Effects of Electrodeposition Conditions and Protocol on ...€¦ · 28.07.2008 · Effects of Electrodeposition Conditions and Protocol on the Properties of Iridium Oxide pH Sensor

Journal of The Electrochemical Society, 156 �1� F1-F6 �2009� F1

Effects of Electrodeposition Conditions and Protocolon the Properties of Iridium Oxide pH Sensor ElectrodesHeather A. Elsen, Christopher F. Monson, and Marcin Majda*,z

Department of Chemistry, University of California, Berkeley, Berkeley, California 94720-1460, USA

The properties of iridium oxide pH sensors produced by electrochemically induced deposition on gold electrodes were examinedas a function of the composition of the deposition solution, as well as the electrochemical deposition protocol. The composition ofthe Ir�IV� deposition solutions, which included oxalate or ethylene diamine tetraacetic acid complexing agent or no complexingagent, had no effect on the slope of the calibration curves. The slope of the calibration curves was shown to increase from ca.49 to 76 mV/pH unit with the fractional coverage of gold substrates with iridium oxide. Increasing the film thickness beyond thefull coverage did not further increase the slope of the calibration curves but resulted in a progressive increase of their interceptvalues. The method of deposition, which involved a constant current, single potential pulse, alternating potential pulse, or cyclicpotential protocol, affected the maximum rate of pH response as well as the capacitance of the iridium oxide sensors. The lattertwo properties of the sensors were investigated using a microelectrochemical time-of-flight method with galvanostatic protongeneration and potentiometric sensing. The alternating potential pulse and cyclic potential methods produced films of smaller rateof pH response and of smaller capacitance relative to the iridium oxide films of the same thickness produced by the other twomethods. This is likely due to a smaller microscopic porosity of the films prepared by the potential pulse and cyclic potentialmethods. The maximum rate of pH response obtained with 50 nm thick iridium oxide sensors varied from ca. 7 to 23 V/s. Thespecific capacitance of the iridium oxide films varied from ca. 900 to 9000 F/cm3.© 2008 The Electrochemical Society. �DOI: 10.1149/1.3001924� All rights reserved.

Manuscript submitted July 28, 2008; revised manuscript received September 22, 2008. Published November 3, 2008.

0013-4651/2008/156�1�/F1/6/$23.00 © The Electrochemical Society

Determining the pH of solutions is important in many fields,from industry to basic chemistry and biology. The most widespreadmethod of measuring the pH of a solution is using a pH electrode.Most pH electrodes are glass membrane electrodes. However, theseelectrodes are not readily amenable to miniaturization, and thusother materials are being widely investigated for use in miniature,lithographically fabricated devices involving pH sensing.1,2 Amongthe most popular alternatives is iridium oxide.

Iridium oxide sensors have many advantages; they have fast re-sponse times and are not prone to chemical interferences. They arestable over long time periods and are relatively inert.3 They havebeen shown to function under high pressure, at high temperatures, innonaqueous solutions, and even in media containing HF.4-10 Theyare sufficiently innocuous to be used in in vivo applications.11,12

Finally, they can be prepared electrochemically on noble metal elec-trodes, making these sensors ideal for microfabricated devices.

Despite these advantages, there are some drawbacks and consid-erable inconsistencies in the literature regarding iridium oxide.Some of the drawbacks can include redox-induced drift �e.g., “oxy-gen drift”� and a somewhat unpredictable response ranging from59 to 90 mV/pH unit. Many of these inconsistencies can be at leastpartially reconciled by understanding the variation within differenttypes of iridium oxide sensors. These can be classified as one of twotypes: hydrated or anhydrous. Anhydrous or, more correctly, lowhydration iridium oxide sensors, are denser and more crystalline,and are usually created in procedures involving heat-treatment orsputtering processes. These electrodes generally exhibit a Nernstianresponse of �59 mV/pH unit. Because of the high temperaturesrequired to create them �600–800 K�, anhydrous iridium oxide sen-sors are not suitable in many microfabricated devices, with the ex-ception of sputtered devices.13-15 sputtered iridium oxide films�SIROFs� can be fabricated into metal oxide semiconductor field-effect transistor-type devices to overcome some of the redox driftand produce a standard 59 mV/pH unit response.16 Unfortunately,there are several instances where SIROFs are not practical to usedue to cost or device compatibility �particularly when proteins orlipid layers are incorporated�.

Hydrated iridium oxide sensors, often called anodic iridium ox-ide films �AIROFs�, are formed in aqueous media. This results inamorphous films with varying hydration and relatively high internal

* Electrochemical Society Active Member.z E-mail: [email protected]

Downloaded 30 Mar 2009 to 128.32.205.142. Redistribution subject to E

microporosity and, thus, low densities. There is a large varietywithin hydrated iridium oxide layers in terms of their pH response�60–90 mV/pH unit�, stability, and compatibility with microfabrica-tion devices.17 Additionally, a variety of deposition solutions con-taining iridium oxide complexes and numerous experimental proce-dures creating AIROFs from iridium solutions are reported in theliterature.18-23

Two types of deposition solutions can be found in the literature:those that do and do not involve oxalate anions as ligands complex-ing Ir�IV�. Oxalate-based deposition solutions are the most com-monly used and were introduced by Yamanaka18 in 1989 and latermodified by Petit and Plichon.24 In both procedures, an aqueousiridium chloride salt, either iridium �IV� tetrachloride �Yamanaka�,or potassium hexachloroiridate �III� �Petit� are the starting materials.Oxalate is then introduced in a 10:1 �Yamanaka� or 5:1 �Petit� molarratio relative to the iridium species and serves to increase the stabil-ity of the iridium complex. The solution is made basic �pH 10–11�using potassium carbonate and allowed to develop over at least twodays, changing from a light yellow or green to a deep blue. Onceprepared, the solution is stable for several months. Deposition iscarried out via an oxidative process, usually following a constantcurrent or a constant potential protocol. The proposed mechanism ofIrO2 deposition is oxidation of the carbon-carbon bond in the ox-alate ligand, producing insoluble iridium �IV� oxide and carbon di-oxide. There is some support for this because carbon dioxide hasbeen observed as a product of the decomposition of tris-oxalatoiridate �IV�, but there are no confirmed reports of mono ox-alate complexes, as Yamanaka claims to have made, only bis and triscomplexes.

