Phosphines as Efficient Dioxygen Scavengers in Nitrous Oxide

Apr 28, 2017 - Danish National Research Foundation: Center for DNA ... The need for aqueous dioxygen scavengers in front of the sensor implies a ...
0 downloads 0 Views 1MB Size
Subscriber access provided by UB + Fachbibliothek Chemie | (FU-Bibliothekssystem)

Article

Phosphines as Efficient Dioxygen Scavengers in Nitrous Oxide Sensors Steffen Gralert Sveegaard, Michael Nielsen, Mikkel Holmen Andersen, and Kurt V. Gothelf ACS Sens., Just Accepted Manuscript • Publication Date (Web): 28 Apr 2017 Downloaded from http://pubs.acs.org on May 2, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Sensors is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

Phosphines as Efficient Dioxygen Scavengers in Nitrous Oxide Sensors Steffen Gralert Sveegaard,a,b† Michael Nielsen,c Mikkel Holmen Andersen,b‡ and Kurt Vesterager Gothelf.a* a

Danish National Research Foundation: Center for DNA Nanotechnology, Interdisciplinary Nanoscience Center (iNANO) and Department of Chemistry, Aarhus University, Gustav Wieds Vej 14, DK-8000 Aarhus C, Denmark b Unisense Environment A/S, Tueager 1, DK-8200 Aarhus N, Denmark c Unisense A/S, Tueager 1, DK-8200 Aarhus N, Denmark KEYWORDS: Amperometric sensors, Nitrous Oxide, Global Warming, Greenhouse Gas, Emission Control, Dioxygen Scavengers, Phosphines

ABSTRACT: A current challenge for development of amperometric sensors for the greenhouse gas nitrous oxide (N2O) is their sensitivity towards dioxygen and trace water. The need for aqueous dioxygen scavengers in front of the sensor implies a background signal from penetrating water vapor. In this paper, we introduce substituted phosphines as dioxygen scavengers and demonstrate the application in an dioxygen-insensitive N2O sensor. Suitably substituted phosphines have been synthesized to achieve good solubility properties in the electrochemically inert solvent propylene carbonate. Several sensors with and without physical separation of the sensing and dioxygen scavenging compartments were made and compared to current commercial sensors. The use of phosphines soluble in organic solvents as dioxygen scavengers yielded a higher sensitivity, albeit with longer response time. Proofof-concept N2O sensors without the physically separated dioxygen scavenger chamber showed a greatly enhanced sensitivity with a comparable response time, thus demonstrating the possibility for greatly simplified sensor construction.

Nitrous oxide (N2O) is a powerful greenhouse gas with a Global Warming Potential (GWP) of nearly three hundred, whereas carbon dioxide is defined to have a GWP of unity.1 The activities of humankind have increased the atmospheric concentration of N2O with 16% compared to its pre-industrial age value, paralleling the rapid increase of atmospheric carbon dioxide.1 Sources of N2O emission include agricultural soils,1 chemical industry,1 automobiles,2 and wastewater treatment plants.3 It is important to develop reliable methods to quantify the emission of N2O, e.g. from wastewater plants, in order to reduce the emission through process optimization. Several operational electrochemical N2O sensors have previously been reported,4–6 and the main challenge has been that the polarization potential required for electroreduction of N2O is higher than that of dioxygen. Hence, electrochemical N2O sensors asis are also sensitive to dioxygen. The issue of dioxygen as interfering species is not restricted to N2O, but also in the case of electrochemical detection of nitrate and nitrite,7,8 where dioxygen scavenging is also key to successful sensor operation. Methods to remove dissolved dioxygen in amperometric applications include the usage of alkaline ascorbate solution,4 enzymes,7 sulfite,9 thiosulfate,10 and by bacterial consumption in e.g. NOx-/NO2- microsensors.11,12 The techniques are, however, only applicable to aqueous solutions of the dioxygen

scavenger. Additionally, since the polarization potential required for the electroreduction of N2O is also high enough to reduce water, the aqueous dioxygen scavenger must be situated in a chamber physically separated by a membrane.4,6 The design of such two-chamber sensor is shown in Figure 1 (also see Figure S-1 in SI).

Figure 1. Overview of existing two-chamber sensor design. Measurements indicate the dimensions of all constructed twochamber sensors and their 95% confidence intervals (n = 6). Dimensions are not to scale.

