Double-Cabin Galvanic Cell-Synthesizing Nanoporous, Flower-like, Pb-Containing Pd–Au Nanoparticles for Nonenzymatic Formaldehyde Sensor

In this work, a novel formaldehyde sensor was constructed based on nanoporous, flower-like, Pb-containing Pd–Au nanoparticles deposited on the cathode in a double-cabin galvanic cell (DCGC) with a Cu plate as the anode, a multiwalled carbon nanotube-modified glassy carbon electrode as the cathode, a 0.1 M HClO4 aqueous solution as the anolyte, and a 3.0 mM PdCl2 + 1.0 mM HAuCl4 + 5.0 mM Pb(ClO4)2 + 0.1 M HClO4 aqueous solution as the catholyte, respectively. Electrochemical studies reveal that the stripping of bulk Cu can induce underpotential deposition (UPD) of Pb during the galvanic replacement reaction (GRR) process, which affects the composition and morphology of Pb-containing Pd–Au nanoparticles. The electrocatalytic activity of Pb-containing nanoparticles toward formaldehyde oxidation was examined in an alkaline solution, and the experimental results showed that formaldehyde mainly caused direct oxidation on the surface of Pb-containing Pd–Au nanoparticles while inhibiting the formation of CO poison to a large degree. The proposed formaldehyde sensor exhibits a linear amperometric response to formaldehyde concentrations from 0.01 mM to 5.0 mM, with a sensitivity of 666 μA mM−1 cm−2, a limit of detection (LOD) of 0.89 μM at triple signal-to-noise, rapid response, high anti-interference ability, and good repeatability.


Introduction
Formaldehyde is a health-hazard chemical widely applied in many industrial processes; for instance [1], it is used as a raw material to manufacture building plates and lacquer materials, and as an intermediate for producing detergents, soaps, cosmetics, and so forth [2].In addition, formaldehyde can also be released into the drinking water by ozonation and chlorination of some natural organic matters during recycling of industrial or domestic wastewater [3].The extensive residue of formaldehyde in the environment causes a risk of chronic intoxication for people exposed in the pollution zone via daily ingestion of the harmful chemical, which is a serious threat to human health [4].The International Agency for Research on Cancer has also classified formaldehyde as a human carcinogen [5].Therefore, it is important to develop a rapid and highly sensitive method to monitor the trace concentrations of formaldehyde in potentially contaminated fields.
Up to now, various analytic methods have been explored to determine formaldehyde concentrations at trace levels; for instance, spectrophotometry [6], gas chromatography [7], and high-performance liquid chromatography [8].However, these methods usually suffer from the drawbacks of high cost, being time consuming, and complex operation, limiting their use in monitoring trace concentrations of formaldehyde in real time.On the other hand, electrochemical sensors have the advantages of low cost, simplicity, sensitivity, and on-the-spot detection.Intensive studies have been devoted to exploiting high-performance formaldehyde electrochemical sensors [9][10][11].For instance, Yinuo Zhang et al. reported a formaldehyde electrochemical biosensor based on the selective electrochemical oxidation of formaldehyde catalyzed by aldehyde dehydrogenases (ALDHs) on a Cu electrode, with a limit of detection (LOD) of 1.46 × 10 −15 M [9].Shimomura et al. developed an enzymatic biosensor based on mesoporous silica materials (FSM8.0 or P123-M) for immobilizing formaldehyde dehydrogenase (FDH), and the LOD was reduced to 1.2 µM [12].Although the enzymatic biosensor exhibits excellent sensitivity and good selectivity toward formaldehyde detection, the inevitable shortcomings that arise from the intrinsic nature of enzymes, including chemical and thermal instability as well as a short lifetime, hinder their wide application.At present, nonenzymatic electrochemical organic small-molecule sensorbased nanomaterials have been widely studied [13,14].Therefore, preference is given to nanomaterial-modified, non-enzymatic sensors, owing to their high electrocatalytic activity towards formaldehyde.For instance, nanostructured Pd-Pt or Pd-Au electrocatalysts have been adopted to construct nonenzymatic formaldehyde sensors [15,16].However, there are two key problems associated with nonenzymatic formaldehyde sensors: first, low sensitivity originating from the sluggish kinetics of formaldehyde electrooxidation on a poisoned interface; and second, poor selectivity derived from the oxidation of formaldehyde and other electroactive species in the same potential range, which should be addressed before practical applications.Hence, improving the performance of electrode materials in many ways is an important idea to address the limitations on the development of non-enzymatic formaldehyde electrochemical sensors.
As is well known, multicomponent metal nanoparticles always exhibit enhanced electrocatalytic performance compared with the corresponding monometallic nanoparticles, owing to the synergistic effects between different components [17,18].Palladium-based nanomaterials have also been widely applied in electrocatalytic oxidation and electrochemical detection of formaldehyde molecules [19,20].Among the various palladium-based nanoparticles, Pd-Au bimetallic nanoparticles can strongly absorb light and efficiently enhance the conversion of several organic reactions under visible light irradiation at ambient temperatures [21].In addition, effectively controlling the morphology of nanoparticles can further improve their performance, as the electrochemical reaction pathways are highly sensitive to the surface atomic arrangement exposed on special facets.A variety of Pd-Au nanoparticles with shapes such as hexoctahedral nanocrystals [22], porous flower [23], coreshell shape [24], etc., have been prepared and investigated.The relatively large surface areas and highly active facets on their surfaces endow the prepared nanoparticles with enhanced catalytic efficiencies.
To date, various strategies, for example, wet-chemical synthesis, self-assembly, electrodeposition, the galvanic replacement reaction (GRR), etc., have been explored to prepare metal micro/nanostructures of specific morphologies [25][26][27][28].Among these, the ecofriendly GRR receives particular attention because of its tunability, economy, and simplicity.Typically, the GRR-based synthesis involves the replacement of relatively active metal templates (for example, Cu and Ag nanostructures) by target noble metal ions, as driven by the difference in electrochemical potential (ECP) between metal templates (lower ECP) and noble metal ions (higher ECP) [29].The conventional GRR synthetic process always works in a manner such that the anode and cathode coexist in a one-pot solution, in which the metal templates' impurities and the time-consuming, post-purifying step are inevitable.However, physically separating the cathodic and anodic processes of the GRR in a double cabin galvanic cell (DCGC) can overcome the above-mentioned disadvantages for synthesizing nanomaterials; moreover, the substrate-supported nanostructured materials synthesized in the DCGC can be used directly as prepared, avoiding the post-purifying step, but such a convenient DCGC step based on the GRR principle is rarely attempted in fabricating highly efficient nanoelectrocatalysts with controllable morphologies and compositions.
Here, a novel formaldehyde sensor is designed and fabricated using the nanoporous, flower-like, Pb-containing Pd-Au nanoparticles synthesized in the DCGC as the electrochemical sensing element.The developed Pb-containing Pd-Au nanoparticles exhibit remarkable catalytic activity toward formaldehyde electrooxidation and efficiently suppress the formation of CO poison, which is beneficial for constructing a high-performance formaldehyde sensor.

