Influence of an active vibration isolator and electrochemical cell design on electrochemical measurements to minimize natural convection

Accepted Manuscript Influence of an active vibration isolator and electrochemical cell design on electrochemical measurements to minimize natural convection

Kyungsoon Park, Eunsung Kim, Jun Hui Park, Seongpil Hwang PII: DOI: Reference:

S1388-2481(17)30200-X doi: 10.1016/j.elecom.2017.07.017 ELECOM 5992

To appear in:

Electrochemistry Communications

Received date: Revised date: Accepted date:

26 April 2017 24 July 2017 25 July 2017

Please cite this article as: Kyungsoon Park, Eunsung Kim, Jun Hui Park, Seongpil Hwang , Influence of an active vibration isolator and electrochemical cell design on electrochemical measurements to minimize natural convection, Electrochemistry Communications (2017), doi: 10.1016/j.elecom.2017.07.017

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.

ACCEPTED MANUSCRIPT

Influence of an active vibration isolator and

PT

electrochemical cell design on electrochemical

SC

RI

measurements to minimize natural convection

†

MA

NU

Kyungsoon Park,†,§, Eunsung Kim, †,§ Jun Hui Park‡* and Seongpil Hwang†,*

Department of Advanced Materials Chemistry, Korea University, Sejong 30019, Republic of

Department of Chemistry Education and Institute of Fusion Science, Chonbuk National

PT E

‡

D

Korea

University, Jeonju 54896, Korea

AC

CE

§ equal contribution

Corresponding Author *E-mail: [email protected]; [email protected]

1

ACCEPTED MANUSCRIPT

Abstract

The influence of an active vibration isolator (AVI) and electrochemical cell design on natural convection was investigated. The natural convection of an electrolyte caused by external vibrations and acoustic sounds is reduced by both an AVI and an acoustic enclosure.

PT

Slow scan voltammetry with an AVI shows a reproducible and predictable diffusion-

RI

controlled current down to 0.4 mV/s. Reducing the distance between the working electrode

NU

SC

and the glass wall also has a considerable effect on the voltammetric signal.

AC

CE

PT E

D

enclosure, Electrochemical cell design

MA

Keywords : Natural convection, Slow scan voltammetry, Active vibration isolation, Acoustic

2

ACCEPTED MANUSCRIPT

1. Introduction The voltammetric [1] or chronoamperometric [2] response predicted by the Randles– Sevcik equation, Shoup and Szabo approximation [3] or the Cottrell equation [4] is the fundamental basis for modern electrochemistry, which assumes diffusional mass transport

PT

without convection or migration. This hypothesis is valid in conventional electrochemical experiments due to negligible migration by the excess supporting electrolyte and the

RI

stationary electrode/electrolyte. For well-defined forced convection, the Koutecký-Levich

SC

equation [5] for a rotating disk electrode (RDE) [6] and the theory of electrochemical flow cells [7,8] offer an insight into electrochemical behavior. Over a longer time scale, however,

NU

for example when using a slow scan rate or a long electrolysis period, uncontrollable factors

MA

generate motion in the solution, called natural convection, which causes a deviation from theoretical results as well as poor reproducibility. Possible contributors to this motion include:

PT E

D

(1) The density gradient during electrolysis can drive the gravity force, resulting in currents dependent on the electrode direction [9]. The Compton group reported a quantitative analysis of this effect, based on a simulation [10].

CE

(2) Thermal convection caused by the heat of a reaction might influence convection.

AC

However, a recent simulation showed a negligible thermal effect on convection [11]. (3) Mechanical vibrations, the movement of air and the cell configuration might induce convection. The Amatore group developed a model for this effect in which the thickness of the limiting diffusion layer is an important parameter [12]. Previous work has been focused on the origin of natural convection, with only a few reports

aiming to

minimize

natural

convection

in

electrochemical

experiments.

Electrochemical behaviour over longer time scales has attracted increasing interest from 3

ACCEPTED MANUSCRIPT

researchers in reaction mechanisms, enhanced sensitivity, and higher resolution scanning electrochemical microscopy (SECM) [13], where unprecedented movement control in experiments is required. Two of the causes of natural convection, mechanical vibration and acoustic motion, can be minimized under well-controlled conditions. Recently, the Aoki

PT

group reduced mechanical convection through enhanced viscosity by adding a gelling agent [14]. Our group used a hydrogel to help reduce the motion in a previous study [15]. To the

RI

best of our knowledge, however, there have been no attempts to minimize the mechanical

SC

motion in a typical electrochemical cell in the absence of additional chemicals which may influence the electrochemistry at the interface or the mass transport within the solution.