1

ACS Paragon Plus Environment

ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The sensing chamber which contains a solution of tetrabutylammonium iodide (TBAI) in propylene carbonate (PC) is separated from the dioxygen scavenger chamber by a silicone membrane as reported previously.6 This physical separation is, however, not always sufficient, since in some cases we have observed that an influx of water vapor into the sensing compartment of the two-chamber sensor has a negative impact on the baseline signal over time.6 This is in particular a challenge when the dimensions of the sensors are increased (i.e. inner membrane diameter). In our search for an alternative efficient dioxygen scavenger that is soluble in organic solvents, we have investigated phosphines; a class of organic compounds which, depending on the substituents can be challenging to handle in the laboratory due to their rapid oxidation in air. Given that many known phosphines are inherently unstable in the atmosphere, and even pyrophoric in nature, it was envisioned that phosphines are likely candidates for efficient dioxygen scavengers in organic solvents. A high number of phosphines with a wide range of solubility and reactivity properties are commercially available. It was considered convenient to target a 1 M concentration of phosphines, since the current design of the dioxygen scavenger consists of a 0.75 M solution of sodium ascorbate in 0.675 M NaOH (pH > 13). Since the life-time of the scavenger solution is directly related to the scavenger concentration, such high concentration will ensure that the dioxygen scavenger is not significantly depleted during the life time of the N2O microsensor, which for current microsensors is expected to be between two and six months. Additionally, the reaction rate between scavenger and dioxygen is dependent on the concentration of both, emphasizing that the phosphine solution must be as concentrated as possible in order to achieve as fast a reaction rate with dioxygen as possible. Results and Discussion In initial experiments, tri-tert-butylphosphine, a commonly known dioxygen-sensitive phosphine, was tested in toluene solution as the dioxygen scavenger in an N2O microsensor. It was however, observed that the outer silicone membrane burst in a matter of minutes, possibly caused by severe swelling of the silicone membrane by toluene. The solvent was hereafter switched to propylene carbonate (PC), which is already used as the electrolyte solvent in the inner electrode chamber. PC is known not to interfere with the silicone membranes. However, tri-tert-butylphosphine was insoluble in PC, which obstructed further experiments with this phosphine candidate. Hence, we started investigating other phosphines that are both highly reactive towards dioxygen, soluble in PC and more benign to the silicone membranes. Many phosphines, particularly commercially available products, were screened for their properties as dioxygen scavengers. Most phosphines turned out to either have a very low solubility in PC, if not completely insoluble, have a very low

reactivity with dioxygen, or both. Eventually, we settled for the phosphines 1-3 (Chart 1) since they all showed high solubility (≥1 M) in PC and proved to be very reactive towards dioxygen.

Chart 1. Phosphines investigated in the current study. Phosphine 1 is commercially available, whereas 2 and 3 are easily prepared in a few steps (See supporting information for details of their synthesis). To assess the reactivity of phosphines towards dioxygen, we used a system as illustrated in Figure 2, where the partial pressure of dioxygen is measured through a solution of the dioxygen scavenger. A dioxygen microsensor (tip diameter 10 µm) was mounted onto a lab stand fitted with a micromanipulator, which enables fine-tuning of the sensor tip placement in the X, Y and Z directions. This micromanipulator is adjustable along the Z-axis by a high precision motor stage capable of 1 µm linear movement. The motor is controlled using a personal computer and specialized software (see experimental section in SI), which also collects data from the dioxygen microsensor.

Figure 2. Schematic of the microprofiling system.

The phosphine, dissolved in PC, is contained in a narrow capillary (Figure 2), which is placed directly below the tip of the microsensor. In this way, the dioxygen partial pressure at any given point along the Z-axis in the phosphine solution can be measured, yielding the dioxygen microprofile. Dioxygen microprofiling The dioxygen profiles for each phosphine and the reference ascorbate solution were normalized with respect to

ACS Paragon Plus Environment

Page 2 of 9

Page 3 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

Figure 3. Dioxygen penetration profiles. Phosphines 1-3 were measured as a 1.0 M solution in PC. Ascorbate was measured as an aqueous and alkaline 0.75 M solution (see text). Top left: Phosphine 1. Top right: Phosphine 2. Bottom left: Phosphine 3. Bottom right: Ascorbate solution as control. Data points indicate mean values and error bars indicate the 95% confidence interval (n = 3, each with three repeated measurements). Some error bars are smaller than the data points and thus invisible.