Fabrication and Characterization of Pb-Containing Pd-Au Nanoparticles
Figure 1 shows the discharging current (i cell ) and potential of the cathode during the deposition of metallic nanoparticles on the MWCNTs/GCE (cathode) based on the GRR principle.After connecting the anode and cathode by a conducting Cu wire in the DCGC, the ECP difference between the anode part (ca.0.01 V for Cu plate in 0.1 M HClO 4 ) and the cathode part (ca.0.78 V for MWCNTs/GCE in a noble metal mixed salts solution) derived the occurrence of GRR, resulting in the stripping of the bulk Cu (Cu bulk ) anode and deposition of noble metal nanoparticles onto the MWCNTs/GCE cathode.During deposition of Pd 3 Au 1 from the Pb(ClO 4 ) 2 -free 3.0 mM PdCl 2 + 1.0 mM HAuCl 4 + 0.1 M HClO 4 catholyte (Figure 1(b)), the steady-state i cell and potential of the MWCNTs/GCE (E MWCNTs/GCE ) were 27 µA and 0.068 V, respectively.However, for deposition of Pb-Pd 3 Au 1 from the 3.0 mM PdCl 2 + 1.0 mM HAuCl 4 + 5.0 mM Pb(ClO 4 ) 2 + 0.1 M HClO 4 catholyte (Figure 1(a)), the steady-state i cell and E MWCNTs/GCE were 48 µA and 0.098 V, both of which are lager than those for deposition of Pd 3 Au 1 , perhaps implying the enhanced deposition of Pd and Au in the presence of Pb(ClO 4 ) 2 .
above-mentioned disadvantages for synthesizing nanomaterials; moreover, the substrate-supported nanostructured materials synthesized in the DCGC can be used directly as prepared, avoiding the post-purifying step, but such a convenient DCGC step based on the GRR principle is rarely attempted in fabricating highly efficient nanoelectrocatalysts with controllable morphologies and compositions.
Here, a novel formaldehyde sensor is designed and fabricated using the nanoporous, flower-like, Pb-containing Pd-Au nanoparticles synthesized in the DCGC as the electrochemical sensing element.The developed Pb-containing Pd-Au nanoparticles exhibit remarkable catalytic activity toward formaldehyde electrooxidation and efficiently suppress the formation of CO poison, which is beneficial for constructing a high-performance formaldehyde sensor.

Fabrication and Characterization of Pb-Containing Pd-Au Nanoparticles
Figure 1 shows the discharging current (icell) and potential of the cathode during the deposition of metallic nanoparticles on the MWCNTs/GCE (cathode) based on the GRR principle.After connecting the anode and cathode by a conducting Cu wire in the DCGC, the ECP difference between the anode part (ca.0.01 V for Cu plate in 0.1 M HClO4) and the cathode part (ca.0.78 V for MWCNTs/GCE in a noble metal mixed salts solution) derived the occurrence of GRR, resulting in the stripping of the bulk Cu (Cubulk) anode and deposition of noble metal nanoparticles onto the MWCNTs/GCE cathode.During deposition of Pd3Au1 from the Pb(ClO4)2-free 3.0 mM PdCl2 + 1.0 mM HAuCl4 + 0.1 M HClO4 catholyte (Figure 1(b)), the steady-state icell and potential of the MWCNTs/GCE (EMWCNTs/GCE) were 27 μA and 0.068 V, respectively.However, for deposition of Pb-Pd3Au1 from the 3.0 mM PdCl2 + 1.0 mM HAuCl4 + 5.0 mM Pb(ClO4)2 + 0.1 M HClO4 catholyte (Figure 1(a)), the steady-state icell and EMWCNTs/GCE were 48 μA and 0.098 V, both of which are lager than those for deposition of Pd3Au1, perhaps implying the enhanced deposition of Pd and Au in the presence of Pb(ClO4)2.