There are two principal examples of deposition solutions madewithout oxalate ligands. Yoshino et al. used a mixture of sulfana-toiridate complexes, predominantly anhydrous iridium �III� sulfate,which is dissolved in deionized �DI� water to saturation and left tostir for a week.25,26 A color change is observed from pale green todeep blue, which is consistent with the conversion of a polynuclearpolysulfanoiridate complex to an aquasulfanoiridate.27 Using thissolution, the electrodeposition involved a reduction followed by anoxidation step. Other authors who used this procedure commentedthat the solution was unstable after deposition.18

Baur and Spaine introduced a similar deposition solution.19 Theybegan with either iridum �III� hexachloride or iridium �IV�hexachloride, which they reduced to iridium �III� under ethanol flux.A color change from light olive green to light brown for the iridium�III� complex and dark reddish brown to light brown for the iridium

CS license or copyright; see http://www.ecsdl.org/terms_use.jsp

F2 Journal of The Electrochemical Society, 156 �1� F1-F6 �2009�F2

�IV� complex was observed. This corresponds to the substitution oftwo water molecules for two chlorides ions, forming a di-aquocomplex.28 In order to prepare the deposition solution, the pH isbrought to �13 with sodium hydroxide, which is supposed to con-vert the iridium complex into a soluble hydrated iridium �III� oxide.Once the base has been added, the solution must be used immedi-ately because it deteriorates in hours. According to the authors, anyoxidation agent will convert the iridium �III� to iridium �IV� and thehydrated iridium oxide will precipitate out. However, hydrated iri-dium �III� complexes are generally less soluble than iridium �IV�complexes.29 Additionally, many aquo/hydoxyl complexes, includ-ing hexahydroxyiridate �III� and �IV� and hexaquoiridate �III� and�IV�, are known to form and remain soluble in alkaline solutions,with the exception of nonhydrated iridium �III� trihydroxides, whichraise some doubts as to the reported deposition mechanism.27,30-32

In addition to several deposition solutions, multiple depositionprocedures have been reported. Yamanaka and Petit both used con-stant current deposition methods, with current densities rangingfrom 35 �A/cm2 to 0.3 mA/cm2, and deposition times rangingfrom 60 to 10,000 s.18,24 Alternatively, constant potential proce-dures can be envisioned, and various such procedures were tested.Baur and Spaine,19 Yoshino et al.,25,26 and others have used pulsedpotential depositions, switching between potentials at which deposi-tion occurs rapidly �0.6 V vs saturated calomel electrode �SCE� orhigher� and more negative potentials �−0.5 V or more negative�with pulse lengths and numbers varying widely, depending on theelectrode being prepared. Similar to pulsed depositions are potentialsweep methods, or the cyclic growth deposition procedures that in-volved similar potential limits to those used in the pulsed methods.Overall, interest in different deposition solutions and methods stemsprimarily from the differences observed in the pH response of theresultant iridium oxide layers.24

The goal of the work reported below is to systematically inves-tigate the effect of solution composition and deposition methodol-ogy on the properties of iridium oxide pH sensing materials. Weinvestigated several iridium solutions containing oxalate or ethylenediamine tetraacetic acid �EDTA� complexing ligands, as well asphosphate buffer alone. We tested four different electrochemicaldeposition protocols involving a single-step constant current, andconstant potential methods, as well as pulsed potential and cyclicpotential growth methods. Within this scope, we controlled thethickness of the electrodeposited oxide films and determined theslope of the pH response in the range of pH 2–12, the rate of pHresponse, and the capacitance of the various iridium oxide materials.We found that the Nernstian slope of the pH response is essentiallyindependent of the deposition method, although other AIROF sensorproperties, such as the rate of response and the observed capaci-tance, do depend on the deposition procedure as well and the thick-ness of the oxide layer. The sensor response rate and capacitancewere measured by our microelectrochemical time-of-flight methodwith galvanostatic ion generation and potentiometric sensing �P-ETOF�. The basic principles of this method are outlined below. Adetailed description of P-ETOF can be found in the literature.33-35

Experimental

Reagents.— Nanopure water was obtained using a three-cartridge Millipore purification system. Potassium chloride, sulfuricacid �reagent grade�, hydrogen peroxide �30%�, perchloric acid�70%�, acetone �reagent grade�, and potassium carbonate were sup-plied by EM Science. Lithium perchlorate �99.99%�, oxalic acid�99%�, and EDTA were supplied by Aldrich. Iridium �IV� chloride�99.95%� and cesium nitrate �99.99%� were supplied by Alfa Aesar.All chemicals were used as received.

Iridium oxide deposition procedures.— Electrochemically in-duced iridium oxide deposition was carried out following the Ya-manaka procedure.18 The same procedure was used when EDTAreplaced oxalate ions as the complexing agent. Additionally, depo-sition solutions were made from both iridium �III� and iridium �IV�


