Electrochemical Modulation of Odorant Molecule: A Study of p-cresol

p-Cresol modulation was for the ﬁ rst time evaluated as an alternative option for odor control in sanitation facilities. Results indicate that the oxidation of p-cresol can generate 4-hydroxybenzaldehyde (4-HB), a molecule with a sweet-woody odor, following the introduction of chloride ions into the supporting electrolyte. In an attempt to impede electrode fouling, pulsed chronoamperometry (CA) was implemented and resulted in ∼ 10% higher p-cresol removal compared to CA at constant potential. Boron doped diamond (BDD) was also explored as an alternative working electrode. p-Cresol oxidation on the diamond surface resulted in higher removal percentages, but the desired oxidation product was not detected by Liquid chromatography – mass spectrometry (LC-MS) likely due to complete combustion. The

Malodor nuisance is a major risk factor for user adoption of effective sanitation technologies because odor can influence the perception of cleanliness and hygiene in sanitation facilities. 1,2 p-Cresol is one of the key odorant molecules found in the headspace of used toilets, 1 and has an animal, nutty, sewage characteristic smell. 1 However, it can be oxidized 3,4 to organic compounds found in vanilla beans 5,6 4-hydroxybenzyl alcohol (4-HBA) and 4-hydroxybenzaldehyde (4-HB). 4-HBA and 4-HB are described as positive odors such as vanilla-like, sweet, 7 and vanilla-like, biscuit, 7,8 respectively. Therefore, this oxidation signifies an improvement in overall odor perception and a positive contribution to the characteristic smell of sanitation facilities by modulating a malodorant molecule to a pleasant-smelling one.
To date, absorbing malodor molecules with activated carbon, 9 masking odors with perfumes, and displacing odors with fans are common techniques used to address malodor nuisance. However, these methods have limited effectiveness and do not irreversibly eliminate the odorant molecules. Most odorant molecules are volatile, small carbon chains with functional groups such as carboxylic acid, aldehyde, or alcohol. A significant change in olfactory perception can be achieved by either modifying the functional group or changing the odorant concentration. [10][11][12] This raises the opportunity to use an electrochemical approach to selectively modify the structure of odorants molecules into their more pleasant-smelling homologs. This paper expands upon preliminary results that led to the patent number WO/2019/099391, 13 which describes the development of a device that can drag ambient air into an electrochemical cell where targeted odorant molecules (i) undergo a reaction and then, (ii) are released back into the air after modification/modulation. Such device could be placed in sanitation facilities as an alternative solution for the remediation of malodor nuisance. Using electrochemistry to mitigate malodor has several advantages such as in situ process, odorant molecules selectivity and modulation, and low maintenance compared to off-the-shelf methods. To explore this possibility p-cresol was chosen here as a case study. p-Cresol can be chemically 5,14-17 and electrochemically 3,4,18,19 oxidized to 4-HBA, 4-HB, and other intermediate side products (e.g. 4-hydroxybenzyl methyl ether and 4-hydroxybenzoic acid).
One of the principal challenges of p-cresol oxidation (as with other phenolic compounds) is electrode fouling that occurs due to the formation of a polymeric film on the electrode surface. [20][21][22][23][24][25][26][27] Fouling has been reported in literature for various electrode materials such as platinum, 25,28,29 boron doped diamond (BDD), 25 glassy carbon, 30 SnO 2 , 31 and graphite. 23,32,33 Despite having great catalytic properties provided by surface functional groups, excellent thermal and chemical stability, and a low fabrication cost, 34 graphite fouls due to the adsorption of insoluble electrogenerated products on its surface. 33,35 Consequently, the reaction rate rapidly diminishes, hindering efficient oxidation during bulk electrolysis. Hence, strategies to inhibit electrode fouling need to be explored. As an alternative to constant potential (chronoamperometry, CA), pulsed techniques have been proposed to achieve higher removal rates 36,37 and to inhibit anodic passivation. 36 In contrast to the electrode fouling in graphite, BDD allows electrode preservation and regeneration. BDD has advantages such as a wide potential window, inert surface with low adsorption properties 38 and surface regeneration by pre-anodization. 18,39,40 When bulk electrolysis is carried out at high anodic potentials electrode deactivation can be inhibited, which is shown by high steady-state current density. 40 Another alternative option to improve oxidation rate is to introduce Cl − into the supporting electrolyte. 41,42 Indirect oxidation with chloride in the supporting electrolyte has been reported to enhance oxidation kinetics. The aromatic ring of phenolic compounds can readily react with chlorine species at low pH during oxidation. 42,43 At low pH, chloride is oxidized at the anode to form chlorine. The gas is further hydrolyzed to form chlorine species (e.g. hypochlorous acid), which are powerful oxidants that contribute to the mediated oxidation of organic compounds. 42 The scope of this study is to investigate the electrochemical oxidation pathway of p-cresol. The specific objectives are: (1) to evaluate graphite as a working electrode for p-cresol oxidation and characterize the oxidation products; (2) to evaluate the contribution of indirect oxidation by chlorine species to p-cresol removal and the p-cresol oxidation pathway; (3) to explore pulsed CA as a strategy to inhibit graphite electrode fouling; and (4) to assess BDD as an alternative electrode material.