the sensor signal of ambient dioxygen partial pressure (at x = 0 µm in Figures 2 and 3). Prior to this normalization, the signal at zero dioxygen pressure was subtracted to obtain zero baseline of the normalized data. The signal from zero dioxygen pressure was taken as the average of several data points after the sensor signal converged to a globally low value at deep penetration depths. The reason for the normalization procedure is, that day-to-day variation of sensitivity and zerocurrent is eliminated, thus make comparisons across the candidates more transparent. Figure 3 illustrates the microprofiles for phosphines 1-3 and also the ascorbate solution currently used in the commercial N2O sensor as a reference. Between measurements, the sensor tip could not reproducibly be placed at exactly the same distance from the liquid surface. This resulted in microprofiles that were similar in shape but shifted parallel in the x-axis direction. All profiles were therefore manually aligned such that the last data point prior to sensor penetration into the liquid (corresponding to ambient dioxygen partial pressure) was set to x = 0 µm. The penetration of the liquid surface is clearly indicated by a steep decrease in the dioxygen signal, which might be noted in Figure 3. From the microprofiles, the dioxygen partial pressure is indistinguisha-

ble from zero-current within the first 5 to 15 micrometers, compared to 20 to 30 micrometers for ascorbate. This is highly encouraging results regarding the use of phosphines for efficient dioxygen scavengers in Clark-type microsensors. The dioxygen profiles for all the tested phosphines are fitting a simple exponential decay (see SI), which suggests that the oxidation of phosphines in the tested concentration is a first order reaction with respect to dioxygen (see justification in SI). The dioxygen microprofile for the ascorbate reference solution also fits an exponential decay, equally suggesting a (pseudo-)first order reaction with dioxygen under the given conditions. The first data point at x = 0 µm was not included in the fits, since plotting the data in semilog plots revealed this data point to be situated systematically higher than would be expected, given that the rest of the data points followed a linear trend (see SI). This indicated that the dioxygen partial pressure generally declined much more rapidly in the first five micrometers that what is expected from the exponential trend. This discrepancy can be explained by the presence of a diffusional film between the air and liquid interface or by a local disturbance due to the presence of a meniscus created by liquid-glass

ACS Paragon Plus Environment

ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

surface tension. A diffusional film would impose a resistance to diffusion for dioxygen and thus imply a large partial pressure gradient across it, explaining the much lower extrapolated dioxygen partial pressure at x = 0 µm. On the other hand, if the liquid adheres significantly to the sensor tip due to liquidglass surface tension, the sensor tip is actually situated deeper into the liquid compared to a completely flat air-liquid interface as depicted in Figure 2. Such a movement of the liquid might also induce convective mixing. Both effects would yield a dioxygen partial pressure lower than expected. Creation of menisci has previously been observed in our laboratory in other non-related experiments. However, since no such severe discrepancy in our data has been observed in the experiments with ascorbate, it may be reasonable to expect a diffusional layer to have a larger impact than an eventual effect from a meniscus. Additionally, PC is hydrophobic, indicating that menisci created in PC solutions may be smaller than in water, further diminishing the impact of a meniscus. Only few studies on the rate of phosphine autoxidation have been performed, despite the fact that oxidation of phosphines in air is a well-known phenomenon. The few investigations are restricted to either trialkylphosphines,13 oxidation rates under the influence of radical initiators,14,15 15,16 triarylphosphines, or at elevated temperatures.17 No studies on the oxidation of secondary phosphines like 1 under the influence of only atmospheric dioxygen have been encountered. Therefore, it cannot be concluded from the dioxygen microprofiles alone that autoxidation always should follow a pseudo-first order rate. The data, however, fit exponential functions expected by a pseudo-first order rate and therefore further support this hypothesis. Table S-3 summarizes the fitted parameters. The dioxygen microprofiles suggest that the required height of the dioxygen scavenging chamber should be around 15-20 µm when using phosphines and 30-40 µm when using ascorbate. We have chosen to use a distance in the range of 85-115 µm for several reasons. This is the same distance as found in the commercial sensors which we wish to compare our results to. The diffusion of fresh scavenger to the tip is spatially restricted by the dimensions of the microsensor tip as compared to the capillary in which we have measured oxygen penetration. Hence, to counteract lesser efficiency under such spatial restriction, the scavenger chamber dimension is set in the range of 85-115 µm in accordance with previous experience in our laboratory. Two-chamber sensor test Based on reactivity, all phosphines were found suitable for the two-chamber N2O sensor version which constitutes a sensing compartment and an outer dioxygen scavenger compartment, according to our previously reported method.6 The outer compartment was filled with a 1.0 M solution of the phosphine in PC and quickly sealed with UV-setting resin. The sensor was allowed to deoxygenate electrochemically at a polarization potential of -1.3 V with respect to the internal Ag/AgI reference electrode for 30 minutes in air in order to remove most residual dioxygen stemming from the sensor fabrication. The polarization was then lowered to -0.8 V and the sensor was allowed to stand until a stable background signal was reached, typically within a few hours.

