Water Research 36 (2002) 85–94
`
`Electrochemical generation of hydrogen peroxide from
`dissolved oxygen in acidic solutions
`
`Zhimin Qiang, Jih-Hsing Chang, Chin-Pao Huang*
`
`Department of Civil and Environmental Engineering, University of Delaware, 137, DuPont Hall, Newark, DE 19716-3120, USA
`
`Received 1 July 2000; received in revised form 1 April 2001; accepted 1 April 2001
`
`Abstract
`
`Hydrogen peroxide (H2O2) was electro-generated in a parallel-plate electrolyzer by reduction of dissolved oxygen
`(DO) in acidic solutions containing dilute supporting electrolyte. Operational parameters such as cathodic potential,
`oxygen purity and mass flow rate, cathode surface area, pH,
`temperature, and inert supporting electrolyte
`concentration were systematically investigated as to improve the Faradic current efficiency of H2O2 generation.
`Results indicate that significant self-decomposition of H2O2 only occurs at high pH (>9) and elevated temperatures
`(>231C). Results also indicate that the optimal conditions for H2O2 generation are cathodic potential of –0.5 V vs.
`saturated calomel electrode (SCE), oxygen mass flow rate of 8.2  102 mol/min, and pH 2. Under the optimal
`conditions, the average current density and average current efficiency are 6.4 A/m2 and 81%, respectively. However,
`when air is applied at the optimal flow rate of oxygen, the average current density markedly decreases to 2.1 A/m2, while
`the average current efficiency slightly increases to 90%. The limiting current density is 6.4 A/m2, which is independent of
`cathode geometry and surface area. H2O2 generation is favored at low temperatures. In the concentration range studied
`(0.01–0.25 M), the inert supporting electrolyte (NaClO4) affects the total potential drop of the electrolyzer, but does not
`affect the net generation rate of H2O2. r 2001 Elsevier Science Ltd. All rights reserved.
`
`Keywords: Hydrogen peroxide; Electrochemical generation; Oxygen; Air; Acidic solutions
`
`1. Introduction
`
`is an environmentally
`Hydrogen peroxide (H2O2)
`leaves no hazardous
`friendly chemical because it
`residuals but oxygen and water after reaction. It has
`been widely applied to the synthesis of organic
`compounds, bleaching of paper pulp,
`treatment of
`wastewater, and destruction of hazardous organic
`wastes. In the environmental field, H2O2 is used as a
`supplement of oxygen source to enhance the bioreme-
`diation of contaminated aquifers [1,2]. Moreover, H2O2
`coupled with ozone or UV radiation can effectively
`decompose aqueous organic contaminants [3–6]. The
`most common environmental application of H2O2 is the
`
`*Corresponding author. Tel.: +1-302-831-2442; fax: +1-
`302-831-3640.
`E-mail address: huang@ce.udel.edu (Chin-Pao Huang).
`
`Fenton’s reagent, an aqueous mixture of H2O2 and
`Fe2+. Under an acidic condition, the reaction between
`H2O2 and Fe2+ generates hydroxyl radicals that are
`strong enough to non-selectively oxidize most organic as
`well as some inorganic compounds. The Fenton’s
`reagent has recently been applied to in-situ remediation
`of contaminated soils and groundwaters [7–9].
`H2O2 is usually produced by electrochemical methods,
`such as electrolysis of inorganic chemicals (H2S2O8,
`KHSO4 and NH4HSO4) and autoxidation of organic
`compounds (alkylhydroanthraquinones and isopropyl
`alcohol) [10]. The electrolysis of inorganics requires
`excessive energy and the autoxidation of organics
`requires non-aqueous solvents for catalyst cycle [11].
`H2O2 may also be directly generated from water,
`hydrogen, and oxygen using thermal, photochemical
`and electrical discharge processes. However,
`these
`processes require specific operational conditions such
`
`0043-1354/01/$ - see front matter r 2001 Elsevier Science Ltd. All rights reserved.
`PII: S 0 0 4 3 - 1 3 5 4 ( 0 1 ) 0 0 2 3 5 - 4
`
`Tennant Company
`Exhibit 1120
`
`

