XPS in development of chemical sensors

Abstract

XPS represents a powerful tool for investigation of chemistry involved in chemical sensors, as analytes and recognition elements interact at a device surface, the region analyzed by the spectroscopic technique. In this way, the processes involved in preparation, operation and limiting life of sensors can be unraveled. Application of XPS to (bio)chemical sensors is reviewed with particular attention to electrochemical sensors, which account for more extensive literature. Systems based on polymers, organic mono/multilayers, carbon nanotubes and graphene, metallic species, also in the nanostructure format, and bioreceptors and relevant composites have been considered with a special focus on the information provided by XPS analysis on such (bio)sensing systems, highlighting the role played by XPS in sensor technology and giving at the same time an idea of what is still missing in this field thus advancing possible future uses of XPS in sensing technology.

---

Elisabetta Mazzotta

Elisabetta Mazzotta is a permanent researcher at the University of Salento since 2011. In 2005 she received her MSc degree in Environmental Sciences at the University of Salento and in 2009 a PhD in Chemistry of Innovative Materials at the University of Bari. From 2009 to 2011 she was a post-doctoral research fellow at the Department of Materials Science of the University of Salento. The scientific activity of E. M. is devoted to the synthesis of novel materials and to their use in the design of biosensors and biomimetic sensors. In particular, her research activity is focused on micro- and nanostructured materials (nanoparticles, conducting polymers and molecularly imprinted polymers) and on the use of XPS methodology for their chemical characterization.

Simona Rella

Simona Rella graduated in materials engineering at the University of Salento in 2009 and obtained her PhD in material and structure engineering in 2013. She worked at National Institute of Nuclear Physics of Lecce for two years carrying out the study and characterization of the mechanical structure of a tracker detector. She works at the Laboratory of Analytical Chemistry of Department of Biological and Environmental Sciences and technologies of Lecce since 2013. Her current research interests are focused on the characterization of materials by XPS.

Antonio Turco

Antonio Turco is a post-doc researcher at the University of Salento. He received his MSc degree (2009) in Biotechnology Science from University of Salento and his PhD (2014) in chemical and pharmaceutical sciences and technologies from Center of excellence for nanostructured materials (CENMAT) of University of Trieste. His research focuses on innovative nanocomposite materials for tissue engineering and environmental remediation, electrochemical sensor and biosensor and surface functionalization with nanostructured materials, polymers and molecularly imprinted polymers for application as sensitive layers in sensors.

Cosimino Malitesta

Cosimino Malitesta received his MSc degree cum laude in Chemistry in 1983 and his PhD in Analytical Chemistry in 1989 from Università di Bari. Since then he has there covered lecturer and associate professor positions in Analytical Chemistry. In 2001 he has been appointed as full professor of Analytical Chemistry at Università del Salento. He is member of Italian Chemistry Society and served as Vice-President of Analytical Chemistry Division and in the Board of Biotechnology Group. His main research interests include XPS development and application, biosensors and biomimetic sensors based on electrosynthesized polymers, electrosynthesis of molecularly imprinted polymers, nanostructures.

---

1. Introduction

Chemical sensors are analytical devices of great interest both in fundamental and applied research, because they represent a valid alternative to complex analytical systems (like those based on chromatographic and mass spectrometric techniques) in term of costs, analysis time, field measurements and expertise of analysts. They consist of an active layer (a recognition element), capable of chemically interacting with target analyte producing a chemical signal (i.e. establishing a chemical bond and/or production of chemical products), which is transformed into an electric signal by an appropriate transducer to which the layer is coupled (Fig. 1). When the recognition element is represented by a biological component (like enzymes, antibodies, etc.) the device is called a biosensor.

Fig. 1 Simplified schematic representation of a generic chemical sensor.

High selectivity, a fundamental characteristic for sensors, relies mostly on recognition capabilities of active layer. Also, degradation of performances are often related to (chemical) modifications of it. Furthermore, it should be considered that sensors work generally as interface devices, because surface of active layer is the locus of chemical interaction between analyte and recognition element.

In this respect, chemical characterization of active layer represents an important task in development of sensors, as comprehension of working mechanism and processes controlling lifetime are of outmost importance in optimization of the device.

Several tools, like spectroscopic, electrochemical, etc. techniques are available for this purpose. Among them, a peculiar role is covered by X-ray photoelectron spectroscopy (XPS), which is able to discover chemical information on surface and near-surface regions, generally without any or little preparation and degradation of samples. Considering that several sensors are now based on thin films and/or nanostructures, XPS can be regarded as a bulk technique in these cases.

XPS is based on photoelectric effect stimulated by X-ray photons (Fig. 2): kinetic energy (KE) spectrum of emitted electrons is recorded. KE is related to binding energy (BE) of levels from which electrons are emitted by Einstein equation (eqn (1)):

BE = hν − KE (1)

Fig. 2 Schematic representation of the photoelectric effect.

Chemical information are generally gathered from peaks originating from core levels, which are subject to a chemical shift depending on chemical state (e.g. oxidation state) of emitting element. All orbital levels except the s levels (l = 0) give rise to a doublet with the two possible states having different binding energies. This is known as spin–orbit splitting (or j–j coupling).¹ For the discover and application of photoelectric effect, Kai Sieghban was awarded by Nobel prize in 1981. He named the technique Electron Spectroscopy for Chemical Analysis (ESCA), a name which is now much less used.

Chemical speciation can be even more accurate if one considers also Auger electrons. Auger electrons are originated from a relaxation process following the creation of a hole in a core level by electron photoemission. Their kinetic energy (Fig. 3) is independent of photon energy source. Auger electrons are represented by additional peaks in the photoelectron spectrum. Considering notation in Fig. 3, the modified Auger parameter (α′) is defined as (eqn (2)):

α′(eV) = E_k(KL₁L₂₃) + E_B(K) (2)

where E_k(KL₁L₂₃) is the kinetic energy of the Auger transition KL₁L₂₃ and E_B(K) is the binding energy of an electron in the photoemitting level K. Since α′(eV) is an empirically determined parameter with a unique value for each chemical state, it can be used as a “fingerprint” for characterizing chemical species.

Fig. 3 Auger transition is indicated using the letters of the involved levels (KL₁L₂₃).

Quantitative analysis is performed using relationship between analytical signal (photoelectronic peak area) and molar ratio of emitting species. The technique can be surprisingly sensitive as elements present in inhomogeneous samples at an average concentration in the order of parts per billion can be detected by XPS if they are concentrated at the surface.² Information depth when analyzing solids is naturally limited to 5–10 nm (i.e. surface and near-surface regions), due to attenuation length characterizing electron motion in solids and relevant scattering events. Non-destructive depth profile within sampled volume can be obtained by angle resolved XPS (ARXPS) (Fig. 4).

Fig. 4 Angle-resolved XPS.