Sensors constructed with phosphines 2 and 3 had an unacceptably high background signal (about 40-100 pA) and had a rapid declining sensitivity towards N2O. These characteristics are presumably due to trace amounts of electroactive impurities together with deposition of the phosphine oxide on the membrane surface, which quickly hindered the diffusion of N2O towards the working electrode. Further work with 2 and 3 was therefore ceased. Insensitivity towards dioxygen was previously observed in sensors constructed with 1, as the baseline signal did not change between alternately inserting the microsensor into dioxygen-free ascorbate solution and then withdrawing it again into air (results not published). This was strongly supported that the constructed sensors with 1 as dioxygen scavenger has a dioxygen sensitivity below the detection limit, like current sensors with ascorbate as dioxygen scavenger. The silicone membrane was somewhat impacted by the solution of 1, but the swelling stabilized within few hours and no rupture of the membrane could be observed under the microscope. A dark band in the middle of the silicone membrane was present, presumably due to precipitating phosphine oxide within the membrane itself. This, however, did not seem to affect the sensor performance significantly during the subsequent experiments. Previous experiments with 1 in two-chambered sensors showed a decrease of sensitivity from about 430 pA/100 µM N2O to about 26 pA/100 µM N2O over a period of 225 days. In the same period, the baseline decreased from about 14 pA to 3.7 pA at day 38, and then slowly increased to 6 pA at day 225. The decline of sensitivity is ascribed to the build-up of blocking layers of phosphine oxide. It shows that such a sensor, in fact, can be useful for the quantification of N2O in up to about seven months. We therefore do not find it immediately reasonable to conclude that the lifetime is any less than the commercial ascorbate-based sensors. The slow increase of the baseline after day 38 is typical and also observed for the commercial sensors, and therefore not necessarily caused by phosphine oxidation products. The sensors constructed with 1 were tested for N2O sensitivity, linearity and response time by immersing the sensor into 500 mL deionized water at room temperature. The water was spiked with N2O from a saturated N2O solution, under gentle stirring, such that the total concentration in the water was increased stepwise from 0 µM to 100 µM in 25 µM steps. Figure 4 illustrates the response curve for a single sensor based on 1 and this behavior was highly reproducible in repeated experiments. The response was found to be perfectly linear in the range of N2O concentrations tested (Figure 4, right). For comparison, the same experiment was also performed with three reference sensors built with ascorbate solution as dioxygen scavenger. A summary of the results is presented in Table 1. The sensor response time was determined by the difference in time between the 90% of the steady-state signal at 25 µM concentration (corrected for baseline signal) and where the signal just began to deflect from the baseline current. This method does induce a small bias of the response time, since no signal will be detected between time of addition and until N2O reaches the working electrode. The method is convenient, though, and the bias should have a minimal influence when comparisons are made

ACS Paragon Plus Environment

Page 4 of 9

Page 5 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

Figure 4. Left: Response curve for a single sensor based on phosphine 1. The aliquots of saturated N2O solution was administered when the signal after previous addition had stabilized. After the last aliquot, the sensor was withdrawn from the water. Right: Fitted calibration curve from the same data.