`

`86
`
`Z. Qiang et al. / Water Research 36 (2002) 85–94
`
`as high temperature, combustion, UV radiation plus
`mercury vapor, or high voltage. In recent years, small
`scale, on-site H2O2 production processes have gained
`increasing attention because of the cost and the hazards
`associated with the transport and handling of commer-
`cial concentrated H2O2 [11]. If H2O2 can be generated
`on-site in an economic and safe way, its field application
`will be largely simplified. For example,
`it would be
`attractive to use the on-site generated H2O2, coupled
`with ozone, UV, or Fe2+, for the detoxification of
`effluents from electro-kinetics, pump and treat, soil
`washing, and soil flushing processes.
`Most electro-generation of H2O2 experiments are
`conducted in alkaline solutions with a high electrolyte
`concentration for the purpose of bleaching paper pulp
` is
`[12–19]. In concentrated alkaline solutions, HO2
`formed (pKa; H2O2 ¼ 11:62 at 251C) which will be
`immediately repelled by the cathode upon its generation.
` to OH is minimized, a
`Because the reduction of HO2
`high current efficiency (about 80–95%) can usually be
`achieved. However,
`if H2O2 is generated in alkaline
`solutions, a substantial amount of acid will be consumed
`for pH adjustment if an acidic condition is required. The
`Fenton’s reagent, which is most commonly applied in
`organic synthesis and effluent treatment, has an optimal
`pH range of 2.5–3.5. Moreover, a high electrolyte
`concentration not only increases the treatment cost,
`but also introduces additional pollutants. Therefore, it is
`desirable to generate H2O2 in acidic solutions only
`containing dilute supporting electrolyte.
`It has been reported that H2O2 can be electrochemi-
`cally generated by reduction of dissolved oxygen (DO)
`in acidic solutions. The H2O2 so generated can be
`coupled with Fe2+ to produce the Fenton’s reagent for
`either degradation or synthesis of organic compounds
`[20–28]. In order to differentiate this process from
`conventional Fenton process that uses commercial
`H2O2, the term ‘‘electro-Fenton process’’
`is applied.
`The major advantages of the electro-Fenton process
`include: (1) H2O2 can be continuously generated on-site
`whenever needed, which eliminates acquisition, ship-
`ment and storage; (2) A dilute H2O2 solution enhances
`safety during material handling; (3) The production
`process can be simply conducted at ambient pressure
`and temperature; (4) Fe2+ can be electrochemically
`regenerated at
`the cathode, which minimizes
`the
`quantity of iron sludge; and (5) Oxygen or air sparging
`enhances the mixing of reaction solution. The disadvan-
`tage is that H2O2 will accumulate at the cathode-
`solution interface and may be partially decomposed.
`Protons at a high concentration may also compete for
`electrons,
`leading to hydrogen gas evolution. Both
`effects will
`reduce the current efficiency of H2O2
`production. Therefore,
`in acidic solutions, cathodic
`potential and solution pH are two essential factors
`controlling the current efficiency.
`
`Though the electro-generation of H2O2 in acidic
`solutions has been studied by a few researchers, there
`exist conflicting results. Sudoh et al. [21] investigated the
`decomposition of aqueous phenol by electro-generated
`Fenton’s reagent. They found that the highest current
`efficiency (85%) was obtained at a cathodic potential of
`0.6 V vs. a saturated Ag/AgCl electrode (SSE) and pH
`3. Tzedakis et al. [23] reported that the electrolysis of an
`oxygen-saturated H2SO4 solution (0.6 M) using a stirred
`mercury pool electrode yielded a current efficiency of
`55% at a cathodic potential of 0.30 V vs. SCE. Chu
`[27] used the electro-Fenton process to remove aqueous
`chlorophenols at a cathodic potential of 0.6 V vs. SCE,
`and reported that the current efficiency increased with
`decreasing pH. Hsiao and Nobe [25] investigated the
`oxidative hydroxylation of phenol and chlorobenzene
`using electro-generated Fenton’s reagent. They reported
`that the optimal cathodic potential was 0.55 V vs.
`SCE, and the generation of H2O2 was favored at low pH
`values.
`The primary objective of the present study is to
`improve the Faradic current efficiency of H2O2 genera-
`tion in acidic solutions. It is also expected that the
`results will clarify existing discrepancies among various
`studies of
`its kind. Influential parameters such as
`cathodic potential, oxygen purity and mass flow rates,
`cathode surface area, solution pH, temperature, and
`inert supporting electrolyte concentration were system-
`atically examined. Considering that a high electrolyte
`concentration is usually not feasible for effluent treat-
`ment by the electro-generated H2O2, the inert support-
`ing electrolyte, i.e., NaClO4, was used only at a low
`concentration.
`
`2. Experimental
`
`Fig. 1 shows the schematic diagram of the reaction
`system. An acrylic parallel-plate electrolyzer was em-
`ployed for the electro-generation of H2O2. The cathodic
`and anodic compartments had a volume of 4.50 and
`3.15 l, respectively. During the experiments, 4.0 l of
`catholyte and 3.0 l of anolyte were used and both were
`completely mixed by a magnetic stir plate. A cation
`exchange membrane (Neosepta CMX, Electrosynthesis
`Company, Lancaster, NY) was used to separate the two
`compartments. This membrane prohibits the penetra-
`tion of anions and H2O2 molecules, but allows cations to
`freely penetrate through it. As a result, H2O2 generated
`at the cathode will be confined in the catholyte, avoiding
`its decomposition at the anode. Moreover, protons
`generated at the anode will be electrically driven to the
`catholyte, partially supplementing the protons con-
`sumed for H2O2 synthesis. Sodium perchlorate (Na-
`ClO4) was used as an inert supporting electrolyte (or
`background ionic strength). Both cathode and anode
`
`