Figure 1.
Simultaneous responses of icell, and Ecathode in the DCGC with a Cu plate as the anode, the MWCNTs/GCE as the cathode, and 0.1 M HClO4 as anolyte (stationary), while the catholytes (stationary) were 3.0 mM PdCl2 + 1.0 mM HAuCl4 +5.0 mM Pb(ClO4)2 + 0.1 M HClO4 aqueous solution (a, for deposition of Pb-Pd3Au1) and 3.0 mM PdCl2 + 1.0 mM HAuCl4 + 0.1 M HClO4 aqueous solution (b, for deposition of Pd3Au1), respectively.Short-circuit discharging of the DCGC (connecting the anode and cathode by a conducting Cu wire) was from 30 s to 270 s. Figure 2 shows SEM images of the MWCNTs/GCE-supported Pd 3 Au 1 and Pb-Pd 3 Au 1 metallic nanoparticles, and SEM image of the MWCNTs/GCE is also shown in Figure S1.Spindly MWCNTs can be observed uniformly dispersed on the GCE before the modification of the nanoparticles.For the Pd 3 Au 1 nanoparticles (panels A and B), walnut kernel-like nanoparticles with rough surfaces were distributed on the MWCNTs, with diameters ranging from 60 nm to 230 nm.However, for Pb-Pd 3 Au 1 nanoparticles (panels C and D), nanoporous flowers of 70~500 nm in diameter were formed on the MWCNTs.In fact, close inspection of the larger Pb-Pd 3 Au 1 nanoparticles shows that flower-like nanoparticles are built up by many smaller nanoparticles with non-isotropic structures, and nanochannels or nanovoids between the smaller nanoparticles are present on the surface.EDS characterization indicated that Pd and Au were deposited on the surfaces of the MWCNTs (Figure 3), and the atomic percentages of Pd and Au for the Pd 3 Au 1 nanoparticles were determined to be 48% and 52% (Figure 3A), respectively, which was un-proportional to that of the feeding solution (mole ration for Pd/Au of 3:1 in the catholyte).Because of the higher ECP of Au versus that of Pd, the reaction kinetics of reducing AuCl 4 − are expected to be much faster than that of reducing PdCl 2 , thus leading to the deposition of more Au in the Pd 3 Au 1 nanoparticles, even at a high Pd/Au mole ratio in the catholyte.In addition, the component of Pb was detected in Pb-Pd 3 Au 1 nanoparticles (Figure 3B), and the atomic percentage of Pd, Au, and Pb were determined to be 35%, 56%, and 9.0%, respectively, which indicates the occurrence of cathodic deposition of Pb from the Pb(ClO 4 ) 2 -containing catholyte during the GRR deposition of Pb-Pd 3 Au 1 nanoparticles, though bulk Pb (Pb bulk ) is more active than Cu bulk (anode here).Here, it is interesting that deposition of Pb not only induces the formation of nanoporous flowers but also changes the content of Pd and Au in Pb-Pd 3 Au 1 nanoparticles (35% Pd and 56% Au in Pb-Pd 3 Au 1 nanoparticles versus 48% Pd and 52% Au in Pd 3 Au 1 nanoparticles).Virtually, previous studies have also reported that the additives of transition metal salts (for instance CuCl 2 , AgNO 3 ) can affect the growth of nanostructured noble metal materials [30,31].
Pb-Pd3Au1 metallic nanoparticles, and SEM image of the MWCNTs/GCE is also shown in Figure S1.Spindly MWCNTs can be observed uniformly dispersed on the GCE before the modification of the nanoparticles.For the Pd3Au1 nanoparticles (panels A and B), walnut kernel-like nanoparticles with rough surfaces were distributed on the MWCNTs, with diameters ranging from 60 nm to 230 nm.However, for Pb-Pd3Au1 nanoparticles (panels C and D), nanoporous flowers of 70~500 nm in diameter were formed on the MWCNTs.In fact, close inspection of the larger Pb-Pd3Au1 nanoparticles shows that flower-like nanoparticles are built up by many smaller nanoparticles with non-isotropic structures, and nanochannels or nanovoids between the smaller nanoparticles are present on the surface.EDS characterization indicated that Pd and Au were deposited on the surfaces of the MWCNTs (Figure 3), and the atomic percentages of Pd and Au for the Pd3Au1 nanoparticles were determined to be 48% and 52% (Figure 3A), respectively, which was un-proportional to that of the feeding solution (mole ration for Pd/Au of 3:1 in the catholyte).Because of the higher ECP of Au versus that of Pd, the reaction kinetics of reducing AuCl4 − are expected to be much faster than that of reducing PdCl2, thus leading to the deposition of more Au in the Pd3Au1 nanoparticles, even at a high Pd/Au mole ratio in the catholyte.In addition, the component of Pb was detected in Pb-Pd3Au1 nanoparticles (Figure 3B), and the atomic percentage of Pd, Au, and Pb were determined to be 35%, 56%, and 9.0%, respectively, which indicates the occurrence of cathodic deposition of Pb from the Pb(ClO4)2-containing catholyte during the GRR deposition of Pb-Pd3Au1 nanoparticles, though bulk Pb (Pbbulk) is more active than Cubulk (anode here).Here, it is interesting that deposition of Pb not only induces the formation of nanoporous flowers but also changes the content of Pd and Au in Pb-Pd3Au1 nanoparticles (35% Pd and 56% Au in Pb-Pd3Au1 nanoparticles versus 48% Pd and 52% Au in Pd3Au1 nanoparticles).Virtually, previous studies have also reported that the additives of transition metal salts (for instance CuCl2, AgNO3) can affect the growth of nanostructured noble metal materials [30,31].

Electrochemical Study of the Evolution of Cathodic Deposition of Pb during the GRR Process
It is important to study the evolution of cathodic deposition of Pb in the DCGC with a Cubulk anode, since the ECP of Pbbulk is lower than that of Cubulk, thus the stripping of Cubulk-induced deposition of Pbbulk is thermodynamically impossible during GRR deposition of Pb-Pd3Au1.Here, we conducted the cyclic voltammetry (CV) experiments to examine the electrochemical behaviors of Pb on electroplated Pd-modified Au (Pdpla/Au, as shown in Figure S2) or Au electrodes via EQCM method, as shown in Figure 4.For the Pdpla/Au electrode, the decrease of f0 started at 0.46 V, corresponding to the occurrence of underpotential deposition (UPD) of Pb onto Pdpla/Au [32], and the UPD of Pb (PbUPD) suppressed hydrogen adsorption/desorption onto Pdpla/Au, to a large degree, in the potential range of 0~−0.22 V.The UPD of Pb ended at −0.45 V where the deposition/stripping of Pbbulk occurred, as reflected by the great decrease/increase of f0 at −0.45 V.For the Au electrode, the decrease of f0 started at 0.06 V, also corresponding to the occurrence of UPD of Pb on Au [33], and two pairs of distinct UPD peaks at 0.01/−0.01V and −0.23/−0.25 V were observed.The deposition/stripping of Pbbulk occurred at ca. −0.45 V, resulting in a great decrease/increase of f0 at −0.45 V. From the CV experimental results, it is obvious that deposition/stripping of PbUPD occurs at more positive potentials versus that of Pbbulk, which means that Pb 2+ /PbUPD has a higher ECP than Pb 2+ /Pbbulk.Furthermore, both the ECP of Pb 2+ /PbUPD on Pd (exclusively for 0.46~0.01V) and ECP of Pb 2+ /PbUPD on Au (exclusively for 0.06~0.01V) are even higher than that of Cu 2+ /Cubulk (0.01 V here, Figure S3), implying that the stripping of Cubulk-induced deposition of PbUPD onto Pd or Au is thermodynamically possible.