Several problems arise in analyzing insulating samples, as emitted electrons leave them charged. If not compensated, charge builds up and time-dependent meaningless spectra are produced. Even when charging state is stabilized, it must be considered that eqn (1) contains an unknown additional charging term. This aspect is generally managed by evaluation of that term by an internal reference (e.g. C 1s signal originating from adventitious carbon contaminant regularly present on the surface of samples manipulated in the air). In addition, if non-uniform distribution of charge occurs, multiple or wider signals could be produced for the same chemical species lying in differently charged regions, which could simulate the occurrence of several chemical species of a certain element. Most recent spectrometers are equipped with charge compensation tools of great help with this drawback.³ For the binding energy calibration and correction in the reviewed papers, readers should refer to original paper details. For extensive information on basics of XPS readers are referred to fundamental books on the topic (see e.g. ref. 1).

Following development of nanotechnology, application of XPS has been extended to this new exciting field, giving important pieces of information in terms of chemistry and electronic properties of nanostructures (see e.g. ref. 4). While XPS application to characterization of materials has been reviewed,⁵ the particular perspective and the peculiarities of its role in development of sensors have not there been the focus. So, the aim of this work is cover this aspect, with particular attention to electrochemical sensors to which XPS has given a more important contribute. The matter has been organized in sections, each describing a specific key role of XPS analysis both in development and application of sensors.

2. The use of XPS in the assembly/fabrication of the sensing material

Since the early 1990's XPS was successfully used to cast light on chemical structure of materials possessing capabilities of selective recognition used as recognition element in the development of electrochemical sensors. Surface analysis technique played a key role in the characterization of surface chemically modified electrode due to the fact that molecular specificity in electrochemical measurements can be dramatically influenced by modifications of electrode material surface composition.

In the field of electrochemical sensors, the use of chemically modified electrodes has significantly grown up over the past three decades, promoted by researcher attempts to exert more direct control over the chemical nature of an electrode. Strategies used for the fabrication of such devices include, along with polymer deposition by chemical or electrochemical polymerization, the immobilization by physisorption or chemisorption of self-assembled monolayer with terminal host or guest groups and of multilayers by layer-by-layer deposition, the covalent bonding of a specific functional group, the anchoring of specific functionalities by electrochemical processes, the deposition of metal species and/or carbon species (carbon nanotubes and graphene) alone or integrated in composite hybrid materials. In all cases, the final aim is to deliberately modify the surface of the electrode with monomeric or polymeric reagents able to determine its electrochemical behavior for tailoring its response to specific analytical needs with improved sensitivity, selectivity and/or stability.

It can be thus easily recognized the preeminent role of the surface analysis in this context, allowing to deeply investigating the chemical structure of functional groups tethered to the electrode surface. Different roles of XPS analysis in the assembly of modified electrodes can be identified, corresponding to different degrees of details in the provided information. The more basic use of XPS analysis consists only in the verification of the successful immobilization of the sensing material by simply tracing a marker XPS; signal without deeply investigating the nature of the bonding with the electrode surface and/or the possible modifications occurring during the stepwise electrode fabrication. In several other examples, surface analysis is effectively used to check the presence and the integrity of the desired functional groups ad hoc introduced on the electrode surface, which are responsible for, or at least involved in, the sensing mechanism. Another application field of XPS analysis in the assembly of modified electrodes consists in assisting the stepwise fabrication process with the aim to verify the successful implementation of each single step and to provide a feedback for its further optimization. Selected literature examples reporting the two latter XPS applications will be discussed below.

2a. XPS analysis of specific functional groups involved in sensing mechanism

The relevance of XPS in the analysis of chemically modified electrodes mainly focused on the investigation of functional groups involved in the sensing process can be inferred from the studies about its application to conducting polymer based electrochemical sensors. Since its first use in studying conducting polymers,⁵ XPS analysis dealt with doping process with the aim to both evaluate doping ions entrapment – to establish if they are inert and incorporated as such within the bulk of the polymer – and to quantify the doping level. The understanding of these points played a key role in the fundamental studies of conducting polymers, but still today represents a useful tool in their sensing applications being in some cases the dopant ion responsible for the interaction with the target analyte and the doping degree critically influencing films conductivity and thus their sensing properties, especially in the case of electrochemical transduction. For instance, the evidence of hydroquinone monosulfonate (HQS) incorporation within polypyrrole (PPy) backbone was provided by XPS analysis in the development of a potentiometric pH sensor using HQS as PPy doping ion⁶ because of its molecular charge-transfer properties. XPS showed the presence of O 1s component at 531.7 eV ascribed to SO₃⁻ and of two species identified in S 2p_3/2 spectrum associated with the R–SO₃⁻ group (168.2 eV) and with the oxidized form of SO₃⁻ (168.8 eV). The dopant incorporation was further confirmed by C 1s spectrum presenting components at 284.7 eV (erroneously indicated by the authors at 248.7 eV) attributed to the aromatic carbon species in the HQS anions and shake-up structure deriving from π → π* transitions in the aromatic structure. A doping level of 28% was calculated by the ratio ([N⁺] + [N²⁺])/[N_total], instead of [S_tot]/[N_tot], as a sulfur content greater than the positively charged nitrogen content was found, evidencing an excess of HQS. Nevertheless, this evaluation did not take into account the bivalent charge of the species N²⁺ as well as the fact that the charge of the oxidized form of SO₃⁻ is not properly evaluated.

The entrapment of a dye, reactive blue-4 (RB4), as a dopant within a film of poly1,5-diaminonaphthalene (poly-DAN) was verified by XPS analysis⁷ in a study devoted to the simultaneous determination of dopamine (DA) and acetaminophen (AP) in urine. The presence of Cl 2p and S 2p at, respectively, 200.6 eV and 163.4 eV, corresponding to the sulfur and chlorine moieties in the structure of RB4, was indicative of the doping by RB4. It was responsible for the electrocatalytic behavior of the system poly-DAN/RB4 towards DA and AP oxidation due to the electrostatic attraction between negatively charged sulfonic group of the dye and positively charged target molecules.

The XPS study of the chemical structure of sensing layer has been extended to several kinds of polymeric material with the aim to verify that functionalization processes for the introduction of specific functionalities within the sensing layers are effectively performed and/or that such functionalities are preserved after deposition step. For instance, XPS characterization was fruitfully performed on hollow microspheres of poly(N-methylpyrrole) (pNMPy) prepared by layer-by-layer deposition of pNMPy on polystyrene (PS) nanoparticles, subsequently removed by chemical etching, and used for the electrochemical detection of dopamine.⁸ High resolution spectra of C 1s, N 1s and Cl 2p for both uncoated PS and pNMPy/PS coated particles were analyzed. No signal associated to the aromatic ring of PS, i.e. π–π* shake-up satellite peaks in the region 291.5–293.0 eV, was observed in the spectra of the coated microspheres, indicating that the deposited 30 nm thickness pNMPy was enough to get a homogeneous coverage of the core–shell. The presence of N 1s and Cl 2p only in the spectra of pNMPy/PS further confirmed the successful coating of the particles. Moreover, there was no evidence of Fe atoms in the Fe 2p region at the surface of the material, indicating the complete purification of the material after the polymerization process.