`

`Z. Qiang et al. / Water Research 36 (2002) 85–94
`
`87
`
`Fig. 1. Schematic diagram of reaction system.
`
`were made of corrosion-resistant graphite (Grade 2020,
`Carbon of America, Bay City, MI). Sudoh et al. [16]
`reported that graphite was the best cathode material for
`the electro-generation of H2O2 in alkaline solutions,
`while metal cathodes such as copper, stainless steel, lead
`and nickel were likely to decompose H2O2. Three
`cathode
`geometries
`(Fig. 1),
`i.e.,
`plain
`plate
`(17.78 cm  15.24 cm, or 7 in  6 in), plate with pro-
`truded short ‘‘fingers’’ (17.78 cm  15.24 cm  1.02 cm,
`or 7 in  6 in  0.4 in) and plate with protruded long
`(17.78 cm  15.24 cm  1.52 cm,
`‘‘fingers’’
`or
`7 in
` 6 in  0.6 in) were employed to investigate the effect
`of cathode surface area. Unless otherwise stated, all
`experiments were conducted using the long-finger plate.
`A graphite plain plate (8.89 cm  15.24 cm, or 3.5 in
` 6 in) was used as the anode. Copper wires were
`connected to both electrodes through Teflon screws. The
`connections were carefully sealed with silicon to prevent
`copper
`electro-corrosion. Compressed oxygen gas
`(99.6%) and air were used as DO sources. The gas was
`sparged into the catholyte through a porous pipe-
`diffuser placed right under the cathode. The catholyte
`was pre-saturated with DO by purging pure oxygen gas
`or air for 15 min before electrolysis was initiated. The
`catholyte pH was controlled by a pH-stat (Model pH-
`40, New Brunswick Scientific Co., Edison, NJ), HClO4
`(1 M) and NaOH (1 M) solutions. A high performance
`combination pH probe (Cat. No. 376490, Corning Inc.,
`Corning, NY) was placed behind the cathode to avoid
`the interference from the electrical field. The solution
`
`temperature was controlled by a thermostat (Model EX-
`200, Brookfield Engineering Laboratories, Inc., Stought-
`on, MA) and a water bath.
`Polarization curves were obtained by cyclic voltam-
`metry using a three-electrode bi-potentiostat (Model
`AFRDE4, Pine Instrument Co., Grove City, PA). A
`saturated calomel electrode (SCE) was used as the
`reference electrode. The SCE was inserted into a Luggin
`capillary filled with saturated KCl solution. By placing
`the tip of the Luggin capillary in contact with the
`cathode surface, the cathodic potential can be accurately
`controlled against the SCE. The cathodic potential of
`the parallel-plate electrolyzer was swept from 0 to –0.8 V
`(vs. SCE) at a linear rate of 33.3 mV/s. Transient current
`response was recorded by a Hewlett Packard X-Y
`recorder (Model 7001A, Moseley Division, Pasadena,
`CA). The electro-generation of H2O2 experiments were
`carried out under either constant potential or constant
`current mode. Usually, constant potential mode is used
`to derive fundamental electrochemical information in
`laboratory scale experiments, while constant current
`mode is commonly used in industrial electrolysis because
`it is technically much easier to control the current than
`the potential [29].
`In this study, the factors significantly affecting the
`limiting current of H2O2 generation, including cathodic
`potential, oxygen purity and mass flow rate, and cathode
`surface area were investigated using the constant
`potential mode. Under this mode, the electrical current
`was monitored on-line by a digital multimeter (Model
`
`

