in the concentration range of 107.0

-10-1.4 M Pb2+. As a result, the pressure between the Pb2+-ISEs and the paper substrates had

no significant influence (between 10-5.0-10-1.4 M) on the Nernstian response of the electrodes. The slight changes in the EMF (within 14 mV) were observed at the lowest analyte concentrations (between 10-7.0-10-5.0 M), where the higher pressure applied resulted in the lower potential readout. This however, had no direct influence on the overall lowering of the detection limit of the Pb2+-ISEs. Furthermore, in order to find out whether there was any mechanical damage to the ISM of Pb2+-ISEs either by residues of ISM on the paper substrate or mechanical cuts/scratches on the ISM of Pb2+-ISEs, FESEM of both the paper substrate and the ISM before and after the pressure test (69.2 N cm-2) were investigated and presented in Fig 1 (B), (C), (D) and (E). To summarize, there was no noticeable differences in surface 9 Page 9 of 32

morphology of both paper and Pb2+-ISM before and after pressure test. Since the ISM of Pb2+-ISEs was found resistant to mechanical stress it can be concluded that pressure does not need to be controlled during potentiometric measurements with liquid polymeric membrane ISEs utilizing paper based sampling. Thus, further on all potentiometric measurements

onto the paper substrates with only gravitational force.

cr

ip t

utilizing paper based microfluidic sampling were conducted with electrodes being pressed

3.2. The effect of paper based microfluidic sampling on potentiometric measurements of Cd2+

us

and Pb2+ using solid-state Cd2+- and solid-contact Pb2+- ISEs.

Fig 2 presents the paper based potentiometric measurements utilizing microfluidic paper

an

sampling protocols 1 and 2 using PS2 with a solid-state Cd2+-ISEs and a solid-contact Pb2+ISEs. For both electrodes, namely Cd2+- and Pb2+-ISEs, the application of the microfluidic paper sampling protocol 1, resulted in a super Nernstian potential response between 10-4.0-10M Cd2+ and Pb2+. The super Nernstian effect is a disturbance caused by the application of

M

3.1

the paper substrates at the sample solution | ISM interfaces. Typically, such effects are associated with unconditioned membranes, where the flux of ion into ISM (enhanced by

d

ionophore-ion binding) results in depletion of ion at the sample solution | ISM interface [36-

te

40]. Also, it was observed before that microfluidic paper based sampling influences the heavy metal measurements [11]. In this situation, the super-Nernstian response is caused by the

Ac ce p

binding of metal ions onto cellulose based paper substrates, e.g. Cd2+ and Pb2+ via hydroxyl groups present in the cellulose [12]. Additionally, cellulose based materials were found to be a good sorbents of heavy metals [41]. In any case, the metal ions interacting with paper substrates at lower analyte concentration depleted the sample solution | ISM interface from heavy metals and the super Nernstian response occurred [11]. Obviously, this is a serious limitation considering usefulness of the microfluidic paper based sampling for the determination of heavy metals in various samples. Thus, the research directions in this work were to eliminate the super Nernst effect, thus extend the Nernstian response of the heavy metal sensing ISEs. One of the ways to eliminate the super Nernstian response for the determination of heavy metals using microfluidic paper based sampling was to condition the paper substrates in each of the standards and sample solution before actual potentiometric measurement, similarly as it was done before [11]. As a result, for both Cd2+- and Pb2+-ISEs, the super Nernstian response was eliminated after application of microfluidic sampling

10 Page 10 of 32

protocol 2. For Cd2+-ISEs the slope of the calibration curve in the concentration range 10-5.010-1.3 M Cd2+ was 33.8 ± 1.0 mV dec−1, while for Pb2+-ISEs the slope of the calibration curve in the concentration range 10-5.0-10-2.2 M Pb2+ was 29.3 ± 1.5 mV dec−1. For both types of ISEs the lower detection limit was 10-5.0 M, tenfold higher than that observed for solution

ip t

based potentiometric measurements (results not shown here). Same approach was investigated in our previous work and was similarly found that conditioning of paper substrates for 30 min in each of the standard solutions (as in protocol 2) resulted in liner

cr

Nernstian response compared to measurements done without paper substrate conditioning. In fact, the performance of Cd2+-ISE was found to be better than previously reported [11]. It is

us

clear that the paper pre-treatment with cations may reduce the paper interaction with the ions in the sample solution during ion concentration determination. In this protocol, however, the

an

paper substrate is simply equilibrated with the sample which has its disadvantages, namely: (i) time for equilibration (30 min) is too long for any practical applications, (ii) substances released from paper matrix during the conditioning process may contaminate the standards

M

and sample solution and (iii) the conditioning partially consumes cations from the standards and sample solution, thus altering the concentration of the analyte. For that reason, there is a clear need of another paper substrate modification protocol that would allow more reliable

d

application of paper based microfluidic sampling in heavy metal sensing.

te

All paper substrates, except of PS3, coupled with Cd2+- and Pb2+-ISEs in microfluidic