NU

The objective of our work was to minimize the vibrational and acoustic motion of the

MA

electrolyte in natural convection. Firstly, an active vibration isolator (AVI) was used to control and reduce the vibration precisely. While a passive vibration isolation system such as an optical table can reduce vibrations above 10 Hz, AVI can work down to 0.5 Hz with

PT E

D

enhanced performance. It can cancel out external vibrations using accelerometer-controlled actuators, and is used in atomic force microscopes (AFM), interferometers [16] and other devices. Secondly, an acoustic enclosure served as a faraday cage to reduce acoustic

CE

vibrations. Thick (~10 mm) plates of metal were assembled into a box, and soundproof foam

AC

panels covered the inside, as shown in Scheme 1(A). A disk electrode and 1 mM K4Fe(CN)6 aqueous electrolyte were used to keep the density-gradient convection low [10]. Thirdly, the distance of the disk electrode from the wall of the glassware was adjusted. We speculate that the velocity of the laminar flow parallel to the plates in the middle of the gap is slower when the gap is smaller, as shown in Scheme 1(B). Thus, natural convection could be minimized by decreasing the gap between the disk electrode and the glass wall. The precision with which the distance is controlled is of the order of mm, using the simple setup shown in Scheme 1(A). 4

ACCEPTED MANUSCRIPT

2. Experimental

PT

All chemicals (obtained from Aldrich) were used as received. The polished Pt disk electrode (1.6 mm in diameter, BAS Inc) was fixed vertically with a ruler (Scheme 1A) within a

RI

homemade faraday cage/acoustic enclosure on an active vibration isolator (DVIA-T67, Daeil

SC

Systems, Korea). The gap distance was adjusted using a support jack. The aqueous electrolyte contained 1 mM Fe(CN)64-and 0.5 M KCl in all experiments. Electrochemical experiments

NU

were performed with a CHI 730E after 1 hour to allow the natural convection to settle down.

MA

A Pt wire and Ag/AgCl (sat’d KCl) served as the counter and reference electrode, respectively. To investigate the effect of vibration, an electric toothbrush (Oral-B DB4010,

D

Braun) acted as a vibration source, and the vibrations were measured by a VM-6360 vibration

PT E

meter (Landtek Instruments, China).

CE

3. Results and Discussion

AC

Firstly the effect of the gap size in the electrochemical cell was investigated. The Pt disk electrode was positioned at 0.5, 1 or 2 mm from the wall of the glass beaker. Cyclic voltammetry was performed in the absence of both the AVI and the acoustic enclosure. Figure 1(A) shows the cyclic voltammograms (CVs) obtained with a gap of 2 mm. While conventional CVs were observed at a scan rate over 3 mV/s, the CVs observed with a scan rate under 1 mV/s showed (1) a sigmoidal shape, (2) very poor reproducibility in a pseudosteady-state current, and (3) independence of the scan rate, assumed to be caused by natural 5

ACCEPTED MANUSCRIPT

convection [12]. With a smaller gap of 1 mm, predictable behavior was observed with scan rates as low as 0.7 mV/s. At a scan rate of 0.5 mV/s, however, a fluctuating pseudo steadystate current was observed (Figure 1(B)). CVs taken with a gap of 0.5 mm, shown in Figure 1(C), show conventional CVs for scan rates down to 0.7 mV/s and then change to a sigmoidal

PT

shape at 0.5 mV/s. The peak currents depend on v1/2, as predicted by the Randles-Sevcik’s equation, down to 0.7 mV/s, indicating the absence of natural convection. The natural

RI

convection is reduced by the disk electrode approaching the glass wall, probably due to the

SC

slower speed of the laminar flow as shown in Scheme 1(B). The higher ratio of the stationary area over the electrolyte volume within the thin layer between the electrode and the wall may

NU

cause higher frictional forces.