`

`88
`
`Z. Qiang et al. / Water Research 36 (2002) 85–94
`
`22-183A, Tanday Co., Fort Worth, Texas). The effect of
`pH was studied using both constant potential and
`constant current modes. The effects of temperature and
`inert supporting electrolyte concentration were investi-
`gated by the constant current mode. A regulated DC
`power
`supply (Model WP-705B, Vector-Vid Inc.,
`Horsham, PA) was employed to provide constant
`current. The concentration of H2O2 was determined by
`the titanic sulfate [Ti(SO4)2] method [16]. A diode array
`spectrophotometer (Model 8452A, Hewlett Packard)
`was used to measure the light absorbance of the Ti4+–
`H2O2 orange complex at 410 nm. The DO concentration
`was determined by an oxygen electrode (Model 97-08-
`99, Orion Research Inc., Beverly, MA). A pH meter was
`used to record the DO concentration in the range of 0–
`14 mg/l. When the concentration exceeded the upper
`response limit, dilution was made with deoxygenated
`distilled water.
`
`3. Results and discussion
`
`3.1. Stability of hydrogen peroxide
`
`In a thoroughly clean container without the presence
`of any catalysts, H2O2 is very stable at any concentra-
`tion. However, the presence of a trace amount of metal
`ions in the solution or on the container surface will lead
`to the decomposition of H2O2. In practice, stabilizing
`agents such as sodium stannate, 8-hydroxyquinoline and
`sodium pyrophosphate are commonly used to stabilize
`H2O2 for long-term storage [10]. The self-decomposition
`rate of H2O2
`is primarily influenced by pH and
`temperature. Sudoh et al. [16] attributed the low current
`efficiency (6.88%) of H2O2 generation in alkaline
`solutions at 301C to a high self-decomposition rate of
`H2O2. Solution pH influences the chemical speciation of
`both H2O2 and trace metals. The trace metals that can
`catalytically initiate the self-decomposition of H2O2 may
`be introduced from acid and base used for pH
`adjustment, from container surfaces, and even from
`distilled water. At low pH, H2O2 and free metal ions
` and metal-hydroxo
`predominate. At high pH, HO2
`complexes are the major
`species. Therefore,
`it
`is
`necessary to investigate the effects of pH and tempera-
`ture on the self-decomposition of H2O2.
`The H2O2 solution with an initial concentration of
`150 mg/l was prepared by diluting a commercial grade
`H2O2 solution (31.5% by weight) with distilled water. A
`series of plastic bottles (250 ml) were washed with 1 M
`HClO4 solution, and then filled with the H2O2 solution.
`The pH was adjusted by reagent grade HClO4 or NaOH
`to cover the range of 1–13. The effect of temperature was
`investigated at 101C, 231C and 501C. At selected time
`intervals, the concentration of H2O2 was determined.
`Fig. 2 shows the self-decomposition of H2O2 at various
`
`Fig. 2. Stability of hydrogen peroxide.
`
`pH values, temperatures and reaction times. Results
`relatively stable at pHo9.
`indicate that H2O2
`is
`However, above pH 9, H2O2 decomposes markedly
`with increasing pH, temperature and reaction time.
`There is complete H2O2 decomposition at pH 13 and
`temperature 501C after 96 h. According to Schumb et al.
`[10], even with the purest H2O2 and at elevated
`temperatures, the decomposition of H2O2 in the liquid
`phase is not a homogeneous autodecomposition process
`increasing temperature
`of the H2O2 itself. Generally,
`increases the reaction rate. The self-decomposition of
`H2O2 at high pH and elevated temperatures are
`attributed in part to the catalytic effect of the container
`walls and the reagent impurities. Another aspect is the
` in the base catalyzedanion, HO2. The role of HO2
`
`
`H2O2 decomposition was
`suggested by Abel
`[30]
`following the reaction:
`H2O2 þ HO
`-H2O þ O2 þ OH:
`
`2
`However, a low temperature (e.g., 101C) suppresses
`H2O2 self-decomposition, even at high pH. It is also
`noted that
`in the acidic region,
`the highest
`self-
`decomposition rate appears at pH 3. It is known that
`pH 3 is the optimal value for the Fenton’s reagent. The
`possible presence of trace metals in the solution or on
`the bottle surfaces may catalytically stress the decom-
`position of H2O2 at pH 3. It is noted that the electrolysis
`was conducted in acidic conditions, i.e., pHp4.0, and
`the electrolysis time was 2 h. Therefore, based on the
`results shown in Fig. 2,
`it
`is clear that
`the self-
`decomposition of H2O2 would be insignificant.
`
`ð1Þ
`
`3.2. Optimal cathodic potential
`
`In acidic solutions, the dissolved oxygen is electro-
`chemically reduced to H2O2 at the cathode
`O2 þ 2Hþ þ 2e-H2O2; Eo ¼ 0:440 V vs: SCE:
`
`ð2Þ
`
`