Electrochemical Study of the Evolution of Cathodic Deposition of Pb during the GRR Process
It is important to study the evolution of cathodic deposition of Pb in the DCGC with a Cu bulk anode, since the ECP of Pb bulk is lower than that of Cu bulk , thus the stripping of Cu bulk -induced deposition of Pb bulk is thermodynamically impossible during GRR deposition of Pb-Pd 3 Au 1 .Here, we conducted the cyclic voltammetry (CV) experiments to examine the electrochemical behaviors of Pb on electroplated Pd-modified Au (Pd pla /Au, as shown in Figure S2) or Au electrodes via EQCM method, as shown in Figure 4.For the Pd pla /Au electrode, the decrease of ∆f 0 started at 0.46 V, corresponding to the occurrence of underpotential deposition (UPD) of Pb onto Pd pla /Au [32], and the UPD of Pb (Pb UPD ) suppressed hydrogen adsorption/desorption onto Pd pla /Au, to a large degree, in the potential range of 0~−0.22 V.The UPD of Pb ended at −0.45 V where the deposition/stripping of Pb bulk occurred, as reflected by the great decrease/increase of ∆f 0 at −0.45 V.For the Au electrode, the decrease of ∆f 0 started at 0.06 V, also corresponding to the occurrence of UPD of Pb on Au [33], and two pairs of distinct UPD peaks at 0.01/−0.01V and −0.23/−0.25 V were observed.The deposition/stripping of Pb bulk occurred at ca. −0.45 V, resulting in a great decrease/increase of ∆f 0 at −0.45 V. From the CV experimental results, it is obvious that deposition/stripping of Pb UPD occurs at more positive potentials versus that of Pb bulk , which means that Pb 2+ /Pb UPD has a higher ECP than Pb 2+ /Pb bulk .Furthermore, both the ECP of Pb 2+ /Pb UPD on Pd (exclusively for 0.46~0.01V) and ECP of Pb 2+ /Pb UPD on Au (exclusively for 0.06~0.01V) are even higher than that of Cu 2+ /Cu bulk (0.01 V here, Figure S3), implying that the stripping of Cu bulk -induced deposition of Pb UPD onto Pd or Au is thermodynamically possible.

Electrochemical Study of the Evolution of Cathodic Deposition of Pb during the GRR Process
It is important to study the evolution of cathodic deposition of Pb in the DCGC with a Cubulk anode, since the ECP of Pbbulk is lower than that of Cubulk, thus the stripping of Cubulk-induced deposition of Pbbulk is thermodynamically impossible during GRR deposition of Pb-Pd3Au1.Here, we conducted the cyclic voltammetry (CV) experiments to examine the electrochemical behaviors of Pb on electroplated Pd-modified Au (Pdpla/Au, as shown in Figure S2) or Au electrodes via EQCM method, as shown in Figure 4.For the Pdpla/Au electrode, the decrease of f0 started at 0.46 V, corresponding to the occurrence of underpotential deposition (UPD) of Pb onto Pdpla/Au [32], and the UPD of Pb (PbUPD) suppressed hydrogen adsorption/desorption onto Pdpla/Au, to a large degree, in the potential range of 0~−0.22 V.The UPD of Pb ended at −0.45 V where the deposition/stripping of Pbbulk occurred, as reflected by the great decrease/increase of f0 at −0.45 V.For the Au electrode, the decrease of f0 started at 0.06 V, also corresponding to the occurrence of UPD of Pb on Au [33], and two pairs of distinct UPD peaks at 0.01/−0.01V and −0.23/−0.25 V were observed.The deposition/stripping of Pbbulk occurred at ca. −0.45 V, resulting in a great decrease/increase of f0 at −0.45 V. From the CV experimental results, it is obvious that deposition/stripping of PbUPD occurs at more positive potentials versus that of Pbbulk, which means that Pb 2+ /PbUPD has a higher ECP than Pb 2+ /Pbbulk.Furthermore, both the ECP of Pb 2+ /PbUPD on Pd (exclusively for 0.46~0.01V) and ECP of Pb 2+ /PbUPD on Au (exclusively for 0.06~0.01V) are even higher than that of Cu 2+ /Cubulk (0.01 V here, Figure S3), implying that the stripping of Cubulk-induced deposition of PbUPD onto Pd or Au is thermodynamically possible.Since the steady-state E MWCNTs/GCE for deposition of Pb-Pd 3 Au 1 was 0.098 V (Figure 1a), matching well with the potential range for UPD of Pb on Pd (−0.45~0.46V), but higher than that for UPD of Pb on Au (−0.45~0.06V), the occurrence of UPD of Pb on Pd during GRR deposition of Pb-Pd 3 Au 1 is reasonable.Hence, the ECP difference between Pb 2+ /Pb UPD on Pd and Cu 2+ /Cu bulk can account for the cathodic deposition of Pb in the DCGC with a Cu bulk anode.Such a UPD effect of Pb can induce the growth of Pd and Au nanoparticles to form nanoporous, flower-like Pb-containing Pd 3 Au 1 nanostructures, the catalytic performance of which may also be improved via the synergistic interaction between Pd, Au, and Pb.

Electrochemical Behaviors and Electrocatalytic Responses to Formaldehyde on Modified Electrodes
Figure 5 shows the cyclic voltammograms (CVs) of several modified electrodes in 0.1 M KOH aqueous solution.The reduction peak of Pd oxides (P c-Pd ) appeared at −0.34 V on Pb-Pd/MWCNTs/GCE, accompanied by the peaks of hydrogen adsorption/desorption at −0.66~−1.0V.The reduction peak of Au oxides (P c-Au ) was observed at 0.13 V on Pb-Au/MWCNTs/GCE.For Pb-Pd 3 Au 1 /MWCNTs/GCE, the potential of P c-Au is consistent with that on Pb-Au/MWCNTs/GCE; however, the P c-Pd (−0.22 V) positively shifted by 120 mV versus that on Pb-Pd/MWCNTs/GCE (−0.34 V); similar potential shifting of P c-Pd was also observed in the electrochemical behaviors of Pd-Au alloy nanoparticles [22].In addition, compared with Pd 3 Au 1 /MWCNTs/GCE, Pb-Pd 3 Au 1 /MWCNTs/GCE did not produce distinct peaks of hydrogen adsorption/desorption, indicating that the electrochemical behaviors of Pb-Pd 3 Au 1 /MWCNTs/GCE depended on the unique morphology and the atomic arrangement of Pd, Au, and Pb on the surface.
Molecules 2024, 29, x FOR PEER REVIEW 6 of 15 Since the steady-state EMWCNTs/GCE for deposition of Pb-Pd3Au1 was 0.098 V (Figure 1(a)), matching well with the potential range for UPD of Pb on Pd (−0.45~0.46V), but higher than that for UPD of Pb on Au (−0.45~0.06V), the occurrence of UPD of Pb on Pd during GRR deposition of Pb-Pd3Au1 is reasonable.Hence, the ECP difference between Pb 2+ /PbUPD on Pd and Cu 2+ /Cubulk can account for the cathodic deposition of Pb in the DCGC with a Cubulk anode.Such a UPD effect of Pb can induce the growth of Pd and Au nanoparticles to form nanoporous, flower-like Pb-containing Pd3Au1 nanostructures, the catalytic performance of which may also be improved via the synergistic interaction between Pd, Au, and Pb.