Ac ce p

sampling protocol 1 and 2 behaved as described for PS2 (supporting information, Fig S1 and Table S1). Moreover in microfluidic sampling protocol 1, the super Nernstian potential jump was higher for Pb2+-ISEs comparing to Cd2+-ISEs utilizing the same type of paper substrate (supporting information, Fig S1). For example when using Pb2+-ISEs with PS1, PS2, PS3, PS4 and PS5 the super Nernstian jumps were 94.4 ± 5.3, 82.5 ± 32.1, 60.3 ± 17.8, 62.0 ± 25.0 and 62.1 ± 6.3 mV, respectively. On the other hand, when using Cd2+-ISEs with PS1, PS2, PS3, PS4 and PS5 the super Nernstian jumps were 53.8 ± 3.1, 47.9 ± 1.0, 27.8 ± 0.1 (Nernstian response), 38.1 ± 5.1 and 46.6 ± 3.5 mV, respectively. The higher super Nernstian jump for Pb2+-ISEs may be attributed to the fact that Pb2+ stronger binds to cellulose than Cd2+ [42]. Uniquely, Cd2+-ISEs using microfluidic sampling protocol 1 with PS3 did not exhibit super Nernstian response (Fig 3). The potentiometric measurements performed using PS3 for microfluidic sampling, in protocol 1 and 2, resulted in the similar and linear response between 10-5.0-10-1.3 M Cd2+. In both cases, the slopes were close to Nernstian 30.8 ± 0.5 and 30.7 ± 2.1 mV dec−1, respectively. The extended linear range, when performing immediate 11 Page 11 of 32

potentiometric measurements after sample collection, showed a great perspective of this paper to be directly used in potentiometric measurement for Cd2+ concentration determination. On the other hand, the measurements of lead(II) using the same PS3 paper substrate still resulted with occurrence of super Nernstian response (supporting information,

ip t

Fig 1S). Thus it is clear that the paper substrate interacts weaker with cadmium than lead(II) and modifications to paper substrates to block metal active groups in the structure of the paper are necessary to extend the applicability of paper based sampling to measurements of

cr

all heavy metals.

us

To summarize, each paper substrate is made of same material (pure cellulose) while all of them are characterized with different pH and pore sizes. The latter ones indicate the need of customized preparation protocol to adjust the pore sizes through different chemical/physical

an

treatment of each type of substrates. Also, the origin of the cellulose material used for the preparation of the paper substrates may play a role in their unique ion-solution interactions in potentiometric sensing, e.g. as observed for PS3. The paper substrates used in this work were

M

the commercial products and the information about their preparation is unknown. Thus it is important to design and study custom made paper substrates to link their potentiometric

d

behaviour to paper substrates composition (origin of cellulose material), physio-chemical

Ac ce p

manufacturing process.

te

properties (pore size, absorption capability, thickness, pH and chemical composition) and

3.3. Paper substrate pre-treatment with different primary ion concentrations. In microfluidic paper based sampling protocol 2 each of the paper substrate was allowed to stay in contact with corresponding standard solution for 30 min. In this process, all metal active sites at the paper substrates were consumed by the inorganic salt/ions and the measurement could be performed, this protocol however has its serious disadvantages as mentioned in previous chapter. Apart of application of inorganic salts that interact with active sites at the paper substrates, organic salts are expected to have a similar behaviour. For example,

modification

of

paper

substrates

with

lipophilic

salts,

e.g.

tridodecylmethylammonium chloride should allow bulky cation to occupy all active sites at the paper substrates and later on, during potentiometric measurement, make these sites unavailable for metal ions. Although, these seems as a very interesting approach, which we

12 Page 12 of 32

plan to investigate in the future, in this work we solely focus on inorganic salts modified paper based substrates. To investigate the applicability of the inorganic salt modification to the paper substrate and to eliminate the disadvantages of microfluidic paper based sampling protocol 2, the microfluidic

ip t

paper based sampling protocol 3 was developed. In this protocol, paper substrate was treated with different inorganic salt solutions before the microfluidic sampling of the analyte was

cr

performed. Thus the modification/blocking the active sites on the paper substrates was a separate process and was performed before from the calibration/sample determination

us

measurements. For that paper based substrates were pre-treated with different concentrated solutions of Pb(NO3)2 or Cd(NO3)2, with subsequent measurement of lead(II) and cadmium(II), respectively. Fig 4 presents the potentiometric response of Cd2+-ISE and Pb2+-

an

ISE in Cd(NO3)2 solutions using PS1 and PS2 with prior pre-treatment of paper substrates in concentration range of 10-7.0-10-1.3 M of Cd(NO3)2 and Pb(NO3)2, respectively. The results were compared to the ones obtained using microfluidic paper based sampling protocols 1 and

M

2. For the Cd2+-ISEs, the super-Nernstian potential jump between 10-3.1-10-4.0 M Cd2+ as obtained applying the microfluidic paper based sampling protocol 1 was eliminated after the

d

pre-treatment of paper substrates with all concentration of Cd(NO3)2 solutions (microfluidic paper based sampling protocol 3). The paper pre-treatment with 10-4 M Cd(NO3)2 solution,

te

however, resulted in the best potentiometric response (26.6 ± 0.8 mV dec-1), which was