between different sensor versions. Due to the small sample size (n = 3), a high variation of the baselines were observed for both sensor types. As such, there is no statistically difference between the baseline signals in the two sensor types. In order to avoid too much fouling of the silicone membranes in sensors with 1, experiments commenced immediately after a stable baseline signal (change within 1 pA in 15 minutes) following initial polarization procedure. For fair comparison, a stable baseline signal in sensors with ascorbate was also taken as a change within 1 pA in 15 minutes time frame. The baseline signal and its variation in the constructed sensors with ascorbate is, however, higher and not directly comparable with the commercially available microsensors from Unisense A/S. This is due to the longer preconditioning/polarization times and strict quality control for the commercial microsensors, the latter of which has great impact on the e.g. the confidence intervals on both response time and baseline signal. Table 1. Dioxygen scavenger properties in constructed N2O sensors. a Entry

a (pA/µM)

b (pA)

R2

1

5.052 ± 0.176

12.2 ± 10.8

0.9996

Ascorbate

2.455 ± 0.062

6.042 ± 3.784

0.9998

Response time t90 (s)

Baseline (pA)

1

62 ± 30

11.03 ± 10.67

Ascorbate

36 ± 22

4.45 ± 10.42

a

The model fitted is f(x) = ax + b. The uncertainties are the 95% confidence interval (n = 3).

The results show that although the N2O microsensors with phosphine 1 as the dioxygen scavenger have almost twice as long a response time as the current N2O microsensors, their sensitivity have also almost doubled (Table 2). The long response time cannot be explained by different sensor dimensions, as the total diffusion paths are identical for the two sensor types: 491.5 µm for sensors with 1 and 479.6 µm for sensors with reference ascorbate solution (also see Table S-1).

The current is proportional to the diffusion coefficient and the concentration gradient at the electrode surface (eq 1)18

 ∝  

 

(1)

where  is the measured current at any given time,  is the diffusion coefficient of N2O,  is the concentration of dissolved N2O, and  is the distance from the electrode surface to bulk. However, the diffusion of N2O is driven by the gradient of its partial pressure . Since the concentration of dissolved gas can be approximated by the product of the partial pressure and Henrys constant  , the current can be rewritten as in eq 2

 ∝  

  





=  

 

 







(2)

As the partial pressure of N2O per definition is zero at the electrode surface (the sensor is polarized enough to ensure fast and complete reaction), the final expression is (eq 3)

 ∝  

 

(3)

From eq 3 it can be deduced that the diffusion coefficient, the Henry constant and the partial pressure gradient at the electrode surface will influence the steady-state current. The inner compartments are identical between the two-chamber sensors with 1 and with ascorbate, which would give rise to an identical  at the electrode surface. Thus, the only sensible conclusion to the observed increased sensitivity for phosphine sensors is that the resistance to N2O diffusion in the outer chamber varies significantly from ascorbate versions. This reasoning is also supported by the analytical solution for the signal of a simple dioxygen microsensor.19 An increased steady-state current with a longer time before equilibrium is reached may then be explained by a higher solubility of N2O in the solution

ACS Paragon Plus Environment

ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

of 1, i.e. higher value of  , together with a lower diffusion coefficient. The outer silicone membrane was observed to swell after construction, which lowers the packing density of the silicone polymer. This can lead to an increase in diffusivity through the membrane. However, a build-up of phosphine oxide precipitate may hinder the diffusion of N2O, hence such phosphine oxide precipitation may increase the time for N2O diffusion to the working electrode. The mean baseline signal, or zero-current, is somewhat higher for sensors with phosphine 1 than for the reference ascorbate sensors, but despite this they are not significantly different from each other. It is our experience that the zero-current is very dependent on the batch of 1, which suggests that trace impurities are responsible for the zero-current signal and not the phosphine itself. In other experiments, we have observed zero-currents for sensors with 1 ranging from about half to several times the observed average in this study, where the effect size differs from batch to batch of 1. One-chamber sensor test At the negative potential required to reduce N2O, water would be co-reduced at the working electrode. Therefore, the use of aqueous electrolytes and dioxygen scavengers inside the sensing chamber would be prohibited, since the co-reduction of water would give rise to a massively increased baseline signal. The design of the two-compartment sensor allows a physical separation of the aqueous ascorbate solution from the sensing chamber, where the hydrophobic silicone membrane between the sensing and scavenging chambers reduces water vapor flux into the sensing chamber. The manufacture of sensors with this particular design, however, is laborious and the introduction of a second outer chamber increases the diffusional path for N2O, thus negatively affecting the sensor response time and sensitivity. If a single chambered sensor could be constructed, the total diffusion length of the analyte, in this case N2O, can be significantly reduced. Since the electrochemical signal is proportional to the diffusion gradient at the electrode surface (eq 1), a decrease in diffusion length from the bulk solution to the electrode surface will result in an increased diffusion gradient and hence an enhanced signal. Additionally, the response time will also be significantly decreased as the transport time of the analyte from the bulk solution to the electrode surface is decreased.