Experimental
Electrolyte.-In all experiments 250 mM H 2 SO 4 was prepared by mixing deionized water with 1.0 N standardized solution H 2 SO 4 (Alfa Aesar) and used as supporting electrolyte. A pH 0.5 was measured across all solutions. Such low pH allows chlorine species (e.g. hypochlorous acid) to readily react with the aromatic ring of z E-mail: mariana.vasquez@duke.edu *Electrochemical Society Student Member. phenolic compounds. 42,43 In addition, most familiar bacteria in sanitation facilities, like Escherichia coli, 44,45 staphylococci, 46 and Salmonella spp. 47,48 are neutrophiles, which means that they optimally grow at pH values between 5 and 8. Therefore, operating in a low-pH medium could limit bacterial growth and extend the electrochemical cell lifetime when placed in sanitation facilities. p-Cresol (Sigma-Aldrich ⩾99%, FG W233706-Sample-K) was diluted in 250 mM H 2 SO 4 to obtain the following p-cresol concentrations: 23.9, 47.8, and 95.6 mM. To investigate the influence of chloride ions on the reaction kinetics and oxidation pathway, sodium chloride (Macron, crystal purity ⩾99%) was added to the supporting electrolyte in the following concentrations: 50, 157, 500, and 1000 mM.
Boron-doped diamond (BDD) on polished diamond films (thickness ∼1.8 μm doped, 30.1 μm undoped) a piece with geometric area of 0.44 cm 2 was cleaved from a 10.16 cm RT_NP08 wafer purchased from sp 3 Diamond Technologies. The cleaved piece was electrically connected to a thin copper wire using silver paste (Ted Pella, PN#16031); the contact was left to dry for several hours at room temperature. To isolate the contact from the electrolyte, a small glass tube was placed over the copper wire and covered with non-conductive epoxy (Loctite EA 9462 Hysol).
Electrochemical conditions.-Experiments were conducted using a 3-electrode setup in a divided electrochemical cell consisting of two 30 ml compartments (one anodic, and the other cathodic) connected at the bottom by a tube (length ∼ 1 cm, diameter 0.5 cm). Ag/AgCl sat. KCl (Bio-logic RE-1CP) and Pt mesh were used as the reference and the counter electrodes, respectively. All working electrodes (BDD and graphite) were submerged into the electrolyte just a few prior to characterization or electrolysis. Electrode  passivation due to phenol oxidation is due to the irreversible combination of two radicals to form dimeric products leading to the formation of dimers and oligomers. 20,27,49,50 This process results in the formation of a polymeric layer on the electrode surface. Therefore, we do not expect significant (if any) passivation to occur at open circuit. All measurements were made using an SP-300 Bio-logic potentiostat. During the kinetics study, Cottrell's equation (Eq. 1) was used to calculate the apparent number of electrons exchanged during p-cresol oxidation. To use Eq. 1, the experimental set-up ensured linear diffusion on the flat electrode surface in an undisturbed electrolyte. In Eq. 1, n is the apparent number of electrons exchanged, A is the electrode area of ∼2 cm 2 , F is Faraday's constant, D o is the p-cresol diffusion coefficient 9.14 cm 2 s −1 , 51 C o is the p-cresol concentration, and t is the step time of 5 s. Liquid chromatography mass spectrometer (LC-MS) analysis.-All samples were 10× diluted with 50% MeOH before characterization. p-Cresol, 4-hydroxybenzaldehyde (Sigma-Aldrich ⩾ 97%, FG W398403-Sample-K), and 4-hydroxybenzyl alcohol (Sigma-Aldrich ⩾ 98%, FG W398705-Sample-K) calibration curves were made for Agilent LC-MS-TOF with Phenomenex Kinetics EVO C18 column (3 × 30 mm, 2.6 u). The mobile phase A was 100:3:0.3 water: MeOH: formic acid, mobile phase B was 100:3:0.3 MeOH: water: formic acid, and flow rate was 0.5 ml min −1 . Initial hold at 0% for 0.5 min, then 0%-100% from 0.5-8 min, for a total 15 min run time. Column temperature was 40°C, the injection volume was 1 μl, UV was set at 220, 254 nm, and MS positive ion. p-Cresol eluted at ∼5.