Ac ce p

comparable to the microfluidic paper based sampling protocol 2. Similarly, for all the other paper substrates except of PS3 (the optimal paper substrate pre-treatment concentration was 10-5 M Cd(NO3)2) the paper pre-treatment with 10-4 M Cd(NO3)2 solution was found to be the most optimal (supporting information, Fig S2 and Table S2). For the Pb2+-ISEs, the superNernstian potential jump between 10-3.1-10-4.0 M Pb2+ as obtained applying the microfluidic paper based sampling protocol 1 was eliminated after the pre-treatment of the paper substrates with only some of Pb(NO3)2 solutions (microfluidic paper based sampling protocol 3). The paper pre-treatment with 10-4 M Pb(NO3)2 solution resulted in the best potentiometric response (29.4 ± 0.9 mV dec-1), which was comparable to the microfluidic paper based sampling protocol 2. The results using other pre-treatment Pb(NO3)2 solution were gathered in the supporting information Fig S3 and Table S1. The pre-treatment of paper substrates with only 10-3 and 10-4 M Pb(NO3)2 resulted in close to Nernstian response of Pb2+-ISEs utilizing the microfluidic paper based sampling protocol 3. In both cases, the pre-treatments of paper substrates with more concentrated inorganic salts (10-3 and 10-4 M) were more 13 Page 13 of 32

beneficial for overall potentiometric response when such paper substrates were further on utilized for solution sampling. This suggest, in this particular system, that to overcome relatively weak binding of metals to cellulose based hydroxyl groups the high concentration of heavy metals (primary ions) is needed during paper substrates pre-treatment. Furthermore,

ip t

when low concentration of heavy metals (primary ions) during paper substrates pre-treatment were applied, the reduction in binding efficiency of heavy metals to the active sites at the paper substrates was expected. Thus, the concentrations of inorganic salts at 10-5 and 10-6 M

cr

during pre-treatment of paper substrates was not effective for Pb2+-ISEs that suggested the

us

insufficient concentration of the ion to block all metal active sites on the paper substrates.

an

3.4. Paper substrate pre-treatment with different interfering ion concentrations. As presented in the previous chapter, by pre-conditioning of the paper based substrates with inorganic salt solutions and using such substrates for microfluidic paper based sampling

M

(protocol 3) the Nernstian responses of both Cd2+- and Pb2+-ISEs in paper based sampling were attained. It is important to note that the cation in inorganic salt used for modification of

d

paper based substrates was the same as primary ion for ISEs. To investigate the applicability of the different inorganic salt modifications to the paper substrates than the primary ion for

te

ISEs, the microfluidic paper based sampling protocol 4 was introduced. For that paper based substrates were pre-treated with different concentrated solutions of Cd(NO3)2 or Pb(NO3)2

Ac ce p

with subsequent measurement of lead(II) and cadmium(II), by Cd2+- and Pb2+-ISEs, respectively. Fig 5 (A) presents the response Cd2+-ISEs in 10-7.0-10-1.3 M using PS1 utilizing microfluidic paper based sampling protocol 4 when the paper substrates were pre-treated with 10-3, 10-4, 10-5 and 10-6 M interfering ions, Pb2+. For the Cd2+-ISEs, the super-Nernstian potential jump between 10-3.1-10-4.0 M Cd2+ as obtained applying the microfluidic paper based sampling protocol 1 was eliminated after the pre-treatment of paper substrates with application of only 10-6 M Pb(NO3)2 solution (microfluidic paper based sampling protocol 4), 28.9 ± 1.7 mV dec-1. Similarly, for all the other paper substrates except of PS3 for which no Nernstian response was attained the paper pre-treatment with 10-6 M Pb(NO3)2 solution was found to be the most optimal (supporting information, Fig S4). In all other cases, the paper substrate pre-treatment in 10-3, 10-4 and 10-5 M interfering ions, except of PS5 pre-treatment in 10-5 M Pb(NO3)2 solution, resulted in sub-Nernstian potentiometric response. Thus, contrary to the microfluidic paper based sampling protocol 3 where for paper substrate pre-

14 Page 14 of 32

treatment the high concentration of the inorganic salt were needed, here the lowest concentration of interfering ion (10-6 M Pb(NO3)2) was needed to attain Nernstian response. In this case, the selectivity of sensor to the primary ion over interfering ion applied in the paper substrate pre-treatment as well as interfering-primary ion preferential binding to active

ip t

groups in the paper substrates may alter the patterns observed in microfluidic sampling protocol 3. In general, the selectivity of solid-state crystalline membranes suffers from the presence of interfering ions [5, 43, 44]. Thus the results indicated that lead(II) undergoes

cr

possible partial desorption from the paper substrate to the standard solutions or ion-exchange at the ISM | paper substrate interface, negatively affecting the response of Cd2+-ISEs towards

us

sub-Nernstian. Moreover, the fact that only 10-6 M Pb(NO3)2 was found to be suitable pretreatment solution indicates that paper substrate metal ion interactions are concentration

an

dependent, involving not only chemical but also physical interactions of metal ion with the paper substrates (explored in the chapter 3.6). Fig 5 (B) presents the response Pb2+-ISEs in 10-7.0-10-1.4 M using PS2 utilizing microfluidic paper based sampling protocol 4 when the

M

paper substrates were pre-treated with 10-3, 10-4, 10-5 and 10-6 M interfering ions, Cd2+. None of the pre-treatment interfering ion concentrations, namely 10-3, 10-4, 10-5 and 10-6 M Cd2+ were sufficient to attain Nernstian response of Pb2+-ISEs. Contrary to the previous case, all

d

Pb2+-ISEs responses were super-Nernstian, similar to the ones obtained using the microfluidic

te

paper based sampling protocol 1 (supporting information Fig S5). It has been documented before that Pb2+ interacts more strongly with wood fibres than Cd2+ and alkali metals [45].