MA

To study the effect of vibration on the electrochemistry, the CVs were measured using electrical toothbrushes as a vibration source. These were positioned near the electrochemical cell; the amplitude of the vibration could be adjusted by changing the number of toothbrushes

PT E

D

and their distance from the cell. A vibration meter in acceleration mode measured the amplitude of the vibration in the electrochemical cell with a 1 mm gap. Figures 2(A) and 2(B) show the CVs measured with vibrations of 1 and 3 m/s2, respectively. The CV taken at a scan

CE

rate of 0.7 mV/s appears to be more sigmoidal in shape than the CVs recorded at the same

AC

scan rate in Figure 1(B) and Figure 3(A). The plots of v1/2 vs Ip shown in the insets also demonstrate that the deviation from the linear relationship starts at ca. 0.7 mV/s and that the deviation at 0.5 mV/s increases from 0.54 A at 1 m/s2 to 0.63 A at 3 m/s2. A further investigation might be helpful to understand the effect of the vibration on the electrochemistry, which is beyond the scope of this work. Overall, the CVs recorded at a slower scan rate are more affected by the external vibration, and the shorter distance from the wall reduces natural convection. 6

ACCEPTED MANUSCRIPT

Next, we investigated the effect of the acoustic enclosure and AVI at 1 and 0.5 mm gaps from the wall. Figure 3(A) shows the conventional CV shape down to a scan rate of 0.5 mV/s, where the effect of the natural convection is inevitable in the absence of the AVI as shown in Figure 1(B). Although the CVs show irreproducible and higher quasi-steady state

PT

currents below 0.5 mV/s, AVI reduces the effect of the natural convection. With a shorter distance of 0.5 mm, the CVs seem to be more reproducible, even at slow scan rates (Figure

RI

3(B)). The plot of the peak currents versus v1/2 at the 0.5 mm gap (Figure 3(C)) shows the

SC

predicted linear behaviour down to ca. 0.4 mV/s. At slower scan rates of 0.2 mV/s and 0.1 mV/s, the peak currents are higher than the predicted values. It is also worth mentioning that

NU

the quasi-steady-state currents at slow scan rates seem to be random and independent of the

MA

scan rates. Two effects may influence the voltammograms. Firstly, the edge effect of diffusion is not negligible at slow scan rates because both the thickness and shape of the diffusion layer on a macroelectrode are similar to radial diffusion on an ultramicroelectrode.

PT E

D

Assuming that the diffusion coefficient of ferrocyanide is 0.67  10-5 cm2/s [17], the thicknesses of the diffusion layers calculated from the square root of D  t are ca. 0.82, 1.16, and 1.42 mm for 1000, 2000 and 3000 s, respectively. These thicknesses are comparable with

CE

the 1.6 mm diameter of the disk electrode, indicating that radial diffusion must be considered

AC

similar to that of an ultramicroelectrode. Secondly, the thickness of the diffusion layer is similar to the gap thickness, which would hinder the diffusion of chemicals. I vs t-½ chronoamperometry plots in the presence of AVI (Figure 3D) show a linear relationship for longer electrolysis times, according to the Cottrell equations, indicating a decrease in natural convection. Deviations from linearity provide an indication as to whether natural convection or the space confinement effect is the major contributor to the system [12]. Figure 3(D) shows the plots of I vs 1/time for different gaps in the presence of AVI. The 7

ACCEPTED MANUSCRIPT

divergence of the current from the linear trend for a 0.5 mm gap represents the major contribution of confined space, while the steady-state current at larger gaps (≥ 1 mm) represents the contribution of natural convection. Unfortunately, a theoretical explanation for our configuration using a macroelectrode cannot be obtained from classical electrochemical

PT

theory, which assumes planar diffusion under a relatively short electrolysis time. The finite

RI

element method offers a numerical simulation of the ideal response in the absence of natural

SC

convection. The Comsol Multiphysics 5.1 software with electrochemistry module was employed to study the ideal behaviour of the system. The Butler-Volmer formalism was

NU

adopted to consider the electrode kinetics. The charge transfer coefficient and the heterogeneous rate constant used in the Fe(CN)64-/ Fe(CN)63- reaction at a Pt electrode were α

MA

= 0.5 and k0 = 0.1 cm/s, respectively. The electrode potential swept for cyclic voltammetry is ±0.2 V vs E0. The temperature was set to 293.15 K. The diffusion coefficients of Fe(CN)64-/

D

Fe(CN)63- were 0.67×10-5cm2/s and 0.76×10-5cm2/s. The simulation was performed by

CE

be expressed as follows:

PT E

solving Fick’s second law in 2D-axial symmetry. Diffusion of Fe(CN)64- in the solution can

ci    ( Di ci ) t

( Eq. 1. )

AC

Figures 4(A) and 4(C) show the concentration profiles on the disk electrode in free space and in a confined space (0.5 mm gap), respectively. At a slow scan rate of 0.09 mV/s, the diffusion layer thickness of ca. 3 mm is comparable to the diameter of the electrode so that the diffusion-limited current is determined by radial diffusion, which is similar to that on an ultramicroelectrode in free space. It is hard to produce this situation under conventional experimental conditions due to the presence of uncontrolled natural convection. Adjusting the distance between the disk electrode and the glass wall in the presence of both an acoustic 8

ACCEPTED MANUSCRIPT

enclosure and AVI successfully minimizes natural convection down to a time scale of 0.4 mV/s without requiring the addition of any chemicals. Nonetheless, CVs at scans slower than 0.3 mV/s may show the effect of hindered mass-transportation as the edge of the diffusion layer approaches the wall (Figure 4(C)). While the COMSOL simulations showed that the

PT

peak currents in this region have lower values than the values predicted from the RandleSevcik equation shown in Figure 4(B), our experiments had higher values, as shown in Figure

RI

3(C). We are unsure about the explanation of this phenomenon; however, we speculate that

SC

an inhomogeneous glass surface that deviates from an ideal flat surface would affect natural

MA

NU

convection in practical conditions.

D

4. Conclusion

PT E

The influence of an active vibration isolator and electrochemical cell design on natural convection was investigated. The CVs on the Pt disk electrode show reduced natural

CE

convection when there is a smaller distance between the glass wall and the disk electrode. AVI with an acoustic enclosure minimized natural convection down to scan rates of 0.4

AC

mV/s. This well-controlled electrochemical environment may result in a predictable voltammetric response at slow scan rates.

ACKNOWLEDGMENT

9

ACCEPTED MANUSCRIPT

This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) grant funded by the Korea Government (MSIP) (No. NRF-2017R1A2B4012056, NRF-2014R1A2A1A11053268) and by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the

RI

PT

Ministry of Education (No. NRF-2014R1A6A1030732, NRF-2016R1D1A1B03931670).

SC

REFERENCES

NU

[1] N.G. Tsierkezos, Cyclic voltammetric studies of ferrocene in nonaqueous solvents in the temperature range from 248.15 to 298.15 K, J. Sol. Chem. 36 (2007) 289-302.

MA

[2] R.J. Forster, A.J. Kelly, J.G. Vos, M.E.G. Lyons, The effect of supporting electrolyte and temperature on the rate of charge propagation through thin-films of [Os(bipy)2PVP10Cl]Cl

D

coated on stationary electrodes, J. Electroanal. Chem. 270 (1989) 365-379.

PT E

[3] D. Shoup, A. Szabo, Chronoamperometric current at finite disk electrodes, J. Electroanal. Chem. 140 (1982) 237-245.

CE

[4] F.G. Cottrell, The cut off current in galvanic polarisation, considered as a diffusion problem, Zeitschrift für Physikalische Chemie - Stochiometrie und Verwandtschaftslehre 42

AC

(1903) 385-431.

[5] J. Koutecky, B.G. Levich, The application of the rotating disc electrode to studies of kinetic and catalytic processes, Zh. Fiz. Khim. 32 (1958) 1565-1575. [6] X.P. Gao, H.S. White, Rotating microdisk voltammetry, Anal. Chem. 67 (1995) 40574064.

10

ACCEPTED MANUSCRIPT

[7] A.J. Bard, L. R. Faulkner, Electrochemical Methods: Fundamentals and Applications, Wiley, New York, 2001, pp. 448. [8] N.V. Rees, J.A. Alden, R.A.W. Dryfe, B.A. Coles, R.G. Compton, Voltammetry under high-mass transport conditions - the high-speed channel electrode and heterogeneous

PT

kinetics, J. Phys. Chem. 99 (1995) 14813-14818. [9] X.P. Gao, J. Lee, H.S. White, Natural-convection at microelectrodes, Anal. Chem. 67

RI

(1995) 1541-1545.

SC

[10] K. Ngamchuea, S. Eloul, K. Tschulik, R.G. Compton, Advancing from rules of thumb: quantifying the effects of small density changes in mass transport to electrodes.