Results and Discussion
Voltammetry study.-p-Cresol electrochemical activity was evaluated using cyclic voltammetry (CV). Figure 1a shows voltammograms performed in H 2 SO 4 and in the presence of p-cresol. p-Cresol oxidation peak (P a1 ) appears at 0.88 V, and its intensity decreases after the first cycle to remain smaller for all subsequent scans. This P a1 magnitude decrease is attributed to electrode surface fouling. Electrode fouling during the electrochemical oxidation of phenolic compounds has been extensively investigated; this known phenomenon results from the deposition of a polymeric film onto the electrode surface. 20,23,53,54 Figure 1b shows the ratio of p-cresol oxidation charge over the charge transferred during the first cycle, Q/Q first . Oxidation charge was obtained by integrating the area under the oxidation peak identified in Fig. 1a. After a few scans, Q/Q first seems to reach a plateau, suggesting surface stabilization. However, only 2% of the initial charge Q first can be transferred after 10 cycles, which may impede practical application if a fouling mitigation strategy cannot be implemented. Change in p-cresol concentration during the CV experiment was ruled out as a reason for P a1 decrease because assuming a 100% faradaic efficiency, only 0.03% of the initial p-cresol concentration can been oxidized (see supplementary material).
Kinetics study.-p-Cresol oxidation kinetics were studied using double step chronoamperometry/chronocoulometry (DSC) to delineate the effects of mass transport and charge transfer. DSC experiments were carried out in H 2 SO 4 with various p-cresol concentrations and at three different anodic potentials. The apparent number of electrons exchanged, n, was calculated using Cottrell's equation (Eq. 1). The value of n provides information about the possible p-cresol oxidation pathway, oxidation products, and ratelimiting factors. Identifying the rate-limiting factors is crucial to determine the desired parameters for efficient bulk electrolysis and practical application. Figure 2a shows that n decreases with the increasing p-cresol concentration regardless of the applied potential. However, at a fixed p-cresol concentration, n increases as the potential becomes more anodic. Although fouling seems to happen considerably in more concentrated electrolytes (lower n values), these results indicate that the fouling surface deposits (possibly product intermediates) may be further oxidized at higher anodic potential (increased n values). We Scheme 1. p-Cresol oxidation pathway to 4-HBA. also hypothesized that the fouling mechanism involves the blockage of active sites where p-cresol oxidation occurs. To test this hypothesis, the charge ratio of the second step forward over the first one is plotted as a function of p-cresol concentration (Fig. 2b). Lower Q 2 /Q 1 values (i.e., greater fouling) are obtained at higher p-cresol concentrations. Therefore, fouling seems to be less dependent on the applied potential than the increasing p-cresol concentration, which indicates that the latter is the preponderant factor in electrode fouling.
p-Cresol bulk electrolysis.-Bulk electrolysis aims at altering pcresol's olfactive perception by decreasing its concentration 11,12 while generating 4-HBA. Scheme 1 proposes a simple 2-electron reaction to generate 4-HBA from p-cresol oxidation. Values of n shown in Fig. 2a indicate which oxidation products are most likely to be generated by direct oxidation. This p-cresol oxidation pathway to 4-HBA is a 2 e − process (Scheme 1). 4-HBA is hypothesized to be generated under the conditions at which n is greater or equal than 2.