Ac ce p

Thus the pre-treatment of paper substrates with weaker bound Cd2+ resulted in its insufficient adsorption capacity onto the paper substrate that promoted further adsorption of Pb2+ onto the paper substrates when in contact in the lead(II) solutions. This indicates also that paper substrates have high adsorption capacity for metal ions (explored in the chapter 3.6).

3.5. Determination of Pb2+ in artificial samples containing background sample matrix of the environmental surface water.

The artificial samples contained the background sample matrix of the surface water. The ICPOES of the artificial samples, after lead(II) addition and pH acidification, indicated 54.84 µM Mg, 12.66 µM Li, 5.2 µM Fe, 2.82 µM Sr, 0.96 µM Ca and 0.1 µM Cu. The other investigated elements, namely Cd, As, Co, Mn, Ni, Sb, Zn and Cr were found to be below the detection limit of ICP-OES. The results of lead concentration in three artificial samples

15 Page 15 of 32

determined using Pb2+-ISE and ICP-OES are shown in Table 2. The potentiometric measurement utilizing microfluidic paper based sampling protocol 3 with Pb2+-ISE was found comparable to the one done using ICP-OES. In all samples, slightly higher concentration of lead was found using ICP-OES. The fact that ICP-OES measures total lead

ip t

concentration while Pb2+-ISE measures ionized lead(II) concentration usually results in slightly higher results using the latter method. However, in this case owning to acidification of the samples there should not be variations in these two fractions, as it may be assumed

cr

100% ionization of lead in the artificial samples at pH= 3.5, unless e.g. lead was bound in pH stable, organic complexes. This can be valid taken the background electrolyte of the sample

us

was taken from environmental conditions. On the other hand, the calibration of Pb2+-ISE was done in pure lead(II) nitrate solutions and no corrections for the ionic strength between

an

artificial samples and standards were done. Fortunately, the ionic content of artificial samples was found to be relatively low, thus the determinations of the Pb2+ were found not to be/or negligibly affected by changes in ionic strength. Nonetheless, the deviation in determined

M

concentration of lead in artificial samples using Pb2+-ISE and ICP-OES was 5, 4 and 2% for samples 1, 2 and 3, respectively. Thus, the lead determination using inorganic salt modified paper substrates in microfluidic paper based sampling protocol 3 was found to be viable for

te

d

the determination of lead in synthetic samples.

Ac ce p

3.6. Interactions between paper substrates and the metal ions. Paper substrates pre-treated with different concentration of primary and interfering ions as described in microfluidic sampling protocols 3 and 4 showed that the concentration of the inorganic salt used for conditioning of paper substrates influences the overall potentiometric performance of the ISEs coupled with paper based microfluidic sampling. To check the interactions of paper substrates with metal ions during the pre-conditioning with inorganic salts, PS4 was conditioned in various concentrations of pure solutions of Cd(NO3)2 and Pb(NO3)2 (Fig 6). After the conditioning, the excess of unbound and loosely bound metal ions were removed by vigorous washing the paper substrates with ultra-pure water, thus the ICPMS of digested paper substrates represented only bound metal ions to the paper substrates. It was found that the sorption capacity of paper substrate was directly proportional to the concentration of the ion in the solution during the paper substrate conditioning. The interaction, slope of the concentration measured vs. concentration used for conditioning, was

16 Page 16 of 32

higher for Pb2+ sorption than Cu2+, namely 143.9 ± 11.3 and 47.7 ± 0.7, respectively. Thus, the Pb2+ ions bounded to the paper substrate with higher affinity than Cd2+ ions which was in agreement as reported by other authors [45]. The similar trend was observed for all applied paper substrates (supporting information, Fig S6, Tables S3 and S4). Moreover, the paper substrates, e.g. for PS4 were able to adsorb significant heavy metal amount, approx. 450 ppm

ip t

of Pb2+ and 125 ppm of Cd2+ for pre-conditioning in 10-3 M solution. These suggest that if untreated, the paper substrate used for microfluidic sampling or simply being a part of

cr

electrode construction may easily alter the concentration of the analyte in the sample solution, which is very valid for all the research that implements paper for construction of paper

us

devices, e.g. electrodes, especially if those are developed for the intention to measure low analyte concentrations or/and to measure heavy metal concentrations. The cellulose based

an

materials, e.g. sesame husk was found to be an efficient adsorbent for cadmium ion [41]. Both physical and chemical adsorption of heavy metals were recognized to drive sorbentmetal interactions. Since paper substrates used in this study were cellulose based, a similar

M

metal ion-paper properties were expected. For that, as it was previously identified some of the ions are bound to the paper via physisorption that is governed by weak Vander Waal’s forces and they can be a subject of desorption from the paper to the solution [41]. Thus, if the paper

d

is modified with inorganic salt/primary ion, the cation can be desorbed at low analyte

te

concentration and pollute the sample solution, and the same time negatively influence the low detection limit of the sensor. That, possibly is the reason why even after paper substrate

Ac ce p

modifications, the ISEs coupled with microfluidic paper based sampling tend to have higher low detection limit (at least tenfold higher) then the ones used in solution based measurements. This may be a limitation for possible measurements of heavy metals at low analyte concentrations, this report was not directed towards lowering the determination limit of the method which is yet to be investigated.