`

`Z. Qiang et al. / Water Research 36 (2002) 85–94
`
`89
`
`continuously increases with increasing H2O2 concentra-
`tion.
`In the limiting current region, H2O2 generation is
`controlled by the mass transfer of DO through the
`cathode-solution diffusion layer, rather than by the
`electron transfer between DO and cathode. Since the
`DO concentration at
`the cathode surface rapidly
`approaches zero after electrolysis starts, the limiting
`current under a steady-state condition can be expressed
`by the following equation for macroscopic electrodes
`[29]:
`IL ¼ kmnFAeC *;
`where IL represents the limiting current ðAÞ; km is the
`mass transfer coefficient (m/s), n is the stoichiometric
`number of electrons transferred, F is the Farady’s
`constant (96,490 C/mol), Ae
`is the effective cathode
`surface area ðm2Þ; and C is the DO concentration in
`bulk solution ðMÞ: The mass transfer coefficient, km; can
`be determined by Eq. (7):
`km ¼ D=d
`
`ð6Þ
`
`ð7Þ
`
`Two side reactions simultaneously occur at the cathode:
`(1) the reduction of H2O2 to H2O due to the accumula-
`tion of H2O2 at the cathode-solution interface, and (2)
`the hydrogen gas evolution. These reactions are shown
`as follows:
`H2O2þ2Hþþ2e-2H2O; Eo ¼ 1:534 V vs: SCE;
`
`ð3Þ
`
`2Hþþ2e-H2; Eo ¼ 0:242 V vs: SCE:
`
`ð4Þ
`
`At the anode, the oxidation of H2O releases oxygen gas
`and protons
`2H2O-4HþþO2þ4e; Eo ¼ 0:987 V vs: SCE:
`
`ð5Þ
`
`The protons so generated will be driven to the catholyte
`electro-statically and partially supplement the protons
`consumption during the synthesis of H2O2.
`Polarization curves reflect transient current response
`with respect to cathodic potential ðEcÞ applied. Results
`in Fig. 3 indicate that at Eco0:15 V,
`the current
`rapidly with increasing Ec:
`density (i)
`increases
`However, a ‘‘plateau’’ appears in the range of 0.15
`to 0.5 V. This ‘‘plateau’’ represents the limiting current
`region for the electro-generation of H2O2 (Reaction 2).
`When the Ec continues to increase above 0.5 V, the i
`quickly rises again. It implies a significant reduction of
`H2O2 to H2O (Reaction 3) and an enhanced H2
`evolution (Reaction 4). Pure oxygen gas provides a
`higher DO concentration than air, thereby yielding a
`higher i: Fig. 3 also shows that the initial equilibrium
`potential, E0; is 0.075 V at i ¼ 0: If a constant potential
`of 0.5 V is applied,
`the initial overpotential
`is
`calculated as 0.425 V. As electrolysis proceeds, the
`overpotential becomes more negative since the E0
`
`Fig. 3. Polarization curves of pure oxygen and air sparging.
`Experimental
`conditions:
`completely mixing,
`sweeping
`rate=33.3 mV/s; pH=2; T ¼ 231C; QO2 ¼ 8:2  102 mol/min;
`Qair ¼ 8:2  102 mol/min;
`ionic strength=0.05 M NaClO4;
`long-finger plate cathode.
`
`where D represents the diffusion coefficient of oxygen
`(m2/s) and d is the thickness of diffusion layer ðmÞ: Fig. 3
`shows that the maximum limiting current is approxi-
`mately located in the range of 0.4–0.5 V ðEcÞ:
`Current efficiency ðZÞ; defined as the ratio of the
`electricity consumed by the electrode reaction of interest
`over the total electricity passed through the circuit, can
`be calculated by Eq. (8):
`R
`Z ¼ nFCH2O2 V
`I dt
`
`100%;
`
`t 0
`
`ð8Þ
`
`where CH2O2 represents H2O2 concentration in bulk
`solution ðMÞ and V is the catholyte volume ðLÞ: By
`definition, the Z actually represents an overall current
`efficiency over a certain period of electrolysis time.
`The effect of cathodic potential was investigated from
`0.2–0.9 V ðEcÞ: Figs. 4a–c shows the time-dependent
`changes of H2O2 concentration, current density and
`current efficiency at various applied potentials, respec-
`tively. Results indicate that at 0:2p Ecp0:5 V, the
`H2O2 concentration increases linearly with reaction time
`(Fig. 4a). The slope represents a constant net generation
`rate of H2O2 ðgnÞ throughout the whole electrolysis
`course. Correspondingly, Fig. 4b shows that i stabilize
`quickly after the electrolysis is initiated. A steady-state
`condition is rapidly reached because a constant DO
`concentration is maintained in the solution. Fig. 4c
`indicates a slight decrease of Z during electrolysis
`because the reduction of H2O2 to H2O is gradually
`promoted by the accumulation of H2O2. Results further
`indicate that at Ec > 0:5 V, both gn and Z decrease
`notably with reaction time (Figs. 4a and c). It implies
`that a high Ec stresses the reduction of H2O2 as well as
`the evolution of H2. Fig. 4b shows that at 0.6 V, the i
`
`