chlorides without any ligands, using potassium hydroxide, potas-sium carbonate, or sodium phosphate to adjust the final pH to �11.The deposition solution was 4 mM IrCl4 with a tenfold excess ofoxalic acid or EDTA. The pH of the solution was adjusted to 10.5with K2CO3. The resulting solution was allowed to stabilize overtwo days at room temperature. Two types of gold electrodes werecoated with iridium oxide films: macroelectrodes, 0.20 cm2 in area,made by Au vapor deposition on a thin ca. 5 nm adhesive layer ofvapor-deposited chromium on glass substrates, and ca. 10 �m wide,2 mm long photolithographically fabricated microelectrodes. Thelatter type was used in electrochemical time-of-flight �P-ETOF� de-vices described below and in our previous reports.34,35 In both cases,the working area of the electrode was defined and electrically iso-lated with a polymer resin. The iridium oxide deposition was carriedout using either a constant potential, constant current, multiplepulsed potential, or cyclic potential method. In constant currentdepositions, a current density of 0.2–0.3 mA/cm2 was applied for60 s to 10 min, depending on the desired thickness of the oxidefilm. In constant potential depositions, the electrode was generallyheld at 0.6–0.7 V vs SCE for a desired time. Pulsed potential depo-sition method typically alternated the electrode potential between0.7 V vs SCE and −0.5 V with pulse lengths of 0.25 s over100–10,000 cycles. Hydrogen evolution was postulated to takeplace during the negative half of the cycle preventing excessivedecrease of pH.36 In potential cycling depositions, the potential wascycled between 0.7 and −0.5 V vs SCE at a rate of 5–10 V/s formicroelectrodes and 1–2.5 V/s for macroelectrodes over 100 to sev-eral thousand cycles. After AIROF deposition, the electrodes werestored in DI water. This led to a response of 79 mV/pH unit asdescribed elsewhere.37

Electrochemical time-of-flight experiments.— A detailed de-scription of this technique can be found in the literature.33-35,38 Inthe present work, we relied on the open-face mode of protonP-ETOF. A schematic design of a device is shown in Fig. 1. Itconsisted of two parallel, lithographically fabricated gold microelec-trodes on a glass slide. The electrodes were 10 �m wide and ca.2 mm long, functioning as generator and sensor. Conventionalcounter and reference electrodes were used but are not shown in Fig.1. The typical interelectrode gap was 20, 50, or 100 �m. Protonswere generated from aqueous solutions via electro-oxidation of wa-ter on an unmodified gold microelectrode using constant current in arange of 1 �A to 1 mA. The sensor microelectrode was modifiedwith an iridium oxide film of desired thickness produced by one ofthe four methods outlined above. The electrode assembly was im-

G SG S

Figure 1. �Color online� Schematic diagram of an open-face P-ETOF de-vice. Gold electrodes are photolithographically fabricated on a glass slide.Protons are produced via galvanostatic electro-oxidation of water at the gen-erator microelectrode, G, and diffuse hemicylindrically toward the sensormicroelectrode, S. The latter is coated with a pH-sensitive film of iridiumoxide. The interelectrode gap is 50 or 100 �m.


F3Journal of The Electrochemical Society, 156 �1� F1-F6 �2009� F3

mersed in a supporting electrolyte solution �typically 0.1 or 1.0 Mcesium nitrate�. Transport of electrogenerated protons between thegenerator and sensor microelectrodes obeyed hemicylindrical diffu-sion as shown schematically in Fig. 1. Drift of the AIROFs, whetherredox or pH induced, was insignificant in comparison to the mea-sured potential changes over the time scale of the experiment.

Instrumentation.— Electrochemical experiments were carriedout using a CH Instruments model 660B electrochemical analyzer.Recording of P-ETOF transients was made possible through dual-channel recording during chronopotentiometric experiments. Thegenerator was directly controlled, and the open-circuit potential wasmonitored at the sensor electrode. This instrument had a 10 kHzsampling rate and better than 0.5 mV resolution.

Results and Discussion

pH response.— As discussed above, several solutions of differ-ent compositions were reported in the literature for depositing iri-dium oxide films. It has been commonly postulated that the presenceor absence of oxalate ions is of fundamental importance, affectingthe properties of the AIROF films. To test this hypothesis, we carriedout electrochemically induced deposition of iridium oxide films ongold substrates following the Yamanaka procedure18 using eitheroxalate or replacing it with EDTA �see Experimental section�. In thethird set of experiments, neither of these complexing agents wasused, and a pH of 10.5–11.0 of the deposition solution was obtainedwith either trisodium phosphate, potassium carbonate, or KOH. Sub-sequently, the calibration curves of the AIROF sensors were re-corded. As can be seen in Fig. 2, the presence or absence of oxalatein the deposition solution does not change the response of the sen-sors if the devices are made in an otherwise similar fashion. Astatistically identical slope value of 79.1 mV/pH unit was obtainedin the presence of ETDA, oxalate, and phosphate ions. The interceptvalues of 722, 759, and 830 mV, respectively, correlated with thethickness of the oxide film as shown below. The only notable dif-ference between these solutions was the fact that the solutions inwhich KOH was used deteriorated quickly after use, similar to thesolutions made by Baur and Spaine.19

On the basis of this, it is likely that the deposition solutions allproduce a similar iridium oxide layer. Another interesting observa-tion along these lines is that, independent of the ligands presentinitially or added in the process of making the solution, all iridiumoxide deposition solutions are a deep blue or purple color in theiractive form. This is consistent with at least a diaquo hydration stateof Ir�IV� species. Aside from the procedures documented in theliterature, EDTA containing solutions resulted in blue-purple solu-

-0.3-0.2-0.100.10.20.30.40.50.60.7

0 2 4 6 8 10 12 14

pH

PhosphateoxalateEDTA

Potential(V)vsSCE

abc

abc

-0.3-0.2-0.100.10.20.30.40.50.60.7

0 2 4 6 8 10 12 14

pH

PhosphateoxalateEDTA

Potential(V)vsSCE

abc

abc

Figure 2. �Color online� A set of calibration curves obtained with iridiumoxide pH sensors deposited according to the Yamanaka procedure17 from: aligand-free solution of Ir�IV� chloride, pH 10.5 adjusted with Na3PO4 �a�, anoxalate containing Ir�IV� chloride solution �b�, and an EDTA containingIr�IV� chloride solution �c�. The average slope of the three linear fits throughthe data points is 79.2 � 0.2 mV/pH unit.


tions, consistent with a diaquoiridate �IV� complex, whereas solu-tions without ligands all contained a strong base, which is known toproduce a stable hexaaquoiridate species.30-32,39