Conclusions

The crystalline membrane solid-state Cd2+- and a solid-contact Pb2+-ISEs were used to investigate the possibility to facilitate the use of paper based microfluidic sampling coupled with potentiometric sensors for the determination of heavy metals. Solid-contact type of ISEs was used here for the very first time and their use was found to be compatible with paper based sampling. Moreover, paper substrates pre-treatment was applied to reduce the metal

17 Page 17 of 32

ions interaction between the standards and the sample solutions and the paper substrates, thus to eliminate unfavourable super Nernstian response of the ISEs. It was found that relatively high concentration of inorganic salt solution was needed to modify the paper substrates prior to their use for paper based sampling coupled with ISEs. Further on, when the paper substrates were pre-treated with interfering ions, only low concentration of 10-6 M Pb2+

ip t

solution was applicable when coupled with Cd2+-ISEs. The inorganic modified paper based substrates were validated for their use in lead(II) determination in artificial samples and were

cr

found to be comparable to the one done with ICP-OES. Finally, it was found that paper substrates were very good sorbents of metal ions, with stronger binding force given to Pb2+

us

over Cd2+. Moreover, the metal-paper interactions suggest concentration dependent sorptiondesorption of metal ion, which can have direct limitations of the use of paper based sampling

an

for the determination of low analyte concentration of heavy metals. Despite that, we have managed to obtain Nernstian response of ISEs down to micromolar concentration range

d

Declarations of interest and funding

M

without jeopardising the overall time of measurement.

The authors declare no competing financial interest. The NTU-India Connect Research

te

Internship Program provided funding for the internship Ms Tharini Saravana Sundaram from

Ac ce p

Coimbatore Institute of Technology, India at the initial stage of this research project.

Acknowledgements

The authors would like to thank NTU-India Connect Research Internship Program for financial support of Ms Tharini Saravana Sundaram from Coimbatore Institute of Technology, India.

References [1] P.B. Tchounwou, C.G. Yedjou, A.K. Patlolla, D.J. Sutton, Heavy metal toxicity and the environment, EXS. (2012). doi:10.1007/978-3-7643-8340-4_6.

18 Page 18 of 32

[2] L. Järup, Hazards of heavy metal contamination, Br. Med. Bull. (2003). doi:10.1093/bmb/ldg032. [3] G.F. Nordberg, B.A. Fowler, M. Nordberg, L.T. Friberg, Introduction-general and

international

perspectives,

Handb.

Toxicol.

(2007).

cr

doi:10.1016/B978-012369413-3/50056-2. [4] Drinking water contaminants: Standards and Regulations, EPA

Met.

ip t

considerations

us

http://water.epa.gov/drink/contaminants/index.cfm (accessed 30.06.2011).

[5] M. Guziński, G. Lisak, J. Kupis, A. Jasiński, M. Bocheńska, Lead(II)-selective for

ion-selective

electrodes:

review,

Anal.

Chim.

Acta.

(2013).

M

doi:10.1016/j.aca.2013.04.044.

A

an

ionophores

[6] B. Peng, Y. Shen, Z. Gao, M. Zhou, Y. Ma, S. Zhao, Determination of total iron in water

d

and foods by dispersive liquid-liquid microextraction coupled with microvolume UV-vis

te

spectrophotometry, Food Chem. (2015). doi:10.1016/j.foodchem.2014.12.084.

Ac ce p

[7] S. Tanikkul, J. Jakmunee, S. Lapanantnoppakhun, M. Rayanakorn, P. Sooksamiti, R.E. Synovec, G.D. Christian, K. Grudpan, Flow injection in-valve-mini-column pretreatment combined with ion chromatography for cadmium, lead and zinc determination, Talanta. (2004). doi:10.1016/j.talanta.2004.05.060. [8] A. Moghimi, Preconcentration ultra trace of Cd(II) in water samples using dispersive liquid-liquid microextraction with salen(N,N′-bis(salicylidene)- ethylenediamine) and determination graphite furnace atomic absorption spectrometry, J. Chinese Chem. Soc. (2008). doi:10.1002/jccs.200800054. [9] T. Blaz, B. Baś, J. Kupis, J. Migdalski, A. Lewenstam, Multielectrode potentiometry in a one-drop sample, Electrochem. Commun. (2013). doi:10.1016/j.elecom.2013.05.021.

19 Page 19 of 32

[10] J. Cui, G. Lisak, S. Strzalkowska, J. Bobacka, Potentiometric sensing utilizing paperbased microfluidic sampling, Analyst. (2014). doi:10.1039/c3an02157b. [11] G. Lisak, J. Cui, J. Bobacka, Paper-based microfluidic sampling for potentiometric

ip t

determination of ions, Sensors Actuators, B Chem. (2015). doi:10.1016/j.snb.2014.07.044.

cr

[12] S.K. Mahadeva, K. Walus, B. Stoeber, Paper as a platform for sensing applications and other devices: A review, ACS Appl. Mater. Interfaces. (2015). doi:10.1021/acsami.5b00373.

us

[13] C. Arnaud, Making diagnostics affordable, Chem. Eng. News. (2009). doi:10.1021/cen-

an

v087n011.p050.