`

`90
`
`Z. Qiang et al. / Water Research 36 (2002) 85–94
`
`Fig. 4. Generation of H2O2 at various applied cathodic potentials: (a) accumulated concentration; (b) current density; (c) current
`conditions: pH=2; T ¼ 231C;
`efficiency;
`(d)
`average
`current density
`and average
`current
`efficiency. Experimental
`QO2 ¼ 8:2  102 mol/min; ionic strength=0.05 M NaClO4; long-finger plate cathode.
`
`exhibits a slight increase with reaction time. Meanwhile,
`a remarkable increase of i is observed at 0.8 and 0.9 V.
`Though more electricity is consumed at high Ec; a
`higher fraction of which is wasted by side reactions. The
`maximum H2O2 concentration is obtained at 0.6 V, i.e.,
`79 mg/l after electrolysis for 2 h. However, based on Z; a
`constant potential of 0.5 V (vs. SCE) is the optimal
`i at all applied
`cathodic potential. The high initial
`potentials in Fig. 4b is caused by the pre-saturated DO
`concentration on the cathode surface. By plotting the
`average i and average Z vs. Ec (Fig. 4d), it is clear that
`the optimal cathodic potential is 0.5 V vs. SCE, and
`the corresponding i and Z are 6.4 A/m2 and 81%,
`respectively. The i obtained at the optimal potential, i.e.,
`6.4 A/m2, is called ‘‘the limiting current density’’.
`
`3.3. Effect of oxygen purity and mass flow rate
`
`Pure oxygen gas (99.6%) and air were used as the
`sources of DO. Both oxygen purity and mass flow rate
`
`affect the limiting current. The equilibrium DO con-
`centration is proportional to the oxygen partial pressure
`in the supply gas. For example, the equilibrium DO
`concentrations were measured as 8.3 and 39.3 mg/l when
`sparging air and pure oxygen at pH 2 and 0.05 M
`NaClO4, respectively. Fig. 5a shows the accumulation of
`H2O2 as a function of reaction time at various oxygen
`purities and mass flow rates. Results clearly indicate that
`gn increases with increasing oxygen flow rate until a rate
`of 8.2  102 mol/min is achieved. Further increase in
`the flow rate does not change the gn: Results also
`indicate that the gn is much smaller when air is sparged.
`Fig. 5b shows the effect of mass flow rate on the average
`i and average Z: In accordance with Fig. 5a, the average i
`reaches the maximum value of 6.4 A/m2 at the oxygen
`flow rate of 8.2  102 mol/min. Doubling the flow rate
`will not increase the i any more. It is seen that a rate of
`8.2  102 mol/min is adequate to maintain the highest
`steady-state DO concentration during electrolysis. Since
`the optimal potential was applied (Ec ¼ 0:5 V vs.
`
`