If the different solutions produce very similar if not identicaltype of iridium oxide films, then a natural question to ask is whetherthey do so via the same mechanism. As previously mentioned, theoxalate-containing solutions are supposed to function via the oxida-tion of the oxalate ligands, resulting in an insoluble compound thatthen precipitates onto the electrode. However, for all deposition so-lutions, including those containing no ligands other than hydroxideions, the AIROF deposition took place at similar potentials, suggest-ing that the mechanism is indeed the same. Between the varioussolutions, the only common oxidizable species are iridium and wa-ter. Most deposition solutions contain iridium �IV�, and althoughoxidation to iridium �V� is possible, it is not likely as iridium �V�typically exists in concentrated �2 M or higher� acid solutions atpotentials in excess of 1.6 V �vs a normal hydrogen electrode�.32 Itis therefore reasonable to postulate that water is the species under-going oxidation. However, only water coordinated to Ir�IV� fits theexperimental evidence: in constant potential depositions, stirred so-lutions showed a limiting current twice as large as that obtained inunstirred solutions with both micro- and macroelectrodes.

If coordinated water is being oxidized, then the mechanism forAIROF formation likely involves the oxidative release of a protonfollowed by the formation of oxy or hydroxy bridges between iri-dium complexes, resulting in a loose polymeric network that pre-cipitates at the electrode. This type of bridged iridium complex hasbeen described in the literature.40

Next we examined the effect of the specific electrochemical pro-tocols used in AIROF electrodeposition. Single-step constant poten-tial and constant current methods, as well as pulsed potential andcyclic potential sweep methods described in the Experimental sec-tion were used. As can be seen in Fig. 3, the pH response of theresulting AIROF �79 mV/pH unit� is independent of the depositionmethod and the thickness of the deposited film. The latter, however,determines the intercept of the calibration curves.

We also observed that increasing film thickness progressivelyresults in less uniform and more fragile films, particularly whenproduced by the galvanostatic method. In comparison, the three con-trolled potential deposition methods create layers that are more uni-form and exhibit better adhesion. Of the three, the constant potentialprocedure is most suitable for preparation of the thickest layers. Aninteresting feature of this procedure is that it allows one to roughlydetermine the time at which a complete iridium oxide layer isformed. The results of the experiments examining the effect of theoxide film thickness are shown in Fig. 4. Figure 4a shows a typicalcurrent-vs-time trace recorded during a constant potential depositionrun. The initially high current �which we postulated is due to thereduction of water coordinated to Ir�IV�� decreases rapidly duringthe first several seconds of the experiment as a result of the decreas-ing surface area of gold electrode that is progressively coated withiridium oxide. A current minimum is observed after ca. 40 s, whichcorresponds, we postulate, to elimination of uncoated gold substrateand formation of a thin, continuous iridium oxide film. This postu-late was supported by microscopic examination of the electrode sur-faces and by profilometric traces recorded at various stages ofdeposition.37 Subsequently, as the thickness of the microscopicallyporous film increases, the current begins to increase and this trendcontinues until the end of the deposition process �ca. 10 min in Fig.4a�. We define the time necessary to deposit a full, continuous filmof iridium oxide, or the time of the current minimum as T. Next weexamined the pH dependence of the AIROF sensors produced by theconstant potential deposition carried out between 0.4 and 1.5 T.These calibration curves are shown in Fig. 4b. It is interesting tonote that as the deposition time increases between 0.4 and 1 T, theslope of the calibration curves increases from 49 to 76 mV/pH unit.The three AIROF sensors of the highest thickness �obtained in thedepositions of 1, 1.2, and 1.5 T� exhibited the same Nernstian slopeof 76 mV/pH unit. This set of experiments further substantiates our



hypothesis that the pH response �the slope of the calibration curves�depends on the quantity of the deposited iridium oxide in the rangewhere the latter does not form a continuous film. As the depositiontime increases beyond 1 T leading to the progressively greater filmthickness, the intercept also increases as noted above �see Fig. 3b�.

Capacitance and the rate of pH response.— The maximum rateof response of sensor electrodes was examined using P-ETOF �seeExperimental section�. This technique allows us to control the rateof diffusion of protons toward an AIROF microsensor by controllingthe magnitude of the current pulse applied to the generator micro-electrode. In most cases, the rate of pH change recorded at theiridium oxide microsensor is slower than theoretically expectedwhen only the time-dependent, diffusion-controlled concentration ofH+ ions at the microsensor surface is taken into consideration. Asdiscussed below, this is due primarily to a finite capacitance ofAIROF microsensors.34 However, we do observe an increase in therate of the potential change of microsensors with an increasing gen-eration current. The slope of the E vs t transient that no longerincreases upon an additional increase of the generator current is theninterpreted in terms of the maximum rate of response.

A typical series of P-ETOF E vs t transients recorded to deter-mine the maximum rate of response of two AIROF sensors is shownin Fig. 5. Clearly, as the current generating H+ ions is increased therecorded E vs t transients become steeper. Their slope no longerincreases when the generator current is ca. 1 � 10−4 A or greater. Itis also apparent that the maximum rate of response depends on thethickness of the iridium oxide layer. Using this approach, we mea-sured the maximum rate of AIROF response as a function of the filmthickness for the microsensors produced by each of the four electro-chemical methods discussed above �Fig. 3�. These results are sum-

y = -0.0788x+0.843R2 =0.9997

y = -0.0791x + 0. 744R2 = 0.9996

y = -0.0769x+0.634R2 =0.9987

y = -0.0794x+0.694R2 =0.9989

-0.4

-0.2

0

0.2

0.4

0.6

0.8

0 2 4 6 8 10 12 14

pH

Potential(V)vsSCE

constant currentconstant potentialpulsed potentialCV growth

y = 0.0008x + 0.6072R2 = 0.9845

0.6

0.65

0.7

0.75

0.8

0.85

0.9

0 50 100 150 200 250 300 350

thickness (nm)

Intercept(V)vsSCE

A

B

Potential(V)vsSCE

Intercept(V)vsSCE

thickness (nm)

y = -0.0788x+0.843R2 =0.9997

y = -0.0791x + 0. 744R2 = 0.9996

y = -0.0769x+0.634R2 =0.9987

y = -0.0794x+0.694R2 =0.9989

-0.4

-0.2

0

0.2

0.4

0.6

0.8

0 2 4 6 8 10 12 14

pH

Potential(V)vsSCE


y = 0.0008x + 0.6072R2 = 0.9845

0.6

0.65

0.7

0.75

0.8

0.85

0.9

0 50 100 150 200 250 300 350

thickness (nm)

Intercept(V)vsSCE

A

B

Potential(V)vsSCE

Intercept(V)vsSCE

thickness (nm)Thickness (nm)

4321

y = -0.0788x+0.843R2 =0.9997

y = -0.0791x + 0. 744R2 = 0.9996

y = -0.0769x+0.634R2 =0.9987

y = -0.0794x+0.694R2 =0.9989

-0.4

-0.2

0

0.2

0.4

0.6

0.8

0 2 4 6 8 10 12 14

pH

Potential(V)vsSCE


y = 0.0008x + 0.6072R2 = 0.9845

0.6

0.65

0.7

0.75

0.8

0.85

0.9

0 50 100 150 200 250 300 350

thickness (nm)

Intercept(V)vsSCE

A

B

Potential(V)vsSCE

Intercept(V)vsSCE

thickness (nm)

y = -0.0788x+0.843R2 =0.9997

y = -0.0791x + 0. 744R2 = 0.9996

y = -0.0769x+0.634R2 =0.9987

y = -0.0794x+0.694R2 =0.9989

-0.4

-0.2

0

0.2

0.4

0.6

0.8

0 2 4 6 8 10 12 14

pH

Potential(V)vsSCE


y = 0.0008x + 0.6072R2 = 0.9845

0.6

0.65

0.7

0.75

0.8

0.85

0.9

0 50 100 150 200 250 300 350

thickness (nm)

Intercept(V)vsSCE

A

B

Potential(V)vsSCE

Intercept(V)vsSCE


y = -0.0788x+0.843R2 =0.9997

y = -0.0791x + 0. 744R2 = 0.9996

y = -0.0769x+0.634R2 =0.9987

y = -0.0794x+0.694R2 =0.9989

-0.4

-0.2

0

0.2

0.4

0.6

0.8

0 2 4 6 8 10 12 14

pH

Potential(V)vsSCE


y = 0.0008x + 0.6072R2 = 0.9845

0.6

0.65

0.7

0.75

0.8

0.85

0.9

0 50 100 150 200 250 300 350

thickness (nm)

Intercept(V)vsSCE

A

B

Potential(V)vsSCE

Intercept(V)vsSCE

thickness (nm)

y = -0.0788x+0.843R2 =0.9997

y = -0.0791x + 0. 744R2 = 0.9996

y = -0.0769x+0.634R2 =0.9987

y = -0.0794x+0.694R2 =0.9989

-0.4

-0.2

0

0.2

0.4

0.6

0.8

0 2 4 6 8 10 12 14

pH

Potential(V)vsSCE


y = 0.0008x + 0.6072R2 = 0.9845

0.6

0.65

0.7

0.75

0.8

0.85

0.9

0 50 100 150 200 250 300 350

thickness (nm)

Intercept(V)vsSCE

A

B

Potential(V)vsSCE

Intercept(V)vsSCE


4321

Figure 3. �Color online� �A� set of calibration curves obtained with iridiumoxide pH sensors deposited from oxalate containing Ir�IV� chloride solutionsby a constant current method �1�, a single potential step method �2�, a cyclicpotential sweep method �3�, and an alternating pulsed potential method �4�.In all cases, the pH response is ca. 79 mV/pH unit. �B� shows the depen-dence of the intercept of the four calibration curve of �A� on the thickness ofthe iridium oxide films.


marized in Fig. 6. It is apparent that regardless of the electrochemi-cal protocol used to produce AIROF sensors, thinner films yield fastresponse. Also, in all cases, a plateau of low response rate is ob-tained in the range of high film thicknesses. A careful analysis of thedata in Fig. 6 further shows that while the plateau region is assumedat thicknesses in excess of 200 nm for the films produced by theconstant current and constant potential methods, such a plateau isapproached at significantly lower film thicknesses in the case offilms prepared by the pulsed potential �ca. 60 nm� and the cyclicpotential �150 nm� methods. In view of the fact that pH response ofAIROFs is associated with proton diffusion into the microporousAIROF films, it is reasonable to infer that the pulsed potentialmethod and, to some extent, the cyclic potential method yield filmswith more constricted and narrow pore structures impeding protondiffusion compared to films obtained by the other two methods. Thisgeneral conclusion can also be illustrated by the maximum rate ofresponse measured at a specific thickness. For example, the maxi-mum rate of response exhibited by the 50 nm sensors obtained bythe four methods can be estimated to be 16, 23, 7, and 14 V/s,respectively. Again, those sensors produced by the pulsed potentialand cyclic potential methods exhibit slower response relative to thesensors produced by the other two methods.

A second electrochemical property that was found to vary withdeposition method is the capacitance of the AIROF potentiometricsensors. Capacitance of potentiometric sensors has been addressedin our previous work.34,35 Briefly, when potentiometric sensors areemployed in relatively small volume samples, the finite capacitanceof a sensor may lead to an erroneous reading. This is because achange of a sensor’s potential itself decreases the concentration of

Figure 4. �Color online� �A� Current vs time trace obtained during a con-stant potential �E = 0.7 V vs SCE� deposition of iridium oxide film on a Auelectrode �A = 0.2 cm2�. The current minimum at 40 s is referred to as T. �B�compares the potential vs pH calibration curves obtained with the iridiumoxide films obtained in �A� using the deposition times of �a� 0.4 �b� 0.5, �c�0.6, �d� 0.8, �e� 1.0, �f� 1.2, and �g� 1.5 T. The pH response increases from49 mV/pH unit �a� to a constant value of 76 mV/pH unit �e–g�.


3

F5Journal of The Electrochemical Society, 156 �1� F1-F6 �2009� F5


the primary ion near its surface. To minimize this problem it isadvisable to miniaturize sensors in order to reduce their total capaci-tance. We previously showed that, due to their internal porosity, theapparent capacitance of AIROF sensors scales not only with theirarea but also with their thickness.34,35 In cases mentioned above, it isthus advantageous to miniaturize sensors both in terms of their sur-face area and their thickness. Below we explore the effect of thedeposition method on the intrinsic 3D capacitance �in F/cm3� ofiridium oxide sensors. Our approach relies on H+ P-ETOF recordedunder conditions of moderate generator current where the E vs ttransients are not limited by the rate of sensor response. Because, asmentioned earlier, the recorded response is slower than that deter-mined by the proton diffusion between the generator and sensorelectrode, the delay is assigned to the capacitance of a sensor. Spe-cifically, its value is obtained by fitting experimental transients usingan explicit finite difference algorithm that takes into considerationsensor’s capacitance.37

In Fig. 7, we show three H+ P-ETOF transients obtained underidentical conditions with three sensors of different thickness. Notsurprisingly, the rise time of the transients correlates inversely withtheir thickness. The thick black trace exhibiting a minimal delay is aresult of a simulation in which sensor’s capacitance was assumed tobe zero. Clearly, in comparison to that theoretical transient, or rela-tive to the proton diffusion rate, all experimental transients show asubstantially delayed response. The open-circle transients are theo-retical fits to the experimental data with the sensors’ two-dimensional capacitance shown in Fig. 7. We note that relative toour earlier report, our current fitting algorithm allows us to fit theentire E vs t transient. The resulting capacitance values listed in Fig.7 are also greater than those previously reported for similar typesensors.34,35 This is due to an increased understanding of the sec-ondary effects of large capacitances, such as an increase of ion fluxresulting from a capacitive ion depletion near the sensor electrode.37

The dependence of sensors’ capacitance on their thickness anddeposition method is shown in Fig. 8. Approximately linear depen-dence of capacitance on film thickness was observed in all cases. Tointerpret data in Fig. 8 more insightfully, we calculated the slope ofthe linear plots, or the intrinsic 3D capacitance of the sensors ob-tained by the four methods. These were 3.0 � 103, 8.7 � 103, 9.2

0

0.05

0.1

0.15

0.2

0 0.2 0.4 0.6 0.8 1

time (seconds)

E(t)-Eint(V)

60 nm 11mF

130 nm 33 mF

330 nm 95 mF

E(t)–E int(V)

0

0.05

0.1

0.15

0.2

0 0.2 0.4 0.6 0.8 1

time (seconds)

E(t)-Eint(V)

60 nm 11mF

130 nm 33 mF

330 nm 95 mF

E(t)–E int(V)

Time (s)

a b c d

0

0.05

0.1

0.15

0.2

0 0.2 0.4 0.6 0.8 1

time (seconds)

E(t)-Eint(V)

60 nm 11mF

130 nm 33 mF

330 nm 95 mF

E(t)–E int(V)

0

0.05

0.1

0.15

0.2

0 0.2 0.4 0.6 0.8 1

time (seconds)

E(t)-Eint(V)

60 nm 11mF

130 nm 33 mF

330 nm 95 mF

E(t)–E int(V)

Time (s)

a b c d

Figure 7. �Color online� Proton P-ETOF potential vs time transients re-corded with 50 �m devices operated at 1.0 � 10−5 A generator current fea-turing the iridium oxide sensors �b� 60, �c� 130, and �d� 330 nm in thick-nesses. Line a is the simulated response with the same generator current,DH+ = 8.5 � 10

−5 cm2/s, and sensor capacitance, C = 0. The small dottedtransients are the simulated responses obtained with sensor capacitance of �b�11, �c� 33, and �d� 95 mF/cm2. The experiments were run in a pH 5.7, 0.1 MCsNO3 solution.

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

E(t)-Eint(V)

5e-61e-55e-51e-4

E(t)–E int(V)

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

time (seconds)

E(t)-Eint(V)

5e-6 1e-5

5e-5 1e-4

E(t)–E int(V)

A

B

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

E(t)-Eint(V)

5e-61e-55e-51e-4

E(t)–E int(V)

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

time (seconds)

E(t)-Eint(V)

5e-6 1e-5

5e-5 1e-4

E(t)–E int(V)

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

E(t)-Eint(V)

5e-61e-55e-51e-4

E(t)–E int(V)

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

time (seconds)

E(t)-Eint(V)

5e-6 1e-5

5e-5 1e-4

E(t)–E int(V)

A

B

Time (s)

ab c

de

ab

c

d

e

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

E(t)-Eint(V)

5e-61e-55e-51e-4

E(t)–E int(V)

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

time (seconds)

E(t)-Eint(V)

5e-6 1e-5

5e-5 1e-4

E(t)–E int(V)

A

B

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

E(t)-Eint(V)

5e-61e-55e-51e-4

E(t)–E int(V)

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

time (seconds)

E(t)-Eint(V)

5e-6 1e-5

5e-5 1e-4

E(t)–E int(V)

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

E(t)-Eint(V)

5e-61e-55e-51e-4

E(t)–E int(V)

0

0.05

0.1

0.15

0.2

0.25

0.3

0 0.2 0.4 0.6 0.8 1

time (seconds)

E(t)-Eint(V)

5e-6 1e-5

5e-5 1e-4

E(t)–E int(V)

A

B

Time (s)

ab c

de

ab

c

d

e

Figure 5. �Color online� Proton P-ETOF potential vs time transients re-corded with the iridium oxide sensors �A� 80 and �B� 320 nm in thicknessprepared by a constant current method. In both sets of data, transient a is thesimulated response with the generator current of 1 � 10−4 A, DH+ = 8.5� 10−5 cm2/s, and sensor capacitance, C = 0. Transients b–e were recordedwith the generator currents of 1 � 10−4, 5 � 10−5, 1 � 10−5, and 5� 10−6 A, respectively. The experiments were run with a 50 �m P-ETOFdevice in a pH 5.7, 0.1 M CsNO solution.

0

5

10

15

0 100 200 300

thickness (nm)

rateofresponse(V/s)

0

5

10

15

20

0 100 200 300 400

thickness (nm)

rateofresponse(V/s)

0

5

10

15

20

0 50 100 150 200 250thickness (nm)

rateofresponse(V/s)

RateofResponse(V/s)

0

10

20

30

0 50 100thickness (nm)

rateofresponse(V/s)

RateofResponse(V/s)

A

DC

B0

5

10

15

0 100 200 300

thickness (nm)

rateofresponse(V/s)

0

5

10

15

20

0 100 200 300 400

thickness (nm)

rateofresponse(V/s)

0

5

10

15

20

0 50 100 150 200 250thickness (nm)

rateofresponse(V/s)

RateofResponse(V/s)

0

10

20

30


rateofresponse(V/s)

RateofResponse(V/s)

A

DC

B

Thickness (nm) Thickness (nm)


0

5

10

15

0 100 200 300

thickness (nm)

rateofresponse(V/s)

0

5

10

15

20

0 100 200 300 400

thickness (nm)

rateofresponse(V/s)

0

5

10

15

20

0 50 100 150 200 250thickness (nm)

rateofresponse(V/s)

RateofResponse(V/s)

0

10

20

30


rateofresponse(V/s)

RateofResponse(V/s)

A

DC

B0

5

10

15

0 100 200 300

thickness (nm)

rateofresponse(V/s)

0

5

10

15

20

0 100 200 300 400

thickness (nm)

rateofresponse(V/s)

0

5

10

15

20

0 50 100 150 200 250thickness (nm)

rateofresponse(V/s)

RateofResponse(V/s)

0

10

20

30


rateofresponse(V/s)

RateofResponse(V/s)

A

DC

B



Figure 6. �Color online� Dependence of the maximum rate of response onthe thickness of the iridium oxide sensors deposited by the �A� constantcurrent, �B� constant potential, �C� pulsed potential, and �D� cyclic potentialgrowth methods.



� 102, and 9.5 � 102 F/cm3, respectively. Thus, the sensors ob-tained by the pulsed potential and the cyclic potential methods ex-hibit significantly lower intrinsic capacitances compared to thoseproduced by the other two methods. This is due to the smaller po-rosity and an apparently smaller internal surface area of the formertwo sensors. This correlates well with their relatively slower rate ofresponse discussed above �Fig. 6�. Finally, it is worth noting that theplots of capacitance vs film thickness of sensors produced by theconstant current and constant potential methods exhibit negative in-tercepts. The plots for the pulsed potential and cyclic potential meth-ods do not. This is due to a less uniform deposition or a greaterroughness of the sensors featured in Fig. 8a and b. Conversely, thesensors produced by the other two methods are more uniform inthickness.

Conclusions

The pH response of iridium oxide sensors was shown to be in-dependent of the type of solution used and the deposition method.Unbuffered deposition solutions made with potassium hydroxidewere unstable beyond a few hours unlike those in which pH wasadjusted with potassium carbonate or sodium phosphate. The maxi-mum rate of pH response for sensors of the same thickness wasfound to be at least somewhat smaller for the sensors produced bythe pulsed potential and cyclic potential methods relative to thoseproduced by the other two methods examined in this project. How-

y = 0.8719x - 62.483R2 = 0.9708

0

50

100

150

200

0 100 200 300

thickness (nm)capacitance(mF)

y = 0.3019x - 7.6251R2 = 0.9803

0

20

40

60

80

100

120

0 100 200 300 400

thickness (nm)

capacitance(mF)

y = 0.0948x+ 0.3399R2 = 0.8907

0

5

10

15

20

25


capacitance(mF)

A B

DC

y = 0.0917x + 6.4841R2 = 0.8977

0

5

10

15

20


capacitance(mF)

y = 0.8719x - 62.483R2 = 0.9708

0

50

100

150

200

0 100 200 300


y = 0.3019x - 7.6251R2 = 0.9803

0

20

40

60

80

100

120

0 100 200 300 400

thickness (nm)

capacitance(mF)

y = 0.0948x+ 0.3399R2 = 0.8907

0

5

10

15

20

25


capacitance(mF)

A B

DC

y = 0.0917x + 6.4841R2 = 0.8977

0

5

10

15

20


capacitance(mF)

Capacitance(mF)

Capacitance(mF)


y = 0.8719x - 62.483R2 = 0.9708

0

50

100

150

200

0 100 200 300


y = 0.3019x - 7.6251R2 = 0.9803

0

20

40

60

80

100

120

0 100 200 300 400

thickness (nm)

capacitance(mF)

y = 0.0948x+ 0.3399R2 = 0.8907

0

5

10

15

20

25


capacitance(mF)

A B

DC

y = 0.0917x + 6.4841R2 = 0.8977

0

5

10

15

20


capacitance(mF)

y = 0.8719x - 62.483R2 = 0.9708

0

50

100

150

200

0 100 200 300


y = 0.3019x - 7.6251R2 = 0.9803

0

20

40

60

80

100

120

0 100 200 300 400

thickness (nm)

capacitance(mF)

y = 0.0948x+ 0.3399R2 = 0.8907

0

5

10

15

20

25


capacitance(mF)

A B

DC

y = 0.0917x + 6.4841R2 = 0.8977

0

5

10

15

20


capacitance(mF)

Capacitance(mF)

Capacitance(mF)


Figure 8. Dependence of the capacitance on the thickness of iridium oxidesensors deposited by the �A� constant current, �B� constant potential, �C�pulsed potential, and �D� cyclic potential growth methods. The slope of thelinear fits or the specific capacitance of the sensors are �A� 3000 � 200, �B�8700 � 600, �C� 900 � 150, and �D� 1000 � 170 F/cm3.

ever, because of the uniformity of the sensing layers produced by the


pulsed potential and cyclic potential methods, the thickness neces-sary for stable sensor function is smaller. We postulated that this canbe explained by a relatively smaller porosity of those films. Consis-tently with this postulate, the sensors produced by the pulsed poten-tial and cyclic potential methods were also found to exhibit smallerintrinsic capacitance relative to those produced by the other twomethods.

AcknowledgmentsWe acknowledge and thank the National Science Foundation for

supporting this work under grants CHE-0416349 and CHE-0719334.

References1. S. Glab, A. Hulanicki, G. Edwall, and F. Ingman, Crit. Rev. Anal. Chem., 21, 29

�1989�.2. A. Fog and R. P. Buck, Sens. Actuators, 5, 137 �1984�.3. M. Wang, S. Yao, and M. Madou, Sens. Actuators B, 81, 313 �2002�.4. T. Katsube, I. Launks, and J. N. Zemel, Sens. Actuators, 2, 399 �1982�.5. J. V. Dobson, P. R. Snodin, and H. R. Thirsk, Electrochim. Acta, 21, 527 �1976�.6. D. Midgley, Talanta, 37, 767 �1990�.7. K. Izutsu and H. Yamamoto, Anal. Sci., 12, 905 �1996�.8. M. F. Smiechowski and V. F. Lvovich, Sens. Actuators B, 96, 261 �2003�.9. M. L. Hitchman and S. Ramanathan, Analyst (Cambridge, U.K.), 116, 1131

�1991�.10. I. Lauks, M. F. Yuen, and T. Dietz, Sens. Actuators, 4, 375 �1983�.11. S. A. M. Marzouk, S. Ufer, R. P. Buck, T. A. Johnson, L. A. Dunlap, and W. E.

Cascio, Anal. Chem., 70, 5054 �1998�.12. D. O’Hare, K. H. Parker, and C. P. Winlove, Med. Eng. Phys., 28, 982 �2006�.13. K. Kinoshita and M. J. Madou, J. Electrochem. Soc., 131, 1089 �1984�.14. P. J. Kinlen, J. E. Heider, and D. E. Hubbard, Sens. Actuators B, 22, 13 �1994�.15. M. J. Tarlov, S. Semancik, and K. G. Kreider, Sens. Actuators B, 1, 293 �1990�.16. J. Hendrikse, W. Olthuis, and P. Bergveld, Sens. Actuators B, 53, 97 �1998�.17. S. Yao, M. Wang, and M. Madou, J. Electrochem. Soc., 148, H29 �2001�.18. K. Yamanaka, Jpn. J. Appl. Phys., Part 1, 28, 632 �1989�.19. J. E. Baur and T. W. Spaine, J. Electroanal. Chem., 443, 208 �1998�.20. Y. Jung, J. Lee, and Y. Tak, Electrochem. Solid-State Lett., 7, H5 �2004�.21. R. K. Jaworski and J. A. Cox, J. Electroanal. Chem., 325, 111 �1992�.22. K. Passztor, A. Sekiguchi, N. Shimo, N. Kitamura, and H. Masuhara, Sens. Actua-

tors B, 12, 225 �1993�.23. S. C. Mailley, M. Hyland, P. Mailley, J. M. McLaughlin, and E. T. McAdams,

Mater. Sci. Eng., C, 21, 167 �2002�.24. M. A. Petit and V. Plinchon, J. Electroanal. Chem., 444, 247 �1998�.25. T. Yoshino, J. Fukawa, and N. Baba, Surf. Sci., Jpn. Ed., 7, 480 �1986�.26. T. Yoshino, N. Baba, and K. Arai, Jpn. J. Appl. Phys., Part 1, 26, 1547 �1987�.27. W. P. Griffith, The Chemistry of the Rarer Platinum Metals (Os, Ru, Ir, and Rh), p.

227, Interscience Publishers, London �1967�.28. J. C. Change and C. S. Garner, Inorg. Chem., 4, 209 �1965�.29. J. Juodkazytė, B. Šebeka, I. Valsiunas, and K. Juodkazis, Electroanalysis, 17, 947

�2005�.30. G. van Loon and J. A. Page, Can. J. Chem., 44, 515 �1966�.31. P. Beutler and H. Gamsjäger, J. Chem. Soc., Chem. Commun., 1976, 554.32. S. E. Castillo-Blum, D. T. Richens, and A. G. Sykes, Inorg. Chem., 28, 954 �1989�.33. K. Slowinska, S. W. Feldberg, and M. Majda, J. Electroanal. Chem., 554-555, 61

�2003�.34. H. A. Elsen, K. Slowinska, E. Hull, and M. Majda, Anal. Chem., 78, 6356 �2006�.35. K. Slowinska and M. Majda, J. Solid State Electrochem., 8, 763 �2004�.36. C. Terashima, T. N. Rao, B. V. Sarada, N. Spataru, and A. Fijishima, J. Electroanal.

Chem., 544, 65 �2003�.37. H. A. Elsen, Ph.D. Thesis, University of California, Berkeley, CA �2007�.38. C. F. Monson and M. Majda, Anal. Chem., 24, 9315 �2007�.39. S. E. Castillo-Blum and A. G. Sykes, Polyhedron, 6, 101 �1987�.40. F. Galsbol, F. Larsen, B. Rasmussen, and J. Springborg, Inorg. Chem., 25, 290

�1986�.


Documents

Effects of Electrodeposition Conditions and Protocol on ...€¦ · 28.07.2008 · Effects of Electrodeposition Conditions and Protocol on the Properties of Iridium Oxide pH Sensor