[14] A. Määttänen, U. Vanamo, P. Ihalainen, P. Pulkkinen, H. Tenhu, J. Bobacka, J.

M

Peltonen, A low-cost paper-based inkjet-printed platform for electrochemical analyses, Sensors Actuators, B Chem. (2013). doi:10.1016/j.snb.2012.10.113.

d

[15] D.M. Cate, J.A. Adkins, J. Mettakoonpitak, C.S. Henry, Recent Developments in Paper-

te

Based Microfluidic Devices., Anal. Chem. (2015). doi:10.1021/ac503968p.

Ac ce p

[16] T.H. da Costa, E. Song, R.P. Tortorich, J.-W. Choi, A Paper-Based Electrochemical Sensor Using Inkjet-Printed Carbon Nanotube Electrodes, ECS J. Solid State Sci. Technol. (2015). doi:10.1149/2.0121510jss.

[17] P. Sjöberg, A. Määttänen, U. Vanamo, M. Novell, P. Ihalainen, F.J. Andrade, J. Bobacka, J. Peltonen, Paper-based potentiometric ion sensors constructed on ink-jet printed gold electrodes, Sensors Actuators, B Chem. (2016). doi:10.1016/j.snb.2015.10.051. [18] S. Ge, L. Ge, M. Yan, X. Song, J. Yu, J. Huang, A disposable paper-based electrochemical sensor with an addressable electrode array for cancer screening, Chem. Commun. (2012). doi:10.1039/c2cc34887j.

20 Page 20 of 32

[19] K. Huang, Y. Chen, F. Zhou, X. Zhao, J. Liu, S. Mei, Y. Zhou, T. Jing, Integrated ion imprinted polymers-paper composites for selective and sensitive detection of Cd(II) ions, J. Hazard. Mater. (2017). doi:10.1016/j.jhazmat.2017.03.035.

ip t

[20] Y. Xu, B. Lou, Z. Lv, Z. Zhou, L. Zhang, E. Wang, Paper-based solid-state electrochemiluminescence sensor using poly(sodium 4-styrenesulfonate) functionalized

cr

graphene/nafion composite film, Anal. Chim. Acta. (2013). doi:10.1016/j.aca.2012.12.009.

us

[21] E. Bihar, Y. Deng, T. Miyake, M. Saadaoui, G.G. Malliaras, M. Rolandi, A Disposable paper breathalyzer with an alcohol sensing organic electrochemical transistor, Sci. Rep.

an

(2016). doi:10.1038/srep27582.

M

[22] M. Cuartero, G.A. Crespo, E. Bakker, Paper-based thin-layer coulometric sensor for halide determination, Anal. Chem. (2015). doi:10.1021/ac504400w.

d

[23] J.C. Cunningham, K. Scida, M.R. Kogan, B. Wang, A.D. Ellington, R.M. Crooks, Paper

te

diagnostic device for quantitative electrochemical detection of ricin at picomolar levels, Lab

Ac ce p

Chip. (2015). doi:10.1039/c5lc00731c.

[24] C. Kokkinos, A. Economou, D. Giokas, Paper-based device with a sputtered tin-film electrode for the voltammetric determination of Cd(II) and Zn(II), Sensors Actuators, B Chem. (2018). doi:10.1016/j.snb.2017.12.182. [25] X. Wang, J. Sun, J. Tong, X. Guan, C. Bian, S. Xia, Paper-based sensor chip for heavy metal ion detection by SWSV, Micromachines. (2018). doi:10.3390/mi9040150. [26] A.M. López Marzo, J. Pons, D.A. Blake, A. Merkoçi, All-integrated and highly sensitive paper based device with sample treatment platform for Cd2+ immunodetection in drinking/tap waters, Anal. Chem. (2013). doi:10.1021/ac3034536.

21 Page 21 of 32

[27] F. Xiao, Y. Li, X. Zan, K. Liao, R. Xu, H. Duan, Growth of metal-metal oxide nanostructures on freestanding graphene paper for flexible biosensors, Adv. Funct. Mater. (2012). doi:10.1002/adfm.201200191.

ip t

[28] X. Wang, G. Lin, G. Cui, X. Zhou, G.L. Liu, White blood cell counting on smartphone

cr

paper electrochemical sensor, Biosens. Bioelectron. (2017). doi:10.1016/j.bios.2016.10.017.

[29] J. Ding, N. He, G. Lisak, W. Qin, J. Bobacka, Paper-based microfluidic sampling and

us

separation of analytes for potentiometric ion sensing, Sensors Actuators, B Chem. (2017).

an

doi:10.1016/j.snb.2016.11.128.

[30] Bobacka, Potential stability of all-solid-state ion-selective electrodes using conducting

M

polymers as ion-to-electron transducers, Anal. Chem. (1999). doi:10.1021/ac990497z.

[31] N.K. Joon, N. He, M. Wagner, M. Cárdenas, J. Bobacka, G. Lisak, Influence of

based

solid-contact

te

d

phosphate buffer and proteins on the potentiometric response of a polymeric membranePb(II)

ion-selective

electrode,

Electrochim.