Electrochemical Behaviors and Electrocatalytic Responses to Formaldehyde on Modified Electrodes
Figure 5 shows the cyclic voltammograms (CVs) of several modified electrodes in 0.1 M KOH aqueous solution.The reduction peak of Pd oxides (Pc-Pd) appeared at −0.34 V on Pb-Pd/MWCNTs/GCE, accompanied by the peaks of hydrogen adsorption/desorption at −0.66~−1.0V.The reduction peak of Au oxides (Pc-Au) was observed at 0.13 V on Pb-Au/MWCNTs/GCE.For Pb-Pd3Au1/MWCNTs/GCE, the potential of Pc-Au is consistent with that on Pb-Au/MWCNTs/GCE; however, the Pc-Pd (−0.22 V) positively shifted by 120 mV versus that on Pb-Pd/MWCNTs/GCE (−0.34 V); similar potential shifting of Pc-Pd was also observed in the electrochemical behaviors of Pd-Au alloy nanoparticles [22].In addition, compared with Pd3Au1/MWCNTs/GCE, Pb-Pd3Au1/MWCNTs/GCE did not produce distinct peaks of hydrogen adsorption/desorption, indicating that the electrochemical behaviors of Pb-Pd3Au1/MWCNTs/GCE depended on the unique morphology and the atomic arrangement of Pd, Au, and Pb on the surface.For the electrooxidation of formaldehyde (Figure 6), during the forward scan, two oxidation peaks appeared at −0.58 and −0.28 V on Pb-Pd3Au1/MWCNTs/GCE, and the oxidation peak potentials were lower than those on Pd3Au1/MWCNTs/GCE (−0.46 and −0.11V) and Pb-Pd/MWCNTs/GCE (−0.34 and −0.14 V).No peaks for the electrooxidation of formaldehyde appeared on MWCNTs/GCE.The maximum oxidation peak current was reached on Pb-Pd3Au1/MWCNTs/GCE (3.06 mA cm −2 , −0.28 V), indicating the For the electrooxidation of formaldehyde (Figure 6), during the forward scan, two oxidation peaks appeared at −0.58 and −0.28 V on Pb-Pd 3 Au 1 /MWCNTs/GCE, and the oxidation peak potentials were lower than those on Pd 3 Au 1 /MWCNTs/GCE (−0.46 and −0.11V) and Pb-Pd/MWCNTs/GCE (−0.34 and −0.14 V).No peaks for the electrooxidation of formaldehyde appeared on MWCNTs/GCE.The maximum oxidation peak current was reached on Pb-Pd 3 Au 1 /MWCNTs/GCE (3.06 mA cm −2 , −0.28 V), indicating the highest catalytic activity of Pb-Pd 3 Au 1 /MWCNTs/GCE toward formaldehyde electrooxidation.During the backward scan, an oxidation peak appeared at 0.19 V on Pb-Pd 3 Au 1 /MWCNTs/GCE, similar to that observed on Pb-Au/MWCNTs/GCE.After the reduction of Pd oxides at more negative potential, another oxidation peak appeared at ca. −0.3 V on Pb-Pd 3 Au 1 /MWCNTs/GCE, corresponding to the oxidation of the accumulated CO poisonous intermediate generated through oxidation of formaldehyde during the forward scan [34], and the oxidation peak current (1.01 mA cm −2 ) was lower than that on Pd 3 Au 1 /MWCNTs/GCE (3.19 mA cm −2 , −0.27 V) and Pb-Pd/MWCNTs/GCE (1.56 mA cm −2 , −0.36 V).Previous studies reported that the oxidation mechanism of formaldehyde on palladium follows a dual pathway [24,35], including a direct oxidation pathway and an indirect oxidation pathway.Formaldehyde is directly oxidated to CO 2 at a low potential in the former, but indirectly oxidated to CO 2 via the chemisorbed CO intermediate at a high potential in the latter.The presence of oxidation peaks in the forward and backward scans indicates direct and indirect pathways of formaldehyde oxidation in the overall process.Here, the lower oxidation peak potentials (−0.58 and −0.28 V) and the maximum oxidation peak current (3.06 mA cm −2 , −0.28 V) on Pb-Pd 3 Au 1 /MWCNTs/GCE during the forward scan, as well as the lower oxidation peak current (1.01 mA cm −2 , −0.3 V) during the backward scan, prove that Pb-Pd 3 Au 1 /MWCNTs/GCE shows remarkably improved electrocatalytic activity toward formaldehyde oxidation, mainly through a direct oxidation pathway (suppressing the formation of CO poison) among the examined electrocatalysts.The enhanced performance on Pb-Pd 3 Au 1 /MWCNTs/GCE can be attributed to nanoporous flower-like Pb-Pd 3 Au 1 nanoparticles with highly active sites in greater quantity on the large areal surface, firstly; secondly, the favorable electronic effect and bifunctional effect between Pd, Au, and Pb atoms facilitate the direct oxidation of formaldehyde [36]; thirdly, facile mass transfer for formaldehyde and the oxidized products on the nanoporous surface of Pb-Pd 3 Au 1 nanoparticles.[34], and the oxidation peak current (1.01 mA cm −2 ) was lower than that on Pd3Au1/MWCNTs/GCE (3.19 mA cm −2 , −0.27 V) and Pb-Pd/MWCNTs/GCE (1.56 mA cm −2 , −0.36 V).Previous studies reported that the oxidation mechanism of formaldehyde on palladium follows a dual pathway [24,35], including a direct oxidation pathway and an indirect oxidation pathway.Formaldehyde is directly oxidated to CO2 at a low potential in the former, but indirectly oxidated to CO2 via the chemisorbed CO intermediate at a high potential in the latter.The presence of oxidation peaks in the forward and backward scans indicates direct and indirect pathways of formaldehyde oxidation in the overall process.Here, the lower oxidation peak potentials (−0.58 and −0.28 V) and the maximum oxidation peak current (3.06 mA cm −2 , −0.28 V) on Pb-Pd3Au1/MWCNTs/GCE during the forward scan, as well as the lower oxidation peak current (1.01 mA cm −2 , −0.3 V) during the backward scan, prove that Pb-Pd3Au1/MWCNTs/GCE shows remarkably improved electrocatalytic activity toward formaldehyde oxidation, mainly through a direct oxidation pathway (suppressing the formation of CO poison) among the examined electrocatalysts.The enhanced performance on Pb-Pd3Au1/MWCNTs/GCE can be attributed to nanoporous flower-like Pb-Pd3Au1 nanoparticles with highly active sites in greater quantity on the large areal surface, firstly; secondly, the favorable electronic effect and bifunctional effect between Pd, Au, and Pb atoms facilitate the direct oxidation of formaldehyde [36]; thirdly, facile mass transfer for formaldehyde and the oxidized products on the nanoporous surface of Pb-Pd3Au1 nanoparticles.