In Fig. 2a, it is observed that n ⩾ ∼2 is mostly obtained for 23.9 mM p-cresol, which justifies why this concentration was chosen for bulk electrolysis. Figure 3a shows that the p-cresol concentration decreased by ∼20% across all three potentials. Similar removal percentages after 2 h of electrolysis indicate that electrode fouling (i.e., the blocking of p-cresol oxidation sites) may occur rapidly, irrespective of the applied potential (Fig. 2b). 4-HBA was only detected by LC-MS post-electrolysis at 1.7 V (Fig. 3b). Electrolysis at potentials lower than 1.7 V likely leads to other products not of interest for the present study. To confirm that p-cresol removal was substantially due to electrolysis, gas stripping experiments were conducted with both air and argon and results indicate less than ∼5% decrease in p-cresol concentration. Furthermore, no oxidation products of interest were detected (see Fig. S1 is available online at stacks.iop. org/JES/167/135501/mmedia).
Reaction pathway modulation.-Presence of Cl − in supporting electrolyte.-It has been reported that the oxidation kinetics of phenolic compounds can be enhanced by the presence of chlorine species. 42,43 Cl − was added to the supporting electrolyte to assess its contribution to p-cresol removal by indirect oxidation. In Fig. 4a, Cl − oxidation potential (P a2 ) was identified at ∼1.6 V. Increasing Cl − concentration increases the electrolyte conductivity, thus leading to an increase in current density for all redox peaks. p-Cresol oxidation charge increases and starts to plateau at high Cl − concentrations because chloride oxidation is a mass transfercontrolled process 55 (Fig. 4b). Therefore, the increase in current density with increasing conductivity causes a lower chloride concentration in the electrode vicinity, which causes reactions that are not controlled by mass transfer, such as water oxidation, to be favored. 55 p-Cresol electrolysis was performed at potentials below and above P a2 to confirm chlorine species contribution to p-cresol removal by indirect oxidation. p-Cresol removal after electrolysis at 1.1 V and 1.4 V is slightly higher than that obtained without Cl − in solution (Fig. 5a). The slight increase in p-cresol removal could be attributed to the increase in conductivity (Fig. 4b), which favors charge transfer. p-Cresol removal after electrolysis at 1.7 V with Cl − in the electrolyte results in an additional ∼20% p-cresol removal compared to electrolysis at 1.7 V without Cl − (Fig. 5a). This large difference in p-cresol removal is likely due to the indirect oxidation of p-cresol by chlorine species generated by Cl − oxidation at potentials greater than 1.6 V.
LC-MS was used to quantify the production of 4-HB and 4-HBA from p-cresol oxidation in the presence of chloride in solution. Results from Fig. 5b show that only 4-HB can be detected after electrolysis at 1.7 V. Based on the product characterization of the bulk electrolysis described above, a reaction mechanism is proposed in which p-cresol can be electrochemically oxidized to 4-HBA (k 1 ) and chemically oxidized to 4-HB (k 4 ). Further, 4-HBA can be oxidized by chlorine species to form 4-HB (k 3 ) (Scheme 2). To evaluate this hypothesis the following rate equations were considered: = Chlorine concentration was considered constant because of the high concentration of Equation 6 indicates that p-cresol concentration decreases exponentially as a function of [Cl 2 ], while Eq. 7 indicates that 4-HB concentration increases linearly with time and its production rate depends both on the chloride oxidation rate (k 2 ) and the initial concentration of Cl − in solution. This is in line with the LC-MS results that show the absence of 4-HB in the absence of chloride in the electrolyte (Fig. 3d).
In Fig. 6, LC-MS results show an exponential decrease of p-cresol and a linear increase of 4-HB over time as predicted by Eqs. 6 and 7, respectively. Note that p-cresol concentration is plotted in a log scale, therefore using a linear fit of the experimental values, an initial concentration of 20.1 mM was obtained, which is close to the LC-MS quantified initial p-cresol concentration of 22.8 mM. A value of 4.5 × 10 -3 s −1 (or 1.5 × 10 -2 M s −1 ) was obtained for k 2 (Eq. 7) using the slope value from 4-HB data linear fit. The rate constant for chloride oxidation to chlorine has been reported elsewhere to be 2.19 × 10 -3 s −1 58 and 7.7 × 10 -2 cm 3 mol −1 . 59 Therefore, the proposed rate equations are consistent with LC-MS data and known literature. Cl − inclusion in the supporting electrolyte for electrolysis at potentials greater than 1.6 V not only results in higher p-cresol removal, but also modifies the oxidation pathway to render the generation of 4-HB possible.