Acta.

(2017).

Ac ce p

doi:10.1016/j.electacta.2017.08.126.

[32] D. Fengel, G. Wegener, Wood: chemistry, ultrastructure, reactions., 1983. doi:10.1007/BF02608943.

[33] Quantitative papers/ grade 392 FT-4-105-150, https://www.sartorius.com [34] J. Huang, F. Yuan, G. Zeng, X. Li, Y. Gu, L. Shi, W. Liu, Y. Shi, Influence of pH on heavy metal speciation and removal from wastewater using micellar-enhanced ultrafiltration, Chemosphere. (2017). doi:10.1016/j.chemosphere.2016.12.137.

22 Page 22 of 32

[35] G. Lisak, F. Ciepiela, J. Bobacka, T. Sokalski, L. Harju, A. Lewenstam, Determination of Lead(II) in Groundwater Using Solid-State Lead(II) Selective Electrodes by Tuned Galvanostatic Polarization, Electroanalysis. (2013). doi:10.1002/elan.201200330.

ip t

[36] A. Ceresa, T. Sokalski, E. Pretsch, Influence of key parameters on the lower detection

Electroanal. Chem. (2001). doi:10.1016/S0022-0728(00)00488-5.

cr

limit and response function of solvent polymeric membrane ion-selective electrodes, J.

Limit

of

Carrier-Based

Potentiometric

Sensors,

Anal.

Chem.

(1998).

an

doi:10.1021/ac970690y.

Ion

us

[37] S. Mathison, E. Bakker, Effect of Transmembrane Electrolyte Diffusion on the Detection

M

[38] T. Sokalski, T. Zwickl, E. Bakker, E. Pretsch, Lowering the detection limit of solvent polymeric ion-selective electrodes. 1. Modeling the influence of steady-state ion fluxes, Anal.

d

Chem. (1999). doi:10.1021/ac980944v.

te

[39] G. Lisak, J. Bobacka, A. Lewenstam, Recovery of nanomolar detection limit of solid-

Ac ce p

contact lead (II)-selective electrodes by electrode conditioning, J. Solid State Electrochem. (2012). doi:10.1007/s10008-012-1725-4. [40] G. Lisak, A. Ivaska, A. Lewenstam, J. Bobacka, Multicalibrational procedure for more reliable analyses of ions at low analyte concentrations, Electrochim. Acta. (2014). doi:10.1016/j.electacta.2014.02.091. [41] H.A. El-Araby, A.M.M.A. Ibrahim, A.H. Mangood, A.A.-H. Abdel-Rahman, Sesame Husk as Adsorbent for Copper(II) Ions Removal from Aqueous Solution, J. Geosci. Environ. Prot. (2017). doi:10.4236/gep.2017.57011. [42] P. Su, K. Granholm, L. Harju, A. Ivaska, Binding affinities of different metal ions to unbleached hardwood kraft pulp, in: Holzforschung, 2011. doi:10.1515/HF.2011.080. 23 Page 23 of 32

[43] E. Bakker, P. Bühlmann, E. Pretsch, Carrier-Based Ion-Selective Electrodes and Bulk Optodes. 1. General Characteristics, Chem. Rev. (1997). doi:10.1021/cr940394a.. [44] G. Lisak, M. Wagner, C. Kvarnström, J. Bobacka, A. Ivaska, A. Lewenstam,

ip t

Electrochemical behaviour of poly(benzopyrene) films doped with eriochrome black T as a

cr

Pb2+-sensitive sensors, Electroanalysis. (2010). doi:10.1002/elan.201000459.

[45] S. Peper, C. Gonczy, Potentiometric Response Characteristics of Membrane-Based Cs+-

us

Selective Electrodes Containing Ionophore-Functionalized Polymeric Microspheres, Int. J.

an

Electrochem. (2011). doi:10.4061/2011/276896.

M

Bibliographics

Ding Ruiyu received her B.Sc. (Chem Eng.) from Sichuan University. She is currently doing

d

her Ph.D studies at the School of Civil and Environmental Engineering, Nanyang

te

Technological University, Singapore. Her scientific interests cover electrochemical sensors,

Ac ce p

paper based sampling and environmental monitoring of ionic contaminants.

Vida Krikstolaityte received her Ph.D. degree in Chemistry from Vilnius University, Lithuania.

She did her postdoctoral work at Technical University of Denmark. She is

currently a postdoctoral researcher at Nanyang Technological University, Singapore. Her scientific interests cover biosensors, electrochemistry and development of new materials for batteries.

Grzegorz Lisak obtained his M.Sc. (Chem. Eng.) in physical chemistry in 2007 from Poznań University of Technology, Poland. He obtained his D.Sc. (Tech.) in analytical chemistry in 24 Page 24 of 32

2012 from Åbo Akademi University, Finland. Between 2013 and 2015 he has been a research fellow at University of Wollongong, Australia and Malmö University, Sweden. Since 2016 he is Assistant Professor at Nanyang Technological University, Singapore. Since 2017 he is

ip t

director of Residues and Resources Reclamation Center at Nanyang Environment and Water Research Institute, Singapore. His scientific interests cover electrochemical methods of ion

cr

analysis in environmental samples, paper based sampling, electrocatalysis, waste to energy

Ac ce p

te

d

M

an

us

and value added resources, functional materials and mitigation technology of CO2.