Amperometric Sensing of Formaldehyde
To construct a formaldehyde amperometric sensor based on the Pb-containing Pd-Au nanoparticles, the effect of sensing conditions, such as applied potential and the atomic ratio of Pd to Au in the nanoparticles, were first optimized.Figure S4 shows the

Amperometric Sensing of Formaldehyde
To construct a formaldehyde amperometric sensor based on the Pb-containing Pd-Au nanoparticles, the effect of sensing conditions, such as applied potential and the atomic ratio of Pd to Au in the nanoparticles, were first optimized.Figure S4 shows the effects of the applied potential on the amperometric response of 50 µM HCHO for Pb-Pd 3 Au 1 /MWCNTs/GCE in 0.1 M KOH aqueous solution.From Figure S4, the maximum amperometric response was achieved under −0.27V; thus, we chose −0.27 V as the ap-plied potential to further optimize the atomic ratio of Pd to Au in the nanoparticles by adjusting the molar concentration ratio of PdCl 2 to HAuCl 4 in the catholyte, as shown in Figure S5.When the molar concentration ratio of PdCl 2 /HAuCl 4 was 3:1, the prepared Pb-Pd 3 Au 1 /MWCNTs/GCE gave the largest amperometric response toward formaldehyde electrooxidation.In addition, the amperometric response of Pb-Pd 3 Au 1 /MWCNTs/GCE was also superior to that of Pd 3 Au 1 /MWCNTs/GCE and Pb-Pd/MWCNTs/GCE, in accordance with the results from the above electrocatalytic experiments (Figure 6).
For the electrochemical determination of formaldehyde, the optimized Pb-Pd 3 Au 1 / MWCNTs/GCE electrode was employed as a working electrode, and amperometric measurements were performed at a constant applied potential of −0.27 V. Figure 7A shows the amperometric response of Pb-Pd 3 Au 1 /MWCNTs/GCE toward successive addition of formaldehyde into a stirred 0.1 M KOH aqueous solution.After adding formaldehyde, the steady-state signal was reached within 5 s, indicating the rapid response of Pb-Pd 3 Au 1 /MWCNTs/GCE towards the change in the concentration of formaldehyde.Figure 7B shows the calibration curve of the amperometric response.The sensor exhibits a linear detection range (LDR) to formaldehyde concentration from 0.01 mM to 5.0 mM, and the linear regression equation is ∆i(µA) = 47.31C(mM)+ 3.17 (R 2 = 0.998), with a sensitivity of 666 µA mM −1 cm −2 .The LOD was estimated to be 0.89 µM at the triple signal-to-noise ratio.Comparison with the previous reported non-enzymatic formaldehyde sensors, as listed in Table 1, the sensor proposed here shows good performance in terms of the suitably wide LDR and the lower LOD.effects of the applied potential on the amperometric response of 50 μM HCHO for Pb-Pd3Au1/MWCNTs/GCE in 0.1 M KOH aqueous solution.From Figure S4, the maximum amperometric response was achieved under −0.27V; thus, we chose −0.27 V as the applied potential to further optimize the atomic ratio of Pd to Au in the nanoparticles by adjusting the molar concentration ratio of PdCl2 to HAuCl4 in the catholyte, as shown in Figure S5.When the molar concentration ratio of PdCl2/HAuCl4 was 3:1, the prepared Pb-Pd3Au1/MWCNTs/GCE gave the largest amperometric response toward formaldehyde electrooxidation.In addition, the amperometric response of Pb-Pd3Au1/MWCNTs/GCE was also superior to that of Pd3Au1/MWCNTs/GCE and Pb-Pd/MWCNTs/GCE, in accordance with the results from the above electrocatalytic experiments (Figure 6).
For the electrochemical determination of formaldehyde, the optimized Pb-Pd3Au1/MWCNTs/GCE electrode was employed as a working electrode, and amperometric measurements were performed at a constant applied potential of −0.27 V. Figure 7A shows the amperometric response of Pb-Pd3Au1/MWCNTs/GCE toward successive addition of formaldehyde into a stirred 0.1 M KOH aqueous solution.After adding formaldehyde, the steady-state signal was reached within 5 s, indicating the rapid response of Pb-Pd3Au1/MWCNTs/GCE towards the change in the concentration of formaldehyde.Figure 7B shows the calibration curve of the amperometric response.The sensor exhibits a linear detection range (LDR) to formaldehyde concentration from 0.01 mM to 5.0 mM, and the linear regression equation is Δi(μA) = 47.31C(mM)+ 3.17 (R 2 = 0.998), with a sensitivity of 666 μA mM −1 cm −2 .The LOD was estimated to be 0.89 μM at the triple signal-to-noise ratio.Comparison with the previous reported non-enzymatic formaldehyde sensors, as listed in Table 1, the sensor proposed here shows good performance in terms of the suitably wide LDR and the lower LOD.