The integrity of sensing layer was verified by XPS analysis in a study⁹ proposing a strategy for the reconstruction of a polyoxometalate (POMs) coordination polymer, namely [H₂bpy]₂[{Cu-(bpy)₂}Mo₅P₂O₂₃]4H₂O (bpy = 4,4′-bpyridine), on ITO electrode with the aim to exploit the resulting polymer electroactivity toward the reduction of iodide. XPS spectra of the deposited POMs on ITO electrode showed the same features as the polyoxometalate coordination polymer, implying that the original structure of the deposited POMs was kept. This emerged from the comparison of XPS spectra of the coordination polymer as is and after deposition process both showing peaks at 932.8 eV, corresponding to the binding energy of Cu 2p_3/2, and at 232.5 eV corresponding to Mo 3d_5/2.

An extensive XPS characterization was performed by Micić et al.¹⁰ aiming at understanding how diverse N- and O-containing surface functionalities of structurally and morphologically similar nanomaterials could influence their activity towards electroxidation of nitrite and ascorbic acid (AA). In particular, nanostructured films of polyaniline (PANI) were prepared by the oxidative polymerization in the presence of salicylic acids and then carbonized. The authors observed that the increase of surface N- and O-containing groups determined higher PANI films electrocatalytic activity towards nitrite and AA oxidation. Moreover, considering the effect of particular functional groups, they found that oxidation peak potential of nitrite and AA shifted to lower values with the increase of surface content of quinone type (C=O) and N-oxide species. According to authors these functional groups could be regarded as surface defect sites serving as active centers thus being beneficial for electrocatalytic behavior of PANIs, in a synergic mechanism exploiting also the mesoporosity of the investigated materials.

The use of XPS analysis as a way for investigating specific functional groups in the structure of sensing material has been largely explored also in the field of carbon based materials modified electrodes. It is well known that carbon nanotubes, both single and multiwalled, have shown great results in many applications such as electronics,¹¹ energy storage and conversion,¹² batteries,¹³ fuel cell and biotechnologies^14,15 because of their unique physiochemical properties such as high surface area, excellent thermal and electronic conductivity and strong mechanical strength. These characteristics allow their large and successful use in electrochemical sensors. However to achieve these aims the carbonaceous materials need to be chemically processed in order to be purified and functionalized in order to allow for the binding of specific recognition elements on their surface and/or their controlled deposition on a specific substrate. It is thus necessary to deeply investigate the chemical structure of such carbon-based materials for properly orienting and optimizing their integration with electrochemical transducers. A few techniques are available for this purpose and XPS is a powerful tool in analyzing surface chemistry of such materials. For instance, Yuan et al.¹⁶ during the fabrication of a sensor for nitrite based on phosphotungstic acid-multiwalled carbon nanotubes (PW-MWCNTs) nanocomposite, used XPS analysis for verifying both covalent and non-covalent functionalization of MWCNTs before their deposition on the electrode surface. MWCNTs surface was preliminarily grafted with poly(4-vinylpiridine) (PV4P) by a free radical polymerization to form a positive matrix able to electrostatically attract anionic PW. They observed the appearance of a peak at ca. 400 eV due to N 1s signal and a new shoulder peak at 286.2 eV which was assigned to C–N bind of pyridine group in PV4P chains. After that, MWCNTs–PV4P composite was grafted with PW, already used in the electrochemical detection of nitrite¹⁷ due to its ability to act as an electron reservoir without compromising its structural integrity.¹⁸ They observed the presence of a W 4f peak after this last functionalization step and, as a further proof, the decrease of signals relative to pirrolidinic groups, indicating that these signals were screened off by PW (Fig. 5). Chang et al.¹⁹ functionalized MWCNTs with platinum nanoparticles by femtosecond laser for non-enzymatic glucose electrochemical detection. Pt 4f was detected by XPS after deposition confirming the presence of platinum on MWCNTs surface; by curve fitting the presence of three doublets at 71.4 eV (Pt 0), 72.6 eV (Pt 2+), 75 eV (Pt 4+) was found. In particular, XPS spectra revealed that Pt 0 was the major species (>61%) confirming the effective reduction of platinum in solution and its deposition on MWCNTs.

Fig. 5 (A) XPS survey spectra recorded for (a) MWCNTs, (b) PV4P grafted MWCNTs and (c) PWs/PV4P grafted MWCNTs hybrids. (B) W 4f spectrum on PW/PV4P–MWCNTs hybrids. (C) C 1s, (D) N 1s and (E) O 1s spectra on (a) MWCNTs, (b) PV4P grafted MWCNTs and (c) PW/PV4P–MWCNT hybrids. Reproduced, with permission, from ref. 16.

More recently, surface characterization is being utilized for graphene modified electrodes since it is recognized as a material possessing interesting features for electrochemical sensing applications, although such a use is made complex by the difficulties in its production and by the low amount of graphene typically obtained.²⁰ A common strategy for overcoming such limitations relies in the production of graphene oxide (GO), which allows obtaining a cost-effective and large-scale production of graphene-based materials. However graphene oxide is hindered by lower conductivity with respect to graphene due to the presence of oxygen on the surface causing an impairment of electronic structure. It is thus necessary to preliminarily reduce the oxidized material to make it prone to the integration with electrochemical sensors fully exploiting its electronic characteristics.²⁰ In this context, XPS characterization can be very useful giving the possibility to investigate both oxygen species and sp² hybridized carbon atomic percentage, besides the presence of specific functional groups. Wang and co-workers²¹ proposed the use of XPS analysis for monitoring a one step procedure for the reduction of GO and integration of 4-aminophenylboronic (APBA) acid as recognition element for electrochemical detection of different sugars. After the reaction of GO with APBA they observed an increase of C 1s peak at 283.9 eV relative to sp² hybridized carbon with respect to GO alone suggesting the effective reduction of GO. In addition they observed the presence of N 1s confirming the integration of APBA on graphene but, contrarily to what observed on APBA alone, two components were found, one at 399.7 eV and one at 398.2 eV, due to amine and imine groups, respectively, resulting from APBA oxidation and from the formation of oligomeric side products. Yang and co-workers²² were able to unequivocally demonstrate the effective functionalization of GO with Prussian blue for H₂O₂ detection by analysing the XPS high resolution spectra of Fe 2p_3/2. In particular they found the presence of two components, one at 711.3 eV that corresponded to Fe³⁺ and one at 708.2 eV assigned to Fe(CN)₆⁴⁻. On the basis of XPS results, they affirmed that Fe³⁺ cations combined with Fe(CN)₆⁴⁻ anions to form Prussian blue on GO. Wang et al.²³ focused on the development of Au–reduced graphene oxide (Au–RGO) nanocomposite obtained by UV irradiation under mild conditions allowing the simultaneous reduction of graphene oxide and formation of Au nanoparticles by using HAuCl₄ as precursor. XPS characterization indicated that GO was deoxygenated and reduced after UV treatment to form the Au–RGO nanocomposite. In particular, C 1s peak in Au–RGO showed that the intensity of epoxy groups (286.7 eV), carboxyl (287.1 eV) and carbonyl groups (288.6 eV) were remarkably reduced respect to GO, accompanied by a significant increase of O 1s peak intensity. Consistently, the Au 4f_7/2 peak at 83.7 eV confirmed that Au³⁺ ions were effectively reduced on the surface of graphene oxide sheets. The Au–RGO nanocomposite exhibited much higher electrocatalytic activity toward 2,4,6-trinitrotoluene than RGO and showed great promise for application in electrochemical sensing.