Table 1 Physiochemical properties of paper substrates and paper based microfluidic sampling.

Paper substrate

Liquid absorption capacity (µL cm-2) Evaporation

rate

-1

(µL min ) pH

of

substrates

paper

PS1

PS2

PS3

PS4

PS5

174 ± 0.60

179 ± 0.90

122 ± 1.80

116 ± 0.60

157 ± 1.20

2.30 ± 0.10

2.03 ± 0.40

2.90 ± 0.30

2.83 ± 0.35

2.70 ± 0. 27

5.63 ± 0.10

5.13 ± 0.23

5.57 ± 0.10

5.14 ± 0.38

5.07 ± 0.10

25 Page 25 of 32

Table 2 Lead concentrations in three artificial samples determined using Pb2+-ISE (ionized Pb2+ fraction) and ICP-OES (total Pb fraction). The uncertainty was calculated from three consecutive measurements (n= 3). log CPb by ICP-OES

Sample 1

-2.75 ± 0.02

-2.90 ± 0.004

Sample 2

-3.44 ± 0.05

-3.59 ± 0.004

Sample 3

-4.10 ± 0.01

-4.19 ± 0.005

Ac ce p

te

d

M

an

us

cr

ip t

log CPb2+ by Pb2+-ISE

Fig.1. Potentiometric response of Pb2+-ISEs in 10-7-10-1 M Pb(NO3)2 standard solutions using microfluidic sampling protocol 2 with PS5 when the electrodes were pressed onto the paper substrate with the pressure of 7.0, 13.8, 27.7, 41.5 and 69.2 N cm-2(A). FESEM images of paper substrates, before (B) and after (C) the pressure test and the Pb2+-ISM, before (D) and after (E) the pressure test.

26 Page 26 of 32

50 mV -8

-7

-6

-5

-4

an

us

cr

ip t

Cd2+-ISE utilizing protocol 1 2+ Cd -ISE utilizing protocol 2 2+ Pb -ISE utilizing protocol 1 Pb2+-ISE utilizing protocol 2

-3

-2

-1

0

M

Log ai

Ac ce p

te

d

Fig.2. Potentiometric response of the Cd2+- and Pb2+-ISEs in 10-7.0-10-1.3 M and 10-7.0-10-1.4 M standard solutions of Cd(NO3)2 and Pb(NO3)2 respectively when coupled with microfluidic paper based sampling protocols 1 and 2 using PS2.

27 Page 27 of 32

-20 solution-based measurement protocol 1 protocol 2

-40

ip t

-60

-100

cr

-120 -140

us

EMF / mV

-80

-160

-200 -220 -8

-7

-6

-5

-4

an

-180

-3

-2

-1

0

M

Log aCd

2+

d

Fig.3. Potentiometric response of solution based and paper based sampling coupled with the

te

Cd2+-ISEs and performed in 10-7.0-10-1.3 M standard solutions of Cd(NO3)2 using PS3 in

Ac ce p

microfluidic paper based sampling protocols 1 and 2.

28 Page 28 of 32

ip t cr us an M d te Ac ce p Fig.4. Potentiometric response of the Cd2+- (A) and Pb2+-ISEs (B) in 10-7.0-10-1.3 M and 107.0

-10-1.4 M standard solutions of Cd(NO3)2 and Pb(NO3)2 using PS1 and PS2 respectively,

utilizing microfluidic paper based sampling protocol 3 when the paper was pre-treated with 10-3, 10-4, 10-5 and 10-6 M of primary ion concentrations. The potentiometric response was compared to the ones obtained using the microfluidic paper based sampling protocols 1 and 2.

29 Page 29 of 32

ip t cr us an M d te Ac ce p Fig. 5. Potentiometric response of the Cd2+- (A) and Pb2+-ISEs (B) in 10-7.0-10-1.3 M and 107.0

-10-1.4 M standard solutions of Cd(NO3)2 and Pb(NO3)2 using PS1 and PS2 respectively, in

microfluidic paper based sampling protocol 4 when the paper substrates were pre-treated with 10-3, 10-4, 10-5 and 10-6 M interfering ions, Pb2+ and Cd2+, respectively. The potentiometric response was compared to the ones obtained using the microfluidic paper based sampling protocols 1 and 2.

30 Page 30 of 32

ip t

450

Pb2+ pre-treatment Cd2+ pre-treatment

400

cr

350 300

us

250 200

an

150 100 50

M

Total element concentration after pre-treatment/ppm

500

0 -6

-5

-4

-3

te

d

Log cpre-treatment

Fig.6. The concentration of cadmium and lead in PS4 as measured by ICP-MS after paper

Ac ce p

substrates pre-treatment in the concentration range of 10-6-10-3 M of Cd(NO3)2 (circles) and Pb(NO3)2 (triangles).

31 Page 31 of 32

Research Highlights Microfluidic paper based sampling coupled with solid-state and solid-contact ISEs Development of inorganic salt modified paper substrates for heavy metals analysis.

Ac ce p

te

d

M

an

us

cr

ip t

Metal ions sorption by paper as limiting factor for paper based analytics of metals.

Page 32 of 32