Interference Tests, Repeatability, and Stability of the Sensor
To evaluate the anti-interference ability of the sensor in the accurate determination of formaldehyde, the effects of potentially interfering species, including methanol, ethanol, formic acid, n-propanol, and acetone, were tested.Figure 8 shows the amperometric response of Pb-Pd 3 Au 1 /MWCNTs/GCE to successive addition of 50 µM HCHO (a), 0.5 mM CH 3 OH (b), 0.5 mM C 2 H 5 OH (c), 0.5 mM HCOOH (d), 0.5 mM C 3 H 8 O (e), 0.5 mM C 3 H 6 O (f), and 50 µM HCHO (g) into the stirred 0.1 M KOH aqueous solution at a potential of −0.27 V.A fast current response occurred after adding formaldehyde; however, adding methanol, ethanol, formic acid, n-propanol, and acetone did not cause the obvious current changes, indicating that the tested species showed almost no interference to formaldehyde detection.These results confirmed the highly selective response of the Pb-Pd 3 Au 1 /MWCNTs/GCE toward formaldehyde.mM CH3OH (b), 0.5 mM C2H5OH (c), 0.5 mM HCOOH (d), 0.5 mM C3H8O (e), 0.5 mM C3H6O (f), and 50 μM HCHO (g) into the stirred 0.1 M KOH aqueous solution at a potential of −0.27 V.A fast current response occurred after adding formaldehyde; however, adding methanol, ethanol, formic acid, n-propanol, and acetone did not cause the obvious current changes, indicating that the tested species showed almost no interference to formaldehyde detection.These results confirmed the highly selective response of the Pb-Pd3Au1/MWCNTs/GCE toward formaldehyde.
Repeatability of the sensor was examined by measuring the amperometric responses of Pb-Pd3Au1/MWCNTs/GCE toward 15 successive detections of 50 μM formaldehyde in 0.1 M KOH aqueous solution, and the current response for each detection was closed to 2.6 μA, yielding a relative standard deviation (RSD) of 5.3%, indicating a good repeatability.In addition, the fabrication repeatability was also examined via measuring the amperometric responses of five Pb-Pd3Au1/MWCNTs/GCE electrodes prepared independently by the same procedure; the RSD is 4.6% for the 50 μM formaldehyde amperometric responses, demonstrating the reliability of the fabrication procedure.Repeatability of the sensor was examined by measuring the amperometric responses of Pb-Pd 3 Au 1 /MWCNTs/GCE toward 15 successive detections of 50 µM formaldehyde in 0.1 M KOH aqueous solution, and the current response for each detection was closed to 2.6 µA, yielding a relative standard deviation (RSD) of 5.3%, indicating a good repeatability.In addition, the fabrication repeatability was also examined via measuring the amperometric responses of five Pb-Pd 3 Au 1 /MWCNTs/GCE electrodes prepared independently by the same procedure; the RSD is 4.6% for the 50 µM formaldehyde amperometric responses, demonstrating the reliability of the fabrication procedure.
Long-term stability of the sensor was investigated by regularly measuring the amperometric responses of Pb-Pd 3 Au 1 /MWCNTs/GCE (placed in the refrigerator under 4 • C) toward 50 µM formaldehyde in two weeks.The current response was 97% of its initial response after two days, and 94% of its initial response after one week.With the experiment prolonged, a further decrease in the response current was observed, retaining about 90% of its initial response after two weeks, which demonstrates that the Pb-Pd 3 Au 1 /MWCNTs/GCE has a satisfactory stability for electrochemical detection of formaldehyde (as shown in Figure S6).

Real-Sample Analysis
To illustrate the feasibility of the sensor in a real-sample analysis, Pb-Pd 3 Au 1 /MWCNTs/ GCE was used to detect formaldehyde in tap water and lake water, respectively.The standard addition method was employed here, and the results are summarized in Table 2.The recoveries were in the range of 94.7-108.1%,and the RSD ranged from 3.4% to 5.5%, indicating that the sensor can be successfully applied for the practical electrochemical detection of formaldehyde in real samples.

Instrumentation and Reagents
All electrochemical experiments were conducted on an Autolab PGSTA 30 electrochemical workstation (Eco Chemie BV, Brunssum, The Netherlands) with the GPES 4.9 software.A conventional three-electrode electrolytic cell was used.A glassy carbon electrode (GCE) with a 3.0 mm diameter served as the working electrode (WE), KCl-saturated calomel electrode (SCE) with a salt bridge served as the reference electrode (RE), and a carbon rod electrode served as the counter electrode.All potentials herein are referenced to SCE.A research quartz crystal microbalance (QCM) (Maxtek, Ontario, CA, USA) was used for electrochemical QCM (EQCM) experiments, and 9 MHz QCM Au electrodes (0.29 cm 2 ) of keyhole configuration (Beijing Chenjing Electronic Co., Beijing, China) were used.The mass change on the electrode was calculated according to the Sauerbrey equation, ∆f 0 = −2.264× 10 −6 f 0g 2 ∆m/A, where ∆f 0 in Hz is the shift of the resonant frequency, ∆m in g is the mass change of the electrode, f 0g in Hz is the fundamental frequency in air, and A in cm 2 is the piezoelectrically active area [41].SEM characterization was performed on a Hitachi S-4800 high-resolution scanning electron microscope (Tokyo, Japan).Energy-dispersive X-ray spectroscopy (EDS, Oxford, UK) was used for elemental analysis of the electrocatalyst.
For the experiments conducted in the DCGC, the external-loading-free discharging current of the galvanic cell (i cell ) was dynamically monitored with the electrochemical noise (ECN) module of the Autolab PGSTA 30 electrochemical workstation [42] (Tokyo, Japan).Before the experiments, the veracity of the ECN device was checked in advance by examining the discharging curves of a 1.5 V dry battery at an external loading of 100 kΩ, yielding a V cell value of 1.501 V and an i cell value of 15.01 µA, respectively, in accordance with Ohm's law.A high-resistance (R ≥ 10 12 Ω) potentiometer attached to the pH meter (Leici, Shanghai, China) was used to dynamically monitor the potential of the anode or cathode (vs SCE).All experiments were performed at room temperature (26 ± 2 • C) under air atmosphere.