Also in the assembly of sensors integrating a biological receptor, the use of XPS has been largely exploited being in some cases focused on the biological component aiming at verifying its effective integration with the electrode surface, its chemical integrity and/or quantitatively estimating its amount. One of the first applications of XPS analysis dedicated to the characterization of biosensor component interface was proposed by De Benedetto et al.²⁴ who studied glucose oxidase (GOx) adsorbed on a platinum surface. The results were used to identify the C 1s components and to hypothesize that the enzyme was oriented with the protein portion in contact with the electrode and with the exterior part consisting of carbohydrate chain. Moreover, N/Pt ratio values after different adsorption times were used to estimate the enzyme coverage and to postulate two possible enzyme arrangements on the electrode surface, consisting in partly horizontally and partly vertically oriented enzyme molecules and/or enzyme molecules arranged as multilayer islands over the electrode surface. Starting from the 1990's, while the polymeric systems embedding enzymes were attracting a strong interest in the biosensors panorama,²⁵ in parallel the surface analysis was increasingly recognized as a powerful tool for the study of such biosensor interfaces. For example, Wang and Mu²⁶ performed XPS characterization of a polyaniline (PANI) film before and after the entrapment of cholesterol oxidase, to be used for cholesterol electrochemical sensing. Upon the enzyme entrapment, C 1s signal exhibited the most pronounced modification, presenting a single component at 284.6 eV in PANI and three peaks of 288.0, 286.1 and 284.5 eV in PANI/enzyme. According to the authors, the main peak at 284.5 eV was due to C1s in PANI, while the other two peaks were related to the immobilized enzyme or to a chemical shift of C 1s in PANI itself. The latter point was explained by invoking the effect of the negative charges of cholesterol oxidase decreasing the shielding of PANI carbon atoms by outer electrons. Such an explanation did not sound very convincing being C 1s peaks at 288.0 and 286.1 eV more easily ascribed to C=O, C–N and C–O groups of the enzyme entrapped into polymer. To this aim the XPS analysis of cholesterol oxidase could have been particularly useful. Moreover, the changes of the intensity of N 1s peak components, after enzyme entrapment, as well as the slight shift of O 1s components, were not explained by the authors.

In our research group an intensive research activity devoted to enzyme-polymer biosensing systems is carried out^27–31 and in some cases the possibility of gaining important information about the developed system by XPS analysis has been demonstrated.^29,31 In the assembly of an amperometric glucose biosensor based on GOx embedded in a poly(vinyl alcohol) (PVA) matrix drop casted on a platinum electrode (Pt/GOx–PVA),²⁹ XPS was used to demonstrate the homogeneous film deposition as well as its structure preservation under operative conditions. XPS spectra of Pt/GOx–PVA were analyzed after film preparation and after electrochemical experiments: in all cases, detailed spectra in the N 1s region showed the presence of component peaks characteristic of GOx standard. Moreover, the absence of signals relative to Pt 4f on all analyzed samples even after glucose detection suggested that the integrity of the film was kept after electrochemical measurements. This information was particularly crucial for supporting the hypothesized sensing mechanism that was not directly related to the role of Pt substrate only but was due to a synergistic effect between Pt and GOx–PVA film.

XPS has been successfully exploited in the investigation of biosensing systems based on covalent immobilization of the biological receptor on a properly modified electrode surface.^32,33 The covalent immobilization of GOx onto electrically conductive ultra-nanocrystalline diamond (UNCD) surface was performed by exploiting tethered aminophenyl functional groups previously grafted to UNCD surface by electrochemical reduction of aryl diazonium salt and a subsequent EDC/NHS (1-ethyl-3-[3-dimethylaminopropyl]carbodiimide hydrochloride/N-hydroxysulfosuccinimide) coupling reaction.³² XPS analysis was used to confirm the surface functionalization and to follow the changes in surface chemistry in the course of immobilization procedure. C 1s, O 1s and N 1s signals were monitored in the initial clean UNCD surface, the succinic acid modified surface, and the GOx modified surface. After surface grafting of aminophenyl groups and their subsequent reaction with NHS, N 1s signals at 400 eV and at 406 eV attributed to amide moieties and to nitrogen in the oxidized form, respectively, were observed, accompanied by an increase of O/C ratio. Following UNCD incubation with GOx, enhanced N 1s signal was observed with a further increase of O/C ratio. The authors concluded that these observations, along with the C 1s peak broadening after the stepwise surface modification proved the successful attachment of the protein molecules onto UNCD surface.

2b. Monitoring and optimizing the stepwise fabrication of the sensing material by XPS analysis

A valid example of the application of XPS analysis in monitoring each single step of electrode modification process can be found in the work of Coates and Nyokong³⁴ who prepared an electrochemical sensor for hydrazine based on an iron phthalocyanine (FePc) modified GC electrode, assembled by a three step approach including grafting of the bare GC by electrochemical reduction of 4-azidodiazonium, click chemistry³⁵ with 4-ethynylpyridine and axial ligation of FePc. N 1s XPS signal was analyzed after each step exhibiting the characteristic peak of the azide group at 404 eV on grafted electrode, located at higher binding energy due to its low electron density. Subsequent click chemistry process was also followed by XPS showing a significant reduction of this peak, with the appearance of a broad nitrogen signal at 400.0 eV due to the contribution of both the pyridinic nitrogen of the clicked pyridine as well as the triazole ring, indicating that the majority of the azide groups reacted to form the triazole. Upon ligation to FePc, the nitrogen signal showed the presence of a peak at 399.7 eV attributed to the triazole, a component at 398.8 eV due to the contribution of the pyrrole/aza FePc and the pyridinic nitrogens, as well as the broad peak at 401.1 eV.