NU

Understanding natural convection, Anal. Chem. 87 (2015) 7226-7234.

MA

[11] J.K. Novev, S. Eloul, R.G. Compton, Influence of reaction-induced thermal convection on the electrical currents measured in chronoamperometry and cyclic voltammetry, J. Phys. Chem. C. 120 (2016) 13549-13562.

PT E

D

[12] C. Amatore, S. Szunerits, L. Thouin, J.S. Warkocz, The real meaning of Nernst's steady diffusion layer concept under non-forced hydrodynamic conditions. A simple model based on Levich's seminal view of convection, J. Electroanal. Chem. 500 (2001) 62-70.

CE

[13] B.B. Katemann, A. Schulte, W. Schuhmann, Constant-distance mode scanning

AC

electrochemical microscopy. Part II: High-resolution SECM imaging employing Pt nanoelectrodes as miniaturized scanning probes, Electroanalysis 16 (2004) 60-65. [14] B. Wang, K.J. Aoki, J.Y. Chen, T. Nishiumi, Slow scan voltammetry for diffusioncontrolled currents in sodium alginate solutions, J. Electroanal. Chem. 700 (2013) 60-64. [15] H. Kang, S. Hwang, Electrochemical characterization of the hydrophobic interaction and the natural convection within agarose gel, Int. J. Electrochem. Sci. 10 (2015) 9706-9713.

11

ACCEPTED MANUSCRIPT

[16] D.B. Newell, S.J. Richman, P.G. Nelson, R.T. Stebbins, P.L. Bender, J.E. Faller, J. Mason, An ultra-low-noise, low-frequency, six degrees of freedom active vibration isolator, Rev. Sci. Instrum. 68 (1997) 3211-3219. [17] S.J. Konopka, B. McDuffie, Diffusion coefficients of ferri- and ferrocyanide ions in

PT

aqueous media, using twin-electrode thin-layer electrochemistry, Anal. Chem. 42 (1970)

RI

1741-1746.

AC

CE

PT E

D

MA

NU

SC

.

12

ACCEPTED MANUSCRIPT

PT

Scheme 1. Illustrations of (A) the setup with AVI and acoustic enclosure in place and (B) velocity distribution of laminar flow. The speeds at stationary plates are zero. The red arrows

SC

RI

represent the velocity of the electrolyte parallel to the plates.

NU

Figure 1. Cyclic voltammograms at a Pt disk which is positioned at a distance (A) 2 mm, (B)

MA

1 mm, and (C) 0.5 mm from the wall of glassware, in the absence of an active vibration isolator. The indicated numbers are scan rates in mV s-1. Inset in (A) and (B) are plots of peak

PT E

D

currents against v1/2.

Figure 2. CVs at a Pt disk at a distance 1 mm from the wall under constantly applied external

CE

vibration with acceleration (A) 1m/s2 and (B) 3 m/s2. The indicated numbers are scan rates in

AC

mV s-1. Inset are plots of peak currents against the square root of the scan rate.

Figure 3. CVs at a Pt disk at a distance (A) 1 mm and (B) 0.5 mm with both AVI and an acoustic enclosure present. The indicated numbers are scan rates in mV s-1. (C) Plot of peak currents against v1/2 of (B). (D) I vs t

-1/2

plot of chronoamperograms measured at different

distances from the wall of the glassware. The indicated numbers are distances (mm).

13

ACCEPTED MANUSCRIPT

Figure 4. Simulation results of (A) concentration profile during voltammetry at 0.09 mV/s in free space. (B) Peak current density plotted versus the square root of the scan rate in free space and in a confined space. Concentration profile during voltammetry in a confined space

AC

CE

PT E

D

MA

NU

SC

RI

PT

at a scan rate of (C) 0.1mV/s and (D) 0.1V/s.

14

AC

CE

PT E

D

MA

NU

SC

Scheme 1

RI

PT

ACCEPTED MANUSCRIPT

15

Figure 1

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

16

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

Figure 2

17

Figure 3

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

18

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

Figure 4

19

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

D

MA

NU

SC

Graphical abstract

20

ACCEPTED MANUSCRIPT

Highlights

CE

PT E

D

MA

NU

SC

RI

PT

Vibration affects voltammetry at slow scan rates. Use of an active vibration isolator with an acoustic enclosure reduced natural convection. Reducing the gap between the electrode and the cell wall also reduced natural convection.

AC

  

21