DCGC Synthesis of Pb-Containing Pd-Au Nanoparticles on MWCNTs/GCE
The GCE was polished with 1.0 and 0.05 µm alumina slurry sequentially, and then washed ultrasonically in water and ethanol for 5 min, respectively.Then, the GCE was subjected to potential cycling (−0.3~1.0V, 30 mV s −1 ) in 0.2 M aqueous HClO 4 for sufficient cycles to obtain reproducible CV curves.The treated GCE gave a peak-to-peak separation of ca.70 mV in the CV experiment of 2 mM K 4 Fe(CN) 6 + 0.1 M Na 2 SO 4 solution (−0.1~0.5 V, 50 mV s −1 ), indicating a well-cleaned electrode surface. 1 mg mL −1 MWCNTs dispersed in DMF were prepared and sonicated for 15 min, then 5 µL of the MWCNTs dispersion was cast onto the GCE (MWCNTs/GCE) and air-dried.
For the DCGC synthesis of Pb-containing Pd-Au nanoparticles on MWCNTs/GCE based on the GRR principle, a Cu plate served as the anode, MWCNTs/GCE served as the cathode, the anolyte was 0.1 M HClO 4 aqueous solution, and the catholyte was 3.0 mM PdCl 2 + 1.0 mM HAuCl 4 +5.0 mM Pb(ClO 4 ) 2 + 0.1 M HClO 4 aqueous solution.A salt bridge (a home-made U-tube filled with saturated KCl solution) was used to connect the anolyte and catholyte.Scheme 1 and Scheme 2 show the schematic diagram of the total experimental setup and the co-electroless deposition of Pb-containing Pd-Au nanoparticles onto MWCNTs/GCE, respectively.Figure S7 shows the physical picture display of the formaldehyde sensor with the three-electrode system.After switching on the DCGC, an ECN and pH meter were used to monitor the i cell and the potential of MWCNTs/GCE (vs.SCE), respectively.The total discharging time in the DCGC was 240 s.The as-prepared modified electrode was denoted as Pb-Pd 3 Au 1 /MWCNTs/GCE.Varying the molar concentration ratio of PdCl 2 to HAuCl 4 in the catholyte can fabricate modified electrodes loaded with different atomic ratios of Pd to Au.For comparison, we also fabricated other modified electrodes in a similar procedure just by changing the catholyte into 3.0 mM PdCl 2 +5.0 mM Pb(ClO 4 ) 2 + 0.

Electrochemical Characterization of the Modified Electrodes and Nonenzymatic Formaldehyde Sensor
The modified electrodes were characterized by CV in 0.1 M KOH (50 mV•s −1 , −1.0-0.5 V) aqueous solution.The nonenzymatic formaldehyde detection was conducted potentiostatically at a fixed potential in 0.1 M KOH aqueous solution.All of the solutions were deoxygenated by bubbling high-purity N2 for at least 10 min prior to each measurement.The modified electrodes were kept at 4 °C in a refrigerator when not use.

Electrochemical Characterization of the Modified Electrodes and Nonenzymatic Formaldehyde Sensor
The modified electrodes were characterized by CV in 0.1 M KOH (50 mV•s −1 , −1.0-0.5 V) aqueous solution.The nonenzymatic formaldehyde detection was conducted potentiostatically at a fixed potential in 0.1 M KOH aqueous solution.All of the solutions were deoxygenated by bubbling high-purity N 2 for at least 10 min prior to each measurement.The modified electrodes were kept at 4 • C in a refrigerator when not use.

Conclusions
In summary, nanoporous, flower-like, Pb-containing Pd-Au nanoparticles were deposited onto the MWCNTs/GCE cathode in DCGC with a Cu-plate anode.The presence of Pb 2+ in the catholyte should be responsible for the morphology evolution of the Pb-containing Pd-Au nanoparticles via a UPD effect.The fabricated Pb-containing Pd-Au nanoparticles exhibit high catalytic activity toward formaldehyde electrooxidation, mainly via a direct oxidation pathway, and an optimized modified electrode (Pb-Pd 3 Au 1 /MWCNTs/GCE) was employed to fabricated a formaldehyde sensor with high sensitivity, anti-interference ability, and stability.The procedure for nanomaterial fabrication presented here should have great potential in synthesizing high-performance nanocatalysts for electrocatalysis and electroanalysis applications.

Supplementary Materials:
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/molecules29122772/s1. Figure S1.SEM images of MWCNTs/GCE. Figure S2.CV curves at Pd pla /Au and bare Au electrodes in 0.1 M H 2 SO 4 aqueous solution.Figure S3.CV curves at Cu plate in 0.1 M HClO 4 aqueous solution.Figure S4.Effects of the applied potential on the amperometric response of 50 µM HCHO for Pb-Pd 3 Au 1 /MWCNTs/GCE in 0.1 M KOH aqueous solution.

17 Figure 1 .
Figure 1.Simultaneous responses of i cell , and E cathode in the DCGC with a Cu plate as the anode, the MWCNTs/GCE as the cathode, and 0.1 M HClO 4 as anolyte (stationary), while the catholytes (stationary) were 3.0 mM PdCl 2 + 1.0 mM HAuCl 4 +5.0 mM Pb(ClO 4 ) 2 + 0.1 M HClO 4 aqueous solution (a, for deposition of Pb-Pd 3 Au 1 ) and 3.0 mM PdCl 2 + 1.0 mM HAuCl 4 + 0.1 M HClO 4 aqueous solution (b, for deposition of Pd 3 Au 1 ), respectively.Short-circuit discharging of the DCGC (connecting the anode and cathode by a conducting Cu wire) was from 30 s to 270 s.

Figure 7 .
Figure 7. Amperometric response to successive addition of HCHO into the stirred 0.1 M KOH aqueous solution (A) and the calibration curve (B) at Pb-Pd3Au1/MWCNTs/GCE. Applied potential: −0.27 V.

Figure 7 .
Figure 7. Amperometric response to successive addition of HCHO into the stirred 0.1 M KOH aqueous solution (A) and the calibration curve (B) at Pb-Pd 3 Au 1 /MWCNTs/GCE.Applied potential: −0.27 V.

Scheme 2 .
Scheme 2. Schematic illustration of the co-electroless deposition of Pb-containing Pd-Au nanoparticles onto MWCNTs/GCE, and the electrochemical sensing of the modified electrode toward formaldehyde.

Table 1 .
Comparison of different modified electrodes for formaldehyde electrochemical sensing.

Table 1 .
Comparison of different modified electrodes for formaldehyde electrochemical sensing.

Table 2 .
Analysis results of formaldehyde in real samples.