A similar XPS application was proposed in a study reporting the anchoring of EDTA to GC electrode with the goal to exploit chelating properties of EDTA in Pb²⁺ electrochemical detection.³⁶ To this aim p-aminophenyl (AP) groups were firstly obtained on the electrode surface by electroreduction of nitrophenyl films and an EDC (1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride)-activated reaction was then carried out. The significant increase of nitrogen content in XPS spectra on EDTA–GC compared to AP–GC indicated the appreciable attachment of EDTA moieties to the AP–GC surface. Interestingly, XPS spectra of EDTA–GC surface evidenced the presence of the –NO₂ peak at 406.4 eV, originating from the incomplete reduction of nitrophenyl films, thus evidencing that the electrochemical method was not effective to completely reduce the grafted nitrophenyl groups to aminophenyl moieties, consequently decreasing the amount of anchored EDTA.

An extensive XPS characterization was performed by Ciampi et al.^37,38 who performed XPS analysis of specific functionalities attached to a silicon surface aiming at making it more prone to application in electrochemical sensing. The authors demonstrated that robust oxide-free silicon surface could be produced by hydrosilylation of a H-terminated Si(000) surface using a diyne species and also that such surfaces were sufficiently stable to allow nearly ideal electrochemistry over many redox cycles with a ferrocene attached to the distal end of the monolayer. The first evidence was inferred from Si 2p detailed spectrum as no silicon oxide was detected in the 102–104 eV region of the Si 2p spectrum. The analysis of Si 2p region proved also the high-quality monolayer after cyclic voltammetry experiments in aqueous solutions as well as after long-term storage, containing in both cases no clear evidence of the formation of SiO_x species. XPS data highlighted also the absence of silicon oxide-related signals, although the sample was exposed to aqueous solutions and extensively handled in air during washing steps. Ferrocene assembly on such modified silicon substrates was achieved by alkyne–azide “click” coupling reactions with ferrocene substituted azide. From XPS spectra, the presence of Fe 2p signal from immobilized ferrocene units suggested the successful outcome of the cycloaddition reaction. In particular, curve fitting of the experimental curves for the Fe 2p region showed a Fe 2p_3/2 component at 708.6 ascribed to Fe(II) species without detectable signals at ∼711 eV, generally observed for Fe(III) compounds. This observation indicated that no appreciable oxidation of ferrocene to ferricinium occurred under the coupling conditions. Moreover, the conversion of the acetylene layer to the corresponding ferrocene-based surface after click process was estimated to be approximately 40%.

The use of XPS for monitoring and optimizing the electrode assembly process has been explored also in the applications of carbon nanotubes and graphene in electrochemical sensing. Chen et al.³⁹ proposed the use of XPS for the analysis of oxidized MWCNTs functionalized with cysteamine. Such functionalization aimed at allowing the vertical alignment of nanomaterial on a gold surface, thus significantly increasing the surface/area ratio and consequently enhancing sensor sensitivity. Characterizing the surface of the as prepared electrode by means of XPS, the authors verified the presence of three components in C 1s high resolution region at 285.5, 286.7, and 288.6 eV, ascribed, respectively, to C–C, C–N, and amide bonding, confirming the covalent bonding between oxidized CNTs and cysteamine deposited on gold electrode; in addition the peak at 163.7 eV in the S 2p_3/2 spectrum was attributed to the Au–S bond formation.

Also in a study of Yola and Atar,⁴⁰ the application of XPS analysis to carbon nanotube modified electrode was reported. They fabricated covalently functionalized CNTs with p-aminothiophenol (pATP) in order to insert an anchor point on CNTs surface to bind Au nanoparticles through thiol–gold interactions. In details they firstly functionalized CNTs with carboxylic moieties, on which p-aminothiophenol was bound by amidation reaction. The nanocomposites were successfully used for the electrochemical simultaneous determination of quercetin and rutin. By curve fitting of N 1s narrow region, they observed the presence of a peak at 397.7 eV attributable to the amides group, and a peak at 400.9 eV suggesting the presence of NH groups in unreacted pATP. Sulfur peak at 163.4 eV suggested the binding between sulfur and gold surface of nanoparticles. According to authors the Au 4f_7/2 peaks signal at 84.1 eV confirmed the presence of bonded Au. Nevertheless, the analysis of Au 4f detailed spectrum was not correctly performed as the Au 4f_5/2 component at 88.4 eV was erroneously attributed to a different species, namely to unreacted Au nanoparticles. Moreover, the peak area ratio (4f_5/2 : 4f_7/2) was evidently higher than the theoretically expected value (3 : 4)².

Han and co-workers⁴¹ developed a Pt/TiO₂ nanohybrids modified single-walled carbon nanotubes electrodes for the detection of hydrogen peroxide being assisted by information provided by XPS analysis. TiO₂ nanoparticles (NPs) were firstly electrochemically deposited, followed by subsequent Pt NPs photoinduced deposition. In order to investigate the presence of Pt NPs as well as their oxidation state, Pt NPs at different pH were prepared and characterized by XPS analysis (Fig. 6A–B). Three different platinum oxidation states were detected, identified as Pt(0), PtO and PtO₂ with Pt 4f_7/2 peaks located at 70.8, 72.2 and 74.5 eV, respectively. In particular, it was observed that at pH 1, metallic platinum was the major component, while the increase of pH determined an increase of Pt oxide species with the PtO₂ component becoming predominant at pH 10; at the same time metallic platinum disappeared. Furthermore, the amount of deposited Pt NPs was found to be dependent on pH, with a maximum increase at pH 10. Also the effect of Cl⁻ ions added to the Pt NP precursor solution was evaluated by XPS analysis (Fig. 6C) evidencing that at pH 10 the addition of chlorine influenced only Pt oxidation state without affecting the amount of the deposited materials, while at pH 1 Pt 4f signal nearly disappeared with the increase of Cl⁻ ions confirming the difficult production of platinum nanoparticles at low pH. All these results allowed the optimization of sensor response as an increase in the amount of PtO is preferred in catalytic reactions with H₂O₂.

Fig. 6 XPS spectra for the 4f state of platinum nanoparticles photoinduced in 1 mM H₂PtCl₆ precursor solution with (A) pH 1 and (B) pH 10. The peaks of 4f_5/2 state of PtO (at 74.2 eV) and 4f_7/2 state of PtO₂ (at 74.5 eV) overlapped because of the very similar binding energy. (C) XPS spectra for the 4f states of platinum nanoparticles prepared at pH 10 in the presence of 100 mM KCl. Reproduced, with permission, from ref. 41.