Figure 5. Overview of the new one-chamber sensor design. Numbers indicate the dimensions of all constructed onechamber sensors and their 95% confidence intervals (n = 3). Dimensions are not to scale. The diphenylphosphine 1 in this study is soluble in organic solvents and therefore we investigated the use of 1 in a one-

Page 6 of 9

compartment design to evaluate if it interferes negatively with the electrolyte solution or the electrodes. To assess the onecompartment design with phosphine 1 as dioxygen scavenger, three sensors were constructed as previously reported6 but with the outer chamber omitted (Figure 5, see SI for dimensions). The one-chamber sensor is very similar in design to the inner compartment of the two-chamber version, the main difference being that the working electrode is retracted about 90 µm from the silicone membrane. The sensors were filled with 0.5 M of phosphine 1 in electrolyte solution (see SI). The concentration of 0.5 M was chosen for convenience, as the stock solution of 1 M of 1 was mixed with electrolyte in a syringe immediately before sensor filling. It was not expected that this concentration would greatly impact oxygen sensitivity cf. the previous discussion of oxygen penetration depth and scavenger chamber dimensions. As according to eq S-6, the oxygen penetration depth would be increased by approximately 40%. The results are presented in Table 2. Table 2. Dioxygen scavenger properties in constructed single-chamber N2O sensors.a Entry

a (pA/µM)

b (pA)

R2

1

11.13 ± 0.18

43.32 ± 10.71

0.9999

Response time t90 (s)

Baseline (pA)

38 ± 16

30.86 ± 35.38

a The model fitted is     . Uncertainties are the 95% confidence interval (n = 3).

For the one-chamber sensors, the sensitivity was dramatically increased to about 11 pA/µM N2O which is more than twice that of the two-chamber based sensor. Additionally, the response time was almost halved as compared to the twochambered sensors with 1. This decrease in response time is correlated with the significantly shorter distance from the outer membrane to the electrode surface (188.8 µm versus 491.5 µm, see Table S-2) for N2O from the bulk to the cathode. The baseline level, however, was somewhat increased but also with a larger spread. This increase in baseline signal is most probably due to the beforementioned residual impurities in the phosphine. It cannot be ruled out, though, that slight rereduction of phosphine oxides at the electrode surface may happen as well. Only one study has been found on this topic, which reports the reduction potential for triphenylphosphine oxide in DMF as -2.6 V versus the SCE electrode.20 As 1 is more reactive towards oxygen than triphenylphosphine, it is not expected that electroreduction of the oxide of 1 will occur at the polarization in the microsensor. Also, the sensing chamber is slightly more conical than the sensing chamber in the two-chambered design. This may decrease the guard efficiency due to an enlarged volume for the guard to electrochemically clean. As further evidence that accumulated phosphine oxide is probably not the main determinators of the sensor lifetime, the same sensors had a baseline current of 26.70 ± 11.56 pA after 160 days without polarization (in air; 95% confidence interval), which is not worse than the baseline current immediately after construction (Table 2). This result indicates that cathodic reduction of the phosphine oxide is not likely to happen at the polarization used in the experiments.

ACS Paragon Plus Environment

Page 7 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors Phosphine 1 does not seem to interfere negatively with the electrode surface, as could be expected with adsorption of 1 to

the

Figure 6. Left: Response curve for a single one-chamber sensor based on phosphine 1. The aliquots of saturated N2O solution was administered when the signal after previous addition had stabilized. After the last aliquot, the sensor was withdrawn from the water. Right: Fitted calibration curve from the same data.