Chronoamperometry (CA) pulsed treatment.-Electrode fouling impedes the continuous and efficient electrolysis of p-cresol at a  constant potential. Pulsed chronoamperometry (CA) has been proposed as a mitigation strategy to achieve higher removal rates compared to constant current 36,37 and to inhibit anodic passivation. 36 Pulsed CA was explored as a method to inhibit electrode fouling during p-cresol bulk electrolysis. As shown in Fig. 7, pulsed CA maintains higher current density values than constant CA over time.
Pulsed methods allow reaction products to diffuse away from the electrode surface (preventing accumulation and further electrode fouling) and reactant-depleted regions at the vicinity of the electrode to be replenished with electroactive species. p-Cresol removal after pulsed CA electrolysis increased by 10%-20% compared to constant potential (Fig. 8a), which demonstrates that this alternative method offers a practical strategy to mitigate electrode fouling. LC-MS results do not indicate the presence of 4-HB, which is expected because of the absence of chloride ions, as demonstrated in the previous section. However, it is interesting to note that 4-HBA is detected across all potentials. By using pulses, reaction intermediates seem not to irreversibly adsorb on the electrode surface, thus allowing more p-cresol to be oxidized at low overpotential (e.g., at 1.1 V). An increased concentration of intermediates in the vicinity of the electrode therefore leads to the generation of 4-HBA. A follow-up study is recommended to vary the duty cycle to gain insights into the fouling kinetics. Such knowledge will help optimize the pulse duration with respect to 4-HBA generation.
Boron doped diamond (BDD) electrode.-BDD is an alternative carbon electrode that has been extensively employed for the oxidation of phenolic compounds. [60][61][62] Combustion (i.e. complete conversion to CO 2 and H 2 O) 63 of phenolic compounds is mainly carried out by electrogenerated hydroxyl radicals. 38,64 Scheme 3 proposes a simplified mechanism for p-cresol combustion at high anodic potentials. 38,64 The p-cresol oxidation peak (P a1 ) was identified at 1.3 V (Fig. 9a). The P a1 current intensity decreases after the first cycle and remains low for all subsequent cycles due to electrode surface fouling as shown in the insert of Fig. 9a 25,40,61,62,65 In Fig. 9b, Q/Q o decreases exponentially, indicating continuous fouling over sequential cycles inhibiting p-cresol oxidation. BDD and graphite fouling exhibit similar kinetics. It is important to note that the BDD electrode geometric area was 0.44 cm 2 , 137 times smaller than graphite's geometric area used in all previous experiments. However, as shown in Fig. 9c, the BDD electrode can maintain high current densities at high anodic potentials and generate hydroxyl radicals that inhibit fouling. 39,40,61 Figure 10 shows that the p-cresol concentration decreases by 12%−14% when using BDD as the working electrode. p-Cresol oxidation is carried out mainly by the electro-generated hydroxyl radicals and therefore is limited by their interaction with p-cresol molecules. Cañizares et al. 60 showed that increasing current density (e.g., by increasing the applied potential- Fig. 9c) does not increase the rate of oxidation of organics, but mostly favors the anodic side reactions such as oxygen generation. 60 4-HB and 4-HBA were not detected after oxidation because these side products are further oxidized to CO 2 . Using BDD as the working electrode presents the    advantage of bulk electrolysis without fouling. However, in this case, the reduction of p-cresol does not correlate with the generation of 4-HB or 4-HBA.

Conclusions
Modulation of the p-cresol oxidation pathway was achieved by changing the electrode material, the electrolyte composition, and using a pulsed chronoamperometry method. It was further demonstrated that the generation of 4-HB at a graphite electrode was only possible in the presence of chloride ions in solution. Pulsed chronoamperometry resulted in ∼10% higher p-cresol removal compared to chronoamperometry at a constant potential. Proposed as a fouling mitigation strategy, the pulsed CA method also led to the generation of 4-HBA, otherwise not detected. Although p-cresol  oxidation on BDD resulted in even higher removal percentages, the desired oxidation products were not detected by LC-MS due to complete combustion. These insights pave the way to the development a technology that converts malodorants into their counterparts with improved smell-pleasantness.