Helpful information were provided by XPS by Liu and collaborators,⁴² who developed a sensor for trace detection of Cu(II) based on a layer by layer assembly of GO and NH₂-rich polyelectrolyte (polyallylamine hydrochloride, PAH) on glassy carbon electrode (GC) (Fig. 7). They observed that after the self assembly followed by reduction of prepared film with NaBH₄, the C 1s high resolution spectrum of graphene exhibited a decrease in C–O and C=O components and a corresponding increase of the peak relative to sp² hybridized carbon (284.6 eV) indicating the effective removal of residual oxygen containing groups in GO (Fig. 7A). At the same time the appearance of N 1s peak (Fig. 7B), before and after chemical reduction, confirmed the incorporation of NH₂ enriched electrolyte on prepared film. Additionally, on reduced film (PAH–GS) a new component at 285.7 eV corresponding to carbon bound to nitrogen (C–N) appeared suggesting a covalent interaction between polyelectrolyte and graphene that could favor the stability of the sensor.

Fig. 7 Schematic illustration of ordered nanocomposite electrodes of [PAH–GO]_n/GC by layer by layer self-assembly (left). XPS spectra of (a) GO, (b) [PAH–GO]₁₂ and (c) [PAH–GS]₁₂ in (A) C 1s and (B) N 1s regions (right). Reproduced, with permission, from ref. 42.

Interesting examples of the key role of XPS analysis in monitoring the stepwise electrode modification can be found also in the field of biosensors. Gooding and his research group,⁴³ for instance, extensively used XPS in the fabrication of a label-free immuno-biosensor for biotin comprising a mixed layer of oligo(ethylene glycol) (OEG), and an oligo(phenylethynylene) molecular wire (MW) bearing a ferrocene-based redox probe bound to the biotin epitope (the structural feature the antibody selectively recognizes). Sensing principle was based on the modulation of the ferrocene electrochemistry due to association or dissociation of the antibody with the sensing interface. Firstly, C 1s and N 1s XPS data indicated that the mixed layer of MW/OEG was successfully grafted on the glassy carbon electrode. The second step modification consisting in the covalent attachment of ferrocene-based redox probe was clearly proved by the presence of Fe 2p_3/2 signal at 708.6 eV suggesting a major Fe(II) population. Finally, the attachment of biotin was confirmed by the S 2p peak at 164.0 eV. The authors highlighted that N 1s spectrum, for each monitored step, was characterized by the presence of a component at about 400.4 eV evidencing a high degree of amide bonds, clearly indicating the successful biosensing interface fabrication. Another example of the useful information provided by XPS analysis in the fabrication of biosensing systems can be found in Zhu et al.⁴⁴ who prepared a nanocomposite material consisting of gold nanoparticles (AuNPs) whose surface was covered by self-assembled 2,5-di-(2-thienyl)-1H-pyrrole-1-(p-benzoic acid) (DPB). DPB was electropolymerized and nanoparticles were deposited on a disposable screen-printed electrode (SPE). A DNA aptamer selected for kanamycin was covalently immobilized onto the poly-DPB(AuNP) nanocomposite to make an electrochemical sensor to detect the kanamycin in phosphate buffer and milk samples. The chemical surface composition of DPB(AuNP)/SPE, poly-DPB(AuNP)/SPE, and aptamer/poly-DPB(AuNP)/SPE was assessed by XPS enlightening the successive steps of electrodes preparation. C 1s at 284.6 eV for all samples was attributed to C–C/C–H bonds. Au 4f_7/2 observed at 83.6 eV for the self-assembled DPB(AuNP) disappeared after electropolymerization indicating that AuNPs were covered with poly-DPB film. S 2p and N 1s were observed for all the surfaces at 163.7 eV (C–S) and 400.0 eV (C–N), respectively, confirming that DPB structure was preserved upon electropolymerization. Finally, the successful immobilization of aptamer on the poly-DPB(AuNP) nanocomposite was demonstrated from the increase in the intensity of the N 1s peak, due to the formation of O=C–NH–(C–H) bond and –NH₂ group present in aptamer, and from the appearance of P 2p peak, observed at 131.9 eV.

The significant contribution provided by surface analysis in the design of the sensing material by following fabrication procedure thus giving a feedback for its optimization is clearly demonstrated by the application of XPS to a specific class polymeric materials suitable for sensing applications, namely molecularly imprinted polymers (MIPs), able of acting as artificial sensing layers due to their properties of highly selective molecular recognition.⁴⁵ The use of XPS in the study of imprinted materials offers the possibility to examine not only their chemical structure and to identify functional groups involved in the target molecule recognition, but it represents also a valuable instrument for evaluating the efficacy of the imprinting process. XPS is in fact successfully used for monitoring the template entrapment within the polymer film and its removal after washing step, thus allowing to compare different protocols and selecting the optimal conditions providing the highest template removal. Such an application of XPS analysis is documented in several papers from our research group.^45–47 For instance, XPS investigation was performed on a MIP electropolymerized from a cobalt porphyrin as monomer with the aim to exploit advantages related to metal ion coordination in imprinting procedure.⁴⁸ It was thus particularly crucial to evaluate the integrity of metal complex upon electropolymerization as well as after washing procedures, the latter possibly influencing polymer coordination properties due to the ability of washing solvent to complex metal centers. The integrity of porphyrin structure upon electropolymerization was assessed by the similarity of N/Co ratio in Co porphyrin, in non-imprinted polymer and in pristine MIP, equal to 8.1, 8.5 and 8.6 respectively, in agreement with stoichiometric ratio. For following chemical changes upon MIP washing, the signal of Cl 2p was monitored being a marker for the presence of the target molecule (i.e. the toxic organohalide 4-(2,4-dichlorophenoxy)butyric acid (2,4-DB)). After washing procedures in acetonitrile (ACN), a significant increase of N/Co ratio was observed suggesting an unwanted Co removal mainly due to the ability of ACN to complex Co, thus prompting the use of a different solvent. MIP washing in methanol revealed the ability to remove almost the total amount of chlorinated compound without altering porphyrin structure, as revealed by the analysis of Cl 2p signals and by the evaluation of Cl/N and N/Co atomic ratios. Interestingly, some considerations about MIP–template interaction were drawn by the analysis of Cl 2p detailed spectra. It was observed that pristine MIP exhibited the presence of two distinct Cl species, attributed to the existence of two populations of binding sites in the imprinted cavities and thus to the occurrence of two different interaction mechanisms: one due to coordination with Co centers and another one based on hydrogen bonds with free amino groups of the electropolymerized porphyrin.