electrode surface. If 1 adsorbs to the electrode, assuming complete adsorption and no electro-reduction at the given potential, the sensor would be expected to have a much lower or non-existent sensitivity towards N2O, since N2O would not be able to reach the electrode surface. As the sensitivity is observed to be much higher than in two-chambered sensors, it appears likely phosphine 1 does not adsorb significantly during the experiments. The response curve and corresponding calibration curve for a single sensor is illustrated in Figure 6. The results for one-chamber sensors clearly show that it is possible to markedly improve the sensor characteristics by direct incorporation of the phosphine-based dioxygen scavenger inside the sensing compartment and thereby altogether eliminating the hitherto required two-compartment system. Not only is the sensor a much simpler construction, which enhances reproducibility, but it enables the development of N2O sensors with greatly enhanced sensitivity and possibly lower response time. Under any circumstances, the increased sensitivity of one-chambered N2O sensors would be expected to enable even lower detection limits than the current technology allows. Estimated detection limits (visual, based on background current and signal at 25 µM N2O) were found to be down to 40 nM and 70 nM respectively for the two-chamber sensors with 1 and ascorbate, and 20 nM for the single chamber sensor with 1. Conclusion Phosphines have been found useful as dioxygen scavengers in an organic solvent and it was applied in sensor applications where dioxygen acts as an interferent. Sensors containing 1.0 M diphenylphosphine (1) in propylene carbonate were constructed in a two-chambered setup, where the phosphine solution was physically separated from the sensing chamber. The phosphine successfully removed ambient dioxygen whilst doubling the sensitivity as compared to commercial sensors with an alkaline ascorbate solution as dioxygen scavenger. Albeit the response time was nearly doubled as well. Diphe-

nylphosphine was subsequently successfully integrated into a single-chamber N2O microsensor, achieving zero sensitivity towards dioxygen, high analyte sensitivity and short response time. This is, to the best of our knowledge, the first example of a single chambered N2O sensor with zero dioxygen sensitivity. Single-compartment designs are much simpler to produce than previous two-chamber designs, and the simplicity of construction enables sensor construction with greater reproducibility and in shorter time. The sensitivity increased more than four times compared to that of current designs with two chambers and ascorbate as dioxygen scavenger. This also results in a lower detection limit down to 20 nM of the single chamber sensor. A further reduction of the response time can be achieved by moving the working electrode closer to the membrane, thereby decreasing the diffusion path for N2O. A reduced diffusion distance will also further increase sensitivity and lower the detection limit. The usage of phosphines instead of the otherwise efficient ascorbate has several advantages: Phosphines can be used in organic solvents, there is no need for addition of strong bases, and it seems to allow the design of much faster responding N2O sensors with greatly enhanced sensitivity and much lower detection limits. Introduction of ionic liquids as solvents, together with stringent purification of ionic phosphines such as 2 and 3 may very likely resolve issues with precipitation of phosphine oxides as well as greatly reduce the zero-current. Change to chemically resistant membranes such as microporous PTFE can also widen the range of suitable phosphines. In conclusion, phosphines seem to be very promising candidates for dioxygen scavengers in sensor applications where aqueous and/or alkaline conditions are not an option or optimal.

ASSOCIATED CONTENT Supporting Information. Supplementary figures, experimental details, and mathematical derivations. This material is available free of charge via the Internet at http://pubs.acs.org.

ACS Paragon Plus Environment

ACS Sensors 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

AUTHOR INFORMATION Corresponding Author

(7)

* [email protected]

Present Addresses

(8)

† Present company address: Unisense A/S, Tueager 1, DK-8200 Aarhus N, Denmark. ‡ DHI A/S, Finlandsgade 10, DK-8200 Aarhus N, Denmark. (9)

Author Contributions S. G. Sveegaard co-conceived the project, designed and performed syntheses and experiments, performed mathematical derivations, and authored the manuscript. M. Nielsen constructed all sensors. M. Nielsen and M. H. Andersen revised the manuscript and suggested experiments, methods, and synthetic procedures. K. V. Gothelf co-conceived and supervised the project, and co-authored the manuscript. All authors have given approval to the final version of the manuscript.

(10)

(11)

(12)

ACKNOWLEDGMENT

(13)

This work has been partly funded by the Danish Innovation Fund (grant no. 4135-00089B). Thanks to Dr. Lars Hauer Larsen (Unisense A/S) and Dr. Søren Porsgaard (Unisense A/S) for fruitful discussions.