`

`Z. Qiang et al. / Water Research 36 (2002) 85–94
`
`91
`
`reached the steady-state condition (ca. 20 min). Fig. 6
`shows that a linear relationship is achieved between the
`limiting current and the effective surface area. The slope
`of 6.4 A/m2 is just the limiting current density. Results
`clearly indicate that
`the limiting current density is
`independent of the cathode geometry and surface area
`applied. It is obvious that increasing the surface area is a
`convenient way to raise the limiting current, and
`consequently, the net generation rate of H2O2. The
`km;
`is
`calculated
`as
`mass
`transfer
`coefficient,
`2.70  105 m/s using Eq. (6) and a C value of
`39.3 mg/l. Moreover, by assuming the oxygen diffusion
`coefficient ðDÞ as 2.0  109 m2/s [31], the thickness of
`the diffusion layer ðdÞ is calculated as about 74 mm using
`Eq. (7). Both km and d provide a detailed microscopic
`insight into the electrolytic synthesis of H2O2.
`
`3.5. Effect of pH
`
`From Eq. (2), it seems that a low pH is favorable for
`the electro-generation of H2O2
`since its
`synthesis
`consumes protons. However, a high proton concentra-
`tion may promote H2 evolution and reduce the current
`efficiency. Fig. 7 shows the effect of pH on gn and
`average Z in the constant current mode (i ¼ 6:4 A/m2).
`Results indicate that pH 2 is the optimal condition.
`Above pH 2, the gn decreases due to insufficient protons.
`Below pH 2, the gn decreases again due to enhanced H2
`evolution. Because pH 3 may stress the decomposition
`of H2O2 by trace metals as previously mentioned, the gn
`at pH 3 is a little smaller than at other pH values. Since
`the current was holding constant, the average Z curve
`exhibits a similar trend as the gn curve. The highest Z is
`84% at pH 2, and the lowest is 69% at pH 3.
`
`Fig. 6. Effect of cathode surface area on limiting current.
`conditions: Ec ¼ 0:5 V vs. SCE; pH=2;
`Experimental
`T ¼ 231C; QO2 ¼ 8:2  102 mol/min;
`ionic strength=0.05 M
`NaClO4.
`
`Fig. 5. Generation of H2O2 with pure oxygen or air at various
`mass flow rates: (a) accumulated concentration; (b) effect of
`mass flow rate on average current density and average current
`efficiency. Experimental conditions: Ec ¼ 0:5 V vs. SCE;
`pH=2; T ¼ 231C; ionic strength=0.05 M NaClO4; long-finger
`plate cathode.
`
`SCE), a high Z could be achieved for all flow rates, i.e.,
`80–90%. If air is sparged at a rate of 8.2  102 mol/min
`(or 1.7  102 mol O2/min), the average i decreases to
`2.1 A/m2, while the average Z slightly increases to 90%
`(Fig. 5b).
`
`3.4. Effect of cathode surface area
`
`i.e., plain
`Three graphite electrode configurations,
`plate,
`short- and long-finger plate were used to
`investigate the effect of cathode surface area. It should
`be pointed out that H2O2 was only generated at one side
`of the cathode that faces the applied electrical field. The
`effective surface areas were 271, 415 and 488 cm2 for the
`plain plate, short- and long-finger plate, respectively.
`The potential was controlled at 0.5 V vs. SCE, and the
`oxygen mass flow rate was 8.2  102 mol/min. The
`electrical current was measured after the electrolysis
`
`