More recently we have reported an extensive XPS study of an overoxidized imprinted polypyrrole (PPy) propaedeutic to its application to the electrochemical detection of sulfadimethoxine (SDM).⁴⁶ In this work, overoxidation was used for SDM removal by exploiting electrostatic repulsion with oxygen-containing groups, such as carbonyl and carboxyl, introduced in PPy backbone upon such electrochemical treatment. Firstly, the role of the electrolyte (tetrabuthylammonium perchlorate (TBAP) (in MIP T) and lithium perchlorate (in MIP L)) in influencing the sensing performances of imprinted PPy was demonstrated and supported by XPS analysis. The lack of Li 1s and the presence of N 1s at about 401 eV in MIP L and in MIP T, respectively, suggested that, while Li⁺ was easily entrapped/ejected, TBA⁺ remained embedded in polymer structure, determining a less homogeneous and less compact film structure, more prone to degradation after overoxidation, with consequent declining of sensor performances. Also the estimation of doping degree of two films in terms of Cl 2p percentage content suggested higher doping level of film L (Cl 2p 1.8%) in respect to film T (Cl 2p 1.1%) thus further confirming the formation of a more densely packaged structure in the former case. Moreover, due to the key role of overoxidation treatment in the imprinting process, three different overoxidation conditions were tested and XPS analysis of resulting polymers was performed, evidencing different extent of polymer overoxidation by the comparison of C 1s and N 1s signals in pristine and overoxidized MIPs. Interestingly, the different overoxidation level was evidenced also by the modification of S 2p signal, due to sulfonamidic group of SDM and thus indicating its presence within the polymer. Atomic ratio between pyrrolic N 1s and S 2p was used to quantitatively estimate the template removal after each treatment, thus selecting the optimal overoxidation conditions determining a template removal of 33%.

3. The use of XPS for investigating the structure/architecture of the sensing material

An interesting use of XPS in electrochemical sensing technology concerns the possibility of gaining useful information about the structure of the sensing material with the aim to verify the hypothesized structure and to postulate its formation mechanism, thus achieving the possibility of tailoring its properties on the basis of the desired sensing applications. An example of such XPS application was proposed by Reisberg et al.⁴⁹ who assembled an electrochemical sensor for DNA hybridization by grafting a tripeptide (glutathione, Glu) onto 5-hydroxy-1,4-naphthoquinone (HNQ), an electropolymerizable molecule, with the aim to exploit the ability of tripepetide to covalently graft amino-modified DNA probe strands.⁴⁹ XPS analysis provided the basis for hypothesizing the structure of the resulting film. An experimental atomic ratio N/S of 2.95 was found, in great agreement with theoretical ratio (3 : 1) of the Glu structure as no sulfur and nitrogen belong to the HNQ molecule. As the experimental C/S ratio was equal to 99 (∼5 times larger than the theoretical expected value 20 : 1 for the HNQ–Glu moiety), the authors hypothesized the formation of a copolymer between the HNQ moieties and the HNQ–Glu moieties with a ratio of 4, i.e. 1 HNQ–Glu for 4 HNQ. Fig. 8 presents the structure of poly(HNQ-co-HNQ–Glu) film, proposed on the basis of XPS results.

Fig. 8 Proposed structure for poly(HNQ-co-HNQ–Glu). Reproduced, with permission, from ref. 49.

The structure of a composite film integrating Pt(0) and polypyrrole, used as DNA electrochemical sensor, was investigated by XPS analysis⁵⁰ which provided useful information being Pt(0) used as anchoring points for DNA attachment as well as an enhancer of the film conductivity. The polypyrrole/Pt(0) film was obtained by low-energy platinum ions implantation into electrosynthesized polypyrrole films doped with ClO₄⁻. In such a context, the possibility to perform depth profile analysis by XPS was particularly beneficial providing information on the distribution of Pt through the PPy film. Results of depth profile analysis (Fig. 9) evidenced that the Pt profile concentration began low at the PPy surface but increased to a maximum at about 500 s of etch cycles to ca. 18 at% of the sample. The amount of chlorine was found to remain stable throughout the depth investigated suggesting that ClO₄⁻ doping was fairly homogeneous through the depth of the film. Moreover, the Pt 4f detailed spectra revealed that implanted samples included a mixture of Pt metal and Pt(II) species. For samples implanted with 2 × 10¹⁶ at. cm⁻², 94.0% of the Pt was present as Pt metal. On the contrary, samples implanted with 6 × 10¹⁵ at. cm⁻² displayed a Pt(0) concentration of 80%. Also Pt–O O 1s signal at 530 eV showed a higher contribute for the latter film. Authors stated that large prevalence of the metallic form in both samples was due to its precipitation as a result of the energy of ion implantation and immiscibility of Pt in PPy or to its reduction by PPy film. Finally, authors explained that Pt 4f_7/2 BE values (72.1 and 72.6 eV measured for implanted films) were higher than that reported for bulk Pt (Pt 4f_7/2 = 71.2 eV) as a consequence of the interactions between Pt atoms and PPy.

Fig. 9 Depth profile of a PPy film implanted with 2 × 10¹⁶ Pt at. cm⁻² after etching with argon. Concentrations are based on integrated survey spectra taken after every 30 s of sputtering. Reproduced, with permission, from ref. 50.

In Li et al.,⁵¹ XPS measurements confirmed the formation of Pt@Au nanoparticles with core–shell structure, prepared on a GC electrode, to be used as electrochemical sensor for H₂O₂. Authors used copper under-potential deposition technique (Cu UPD) on the surface of a GC electrode preliminarily functionalized with Au nanoparticles, immobilized on ethylenediamine (Au/EDA/GC). A redox replacement of copper with platinum was then performed to fabricate the sensor, denoted as Pt@Au/EDA/GC. The procedure scheme is illustrated in Fig. 10. Accordingly, high resolution peaks of the Au 4f_7/2 at 84.0 eV confirmed the zero-valent Au and Pt 4f_7/2 peak at 71.2 eV, the zero-valent Pt presence. N 1s peak at 399.8 eV evidenced the EDA grafting on the surface of GCE, while the lack of Cu 2p signals indicated that galvanic exchange was complete and that Au nanoparticle surfaces were Cu-free. Although XPS data thus confirmed that Pt@Au nanostructures were formed through Cu UPD process, they were not enough to prove the Pt@Au core–shell structure. However, authors reported that XPS results, combined with cyclic voltammetry data, proved the formation of core–shell nanostructure, showing the latter few exposed Au sites during the successive cycling and the pinhole-free nature of the Pt coating.

Fig. 10 Schematic representation of the procedure for preparing the ultrathin platinum-coated gold nanoparticles on a glassy carbon electrode surface. Reproduced, with permission, from ref. 51.

Chen et al.⁵² investigated by XPS the surface composition of Pd nanoparticles and flower shaped (FS) Au@Pd nanoparticles, a starting material used to make Au@Pd-ILs–Au@Pd/GCE sensor for glucose, employing ILs (i.e., trihexyltetradecylphosphonium bis(trifluoromethylsulfonyl)imide). Pure Pd nanoparticles showed peaks at 335.35 eV, corresponding to Pd 3d_5/2 of Pd(0). For FS Au@Pd nanoparticles the corresponding peak appeared at 335.06 eV. The negative shift of Pd 3d_5/2, i.e. −0.29 eV, as well as the wider FWHMs for Au@Pd nanoparticles than that for pure Pd, were reported from authors to confirm that the surface of the shell nanoparticles consisted of Pd(II). Nevertheless, it should be remarked that Pd 3d detailed spectra reported by authors, for pure Pd and for Au@Pd nanoparticles, did not show what commented in the text. Also the attribution of Pd 2p components to Pd(II) is not clearly referred to any features of the spectra. For gold, a very weak Au 4f signal was observed (Au 4f_7/2 at 83.90 eV), suggesting the presence of the Au core embedded under the Pd shell, with Au signals coming from the interstices of Pd grains.

The power of XPS analysis in investigating the structure of the sensing layer is clearly demonstrated by its application to nanostructured materials for which the shift in binding energy has been demonstrated to be due to the nanosize effect.⁵³ The formation of silver in nanometer size has been demonstrated by XPS in a study performed in our lab, devoted to the development of a glucose sensor based on a Pt electrode modified by a nanocomposite made of silver nanoparticles capped in a commercial nontoxic polyvinyl alcohol (PVA) matrix, obtained by reduction of AgNO₃ in ethanol.⁵⁴ C 1s, O 1s, and Pt 4f signals were identified and were attributed to PVA and to the substrate. Ag 3d XPS region was studied in detail evidencing the main component at 369.0 eV ascribed to Ag(0) in the nanometer size due to its BE significantly higher than typical values for bulk Ag (368.0 eV). The second minor component detected at around 367.5 eV was ascribed to unreacted AgNO₃ as BEs of Ag(I) species fall generally in this range.

As mentioned above, information drawn by XPS analysis can be used for formulating hypothesis on the formation mechanism of the sensing material, with the subsequent possibility of properly modifying experimental conditions for improving the fabrication and, consequently, the sensing steps. Such an application of XPS can be found in a work of Liu and collaborators⁵⁵ who proposed an electrochemical sensor for acetaminophen based on electrochemically reduced graphene (ERG) loaded nickel oxides (Ni₂O₃–NiO) nanoparticles coated onto glassy carbon electrode (ERG/Ni₂O₃–NiO/GC). In particular, by applying a potential of −1.5 V (vs. SCE) the simultaneous reduction of graphite oxide and deposition of graphene together with nickel oxides on a GC electrode was achieved. The advantage of the method is that high quality graphene can be prepared by electrochemical reduction eliminating the use of excessive reducing agents and thus addressing the concern of contaminant resulting products. From the comparison of C 1s XPS spectra of reduced graphene and graphite oxide the authors concluded that an incomplete reduction process occurred, although a significant decrease of oxygen-bound carbons was observed. The presence of nickel was confirmed by the analysis of Ni 2p high resolution spectra exhibiting two components, one at 856.3 eV attributed to NiO and one at 855.8 eV attributed to Ni₂O₃. On this basis the authors postulated a formation mechanism of nanoparticles based on the following reaction:

NiO + Ni(OH)₂ → Ni₂O₃ + 2H⁺ + 2e⁻

Moreover, they hypothesized that this local transformation occurred at the NiO/Ni(OH)₂ interface. Nonetheless, it should be observed that the hypothesized formation of Ni₂O₃ at a potential of −1.5 V is quite difficult to occur. It could derive instead from the re-oxidation of NiO/Ni(OH) in air and from the incomplete reduction process, evidenced by XPS data.

Wang et al.⁵⁶ reported a multi-layered gold nanoparticles/polyaniline/halloysite nanotubes (AuNPs/PANI/HNTs) nanostructures for electrochemical detection of hydrogen peroxide. The PANI/HNTs were firstly obtained by in situ polymerization employing the thioglycollic acid (TA) as a dopant; thereafter, Au ions were anchored to TA and then reduced by PANI in a HAuCl₄ solution. The XPS measurements were used to detect the surface elemental composition and study the chemical process in depth. The XPS N 1s spectrum of PANI/HNTs exhibited three peaks at 399.2 eV, 399.9 eV and 400.9 eV, which were assigned, respectively, to quinonoid imine (=N–), benzenoid amine (–NH–) and positively charged nitrogen (N⁺). After the fabrication of AuNPs/PANI/HNTs nanostructures, the binding energy of the three different states slightly shifted, with a significant modification of the relative intensity between benzenoid amine and quinonoid imine, implying that the benzenoid amine were oxidized to quinonoid imine. This suggested that it was mainly the nitrogen lone electron pair to interact with the metal matrix during the redox reaction between PANI and HAuCl₄. As for as S was concerned, S 2p_3/2 shifted from 163.7 to 162.4 eV in presence of gold, as reported for sulfur attached to metal, while Au 4f doublet corresponded well to reduced gold.

4. The use of XPS for evaluating the stability of the sensing material

In a few literature works the use of XPS analysis with the aim to evaluate the stability of the sensing material, both during its fabrication and under operative conditions, is reported. Such applications provide further evidence of the close link between XPS analysis and sensor technology representing the former a powerful instrument for selecting optimal conditions not only during sensor fabrication step, but also during its use, with subsequent possibility of improving sensor performances. The stability of CNTs deposited on electrode surface under O₂ plasma treatment was evaluated by XPS.⁵⁷ The authors used O₂ plasma treatment to insert oxygen moieties on carbon nanotubes surface with the aim to anchor DNA on CNTs as probe sensor. Following peaks at 286.4, 287.5, and 288.6 assigned respectively to –C–O (alcohol, ether), [double bond splayed left]C O (ketone, aldehyde), and –COO (carboxylic acid, ester) species on MWCNTs (Fig. 11), they observed an increase of peaks intensity until 30 s of plasma treatment, and a slightly decrease after 60 s suggesting that MWCNTs lost their intrinsic properties due to an impairment of the electronic structure that made impossible their use in sensor applications. In a similar work, Park et al.⁵⁸ used XPS analysis for comparing MWCNTs treated with O₂ plasma for different time intervals with the aim to correlate the modification of chemical composition to the modification of electrochemical performances in the detection of target DNA. Also in this work, the plasma treatment was performed to increase the electrochemical active area and additionally provide chemical functional groups on the surface of the MWCNTs. C 1s peaks at 284.2 eV and at 285.5 eV were observed for all MWCNT films, with the former ascribed to both sp²-hybridized and hydrogen-bound carbon atoms, and the latter to sp³-hybridized carbon atoms. The intensity of these two signals was found to depend on the time of the plasma treatment as shown in Fig. 12A. The results indicated that the sp²-hybridized carbons decreased while the sp³-hybridized carbons increased as the plasma treatment time increased. The peak ratio of the sp³ to sp² increased as the treatment time increased meaning that the plasma treatment developed many defects on the MWCNTs surface. The other three C 1s peaks identified at 286.4, 287.5, and 288.6 eV were ascribed to oxygen-bound carbons (–C–O, [double bond splayed left]C=O, –COO species), and their relative amounts were evaluated as a function of the plasma treatment time (Fig. 12B) evidencing their increase with the increase of the plasma treatment time with saturation in