(14)

REFERENCES (1)

(2) (3)

(4)

(5)

(6)

(15)

(16)

IPCC. Climate Change 2001: The Scientific Basis; Houghton, J. T., Ding, Y., Griggs, D. J., Noguer, M., Van der Linden, P. J., Dai, X., Maskell, K. and Johnson, C. A., Ed.; Cambridge University Press: Cambridge and New York, 2001; Vol. 881. Khalil, M. A. K.; Rasmussen, R. A. The Global Sources of Nitrous Oxide. J. Geophys. Res. 1992, 97, 14651. Shen, L.; Guan, Y.; Wu, G.; Zhan, X. N2O Emission from a Sequencing Batch Reactor for Biological N and P Removal from Wastewater. Front. Environ. Sci. Eng. 2013, 8, 776–783. Andersen, K.; Kjær, T.; Revsbech, N. P. An Oxygen Insensitive Microsensor for Nitrous Oxide. Sensors Actuators B Chem. 2001, 81, 42–48. Tsugawa, W.; Shimizu, H.; Tatara, M.; Ueno, Y.; Kojima, K.; Sode, K. Nitrous Oxide Sensing Using Oxygen-Insensitive Direct-Electron-Transfer-Type Nitrous Oxide Reductase. Electrochemistry 2012, 80, 371–374. Larsen, L. H.; Nielsen, M.; Andersen, M. H. An

(17)

(18)

(19)

(20)

Page 8 of 9 Electrochemical Sensor for Sensing Nitrous Oxide. WO 2014101921 A1, July 3, 2014. Plumeré, N.; Henig, J.; Campbell, W. H. Enzyme-Catalyzed O2 Removal System for Electrochemical Analysis under Ambient Air: Application in an Amperometric Nitrate Biosensor. Anal. Chem. 2012, 84, 2141–2146. Plumeré, N. Interferences from Oxygen Reduction Reactions in Bioelectroanalytical Measurements: The Case Study of Nitrate and Nitrite Biosensors. Anal. Bioanal. Chem. 2013, 405, 3731– 3738. Israel, Y. Polarographic Study of Inhibitors and Catalysts for the Reaction of Dissolved Oxygen with Sulphite Ion. Talanta 1967, 14, 925–931. Gu, Y.; Chen, C.-C. Eliminating the Interference of Oxygen for Sensing Hydrogen Peroxide with the Polyaniline Modified Electrode. Sensors 2008, 8, 8237–8247. Larsen, L. H.; Revsbech, N. P.; Binnerup, S. J. A Microsensor for Nitrate Based on Immobilized Denitrifying Bacteria. Appl. Environ. Microbiol. 1996, 62, 1248–1251. Nielsen, M.; Larsen, L. H.; Jetten, M. S. M.; Revsbech, N. P. Bacterium-Based NO2-Biosensor for Environmental Applications. Appl. Environ. Microbiol. 2004, 70, 6551–6558. Buckler, S. A. Autoxidation of Trialkylphosphines. J. Am. Chem. Soc. 1962, 84, 3093–3097. Floyd, M. B.; Boozer, C. E. Kinetics of Autoxidation of Trialkylphosphines. J. Am. Chem. Soc. 1963, 85, 984–986. Ogata, Y.; Yamashita, M. Kinetics of the Autoxidation of Trimethyl Phosphite, Methyl Diphenyl-Phosphinite, and Triphenylphosphine. J. Chem. Soc. Perkin Trans. 2 1972, 730733. Alfassi, Z. B.; Neta, P.; Beaver, B. Radiolytic Studies of the Mechanism of Autoxidation of Triphenylphosphine and Related Compounds. J. Phys. Chem. A 1997, 101, 2153–2158. Beaver, B. D.; Gao, L.; Fedak, M. G.; Coleman, M. M.; Sobkowiak, M. Model Studies Examining the Use of Dicyclohexylphenylphosphine to Enhance the Oxidative and Thermal Stability of Future Jet Fuels. Energy & Fuels 2002, 16, 1134–1140. Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd Ed.; 2.; John Wiley & Sons, Inc: New York, 2001. Gundersen, J. K.; Ramsing, N. B.; Glud, R. N. Predicting the Signal of O2 Microsensors from Physical Dimensions, Temperature, Salinity, and O2 Concentration. Limnol. Oceanogr. 1998, 43, 1932–1937. Santhanam, K. S. V.; Bard, A. J. Electrochemistry of Organophosphorus Compounds. II. Electroreduction of Triphenylphosphine and Triphenylphosphine Oxide. J. Am. Chem. Soc. 1968, 90, 1118–1122.

ACS Paragon Plus Environment

Page 9 of 9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sensors

TOC graphic

ACS Paragon Plus Environment

9