`

`92
`
`Z. Qiang et al. / Water Research 36 (2002) 85–94
`
`Fig. 7. Effect of pH on net generation rate and average current
`i ¼ 6:4 A/m2; T ¼ 231C;
`efficiency. Experimental conditions:
`QO2 ¼ 8:2  102 mol/min;
`ionic strength=0.05 M NaClO4;
`long-finger plate cathode.
`
`temperature on net generation rate and
`Fig. 8. Effect of
`average current efficiency. Experimental conditions: i ¼ 6:4 A/
`m2; pH=2; QO2 ¼ 8:2  102 mol/min; ionic strength=0.05 M
`NaClO4; long-finger plate cathode.
`
`The effect of pH was also investigated using the
`constant potential mode at Ec ¼ 0:5 V vs. SCE (data
`not shown). Results indicate that both gn and i increases
`with decrease in pH, but the highest Z (81%) is still
`achieved at pH 2. Therefore, it is concluded that pH 2 is
`the optimal value for H2O2 generation in acidic
`solutions.
`
`3.6. Effects of temperature and supporting electrolyte
`
`Temperature exerts conflicting effects on H2O2 gen-
`eration. As temperature rises,
`the oxygen diffusion
`coefficient will increase, resulting in an increase of gn:
`However, increasing temperature will decrease the DO
`solubility and increase the H2O2 decomposition rate,
`thereby decreasing gn: If the current is held constant, the
`formation rate of H2O2 will not change appreciably, but
`the decomposition rate of H2O2 will
`increase as
`temperature rises. The effect of
`temperature was
`investigated from 131C to 331C using the constant
`current mode (i ¼ 6:4 A/m2). Fig. 8 demonstrates that gn
`slightly decreases with increasing temperature. The
`average Z shows a similar trend, decreasing from 92%
`at 131C to 81% at 331C. It is seen that the H2O2
`generation is favored at low temperatures.
`The effect of inert supporting electrolyte concentra-
`tion was also investigated using the constant current
`mode. Fig. 9 shows that
`in the region of NaClO4
`concentration less than 0.1 M, the total potential drop
`between cathode and anode ðDEÞ decreases notably with
`increasing NaClO4 concentration. In the >0.1 M region,
`the DE decreases slowly due to the decrease of ion
`activity coefficients. Since NaClO4 does not participate
`the gn remains almost
`in electrochemical reactions,
`constant in the concentration range studied. Do and
`Chen [24]
`investigated the oxidative degradation of
`
`Fig. 9. Effect of NaClO4 concentration on total potential drop
`and net generation rate. Experimental conditions: i ¼ 6:4 A/m2;
`pH=2; T ¼ 231C; QO2 ¼ 8:2  102 mol/min; long-finger plate
`cathode.
`
`formaldehyde with electro-generated H2O2 and reported
`that the formaldehyde degradation was not affected by
`the concentration of supporting electrolyte, Na2SO4, in
`the range of 0–0.5 M. The results indicate that the
`background ionic strength in wastewater may be directly
`used as supporting electrolyte for on-site H2O2 genera-
`tion.
`It has been reported that chloride or bromide can
`promote H2O2 generation [32,33]. Since the SCE
`reference was filled with saturated KCl solution, the
`diffusion of chloride from the SCE into catholyte might
`enhance H2O2 generation and skew the experimental
`data. The experimental
`results
`from the constant
`potential mode (Ec ¼ 0:5 V vs. SCE, with SCE inserted
`in the catholyte) and from the constant current mode
`(i ¼ 6:4 A/m2, without SCE) are compared. During 2 h
`
`

`

`Z. Qiang et al. / Water Research 36 (2002) 85–94
`
`93
`
`of electrolysis, the total electricity consumed is very
`close, i.e., 2,257 C for the constant potential mode versus
`2,232 C for the constant current mode. The gn and Z are
`0.64 mg/l min and 81% (Figs. 4a and d), and 0.68 mg/
`l min and 84% (Fig. 7), correspondingly. It is seen that
`the presence of the SCE in the catholyte does not
`promote H2O2 generation. The diffusion of chloride is
`negligible, mainly due to the large reactor volume with a
`relatively short electrolysis time.
`
`4. Conclusions
`
`The major objective of this study is to improve the
`Faradic current efficiency of H2O2 generation in acidic
`solutions containing dilute supporting electrolyte by
`optimizing the operational parameters. Based on the
`experimental results presented above,
`the following
`conclusions are drawn:
`
`3.7. Potential profile and energy consumption
`
`Fig. 10 shows the potential profile in the electrolyzer
`at Ec ¼ 0:5 V vs. SCE. Results indicate that the DE
`consists of three parts, i.e., the potential drops at the
`electrode–solution interfaces caused by activation polar-
`ization and concentration polarization of DO, and the
`IR drop in the electrolyte solution. In the solution,
`Ohm’s law is well obeyed since the electrolyte concen-
`tration remains almost constant. The electrical resis-
`tance of the cation exchange membrane is negligible,
`which means a prompt migration of protons from
`anolyte to catholyte. A constant anolyte pH was
`observed during the course of electrolysis although
`protons were being continuously generated at the anode.
`Under the experimental conditions stated in Fig. 10,
`the unit energy consumption for H2O2 generation is
`calculated as 7.8 kW h/kg H2O2.
`
`* Significant self-decomposition of H2O2 is observed
`only at high pH (>9) and elevated temperatures
`(>231C).
`* The optimal conditions for H2O2 generation are
`cathodic potential of 0.5 V (vs. SCE), oxygen mass
`flow rate of 8.2  102 mol/min, and pH 2. Under the
`optimal conditions, the average current density and
`average current efficiency are 6.4 A/m2 and 81%,
`respectively. When air is used, the average current
`density decreases to 2.1 A/m2, while the average
`current efficiency slightly increases to 90%.
`i.e., 6.4 A/m2,
`* The limiting current density,
`is
`independent of the cathode geometry and surface
`area applied.
`* H2O2 generation is
`tures.
`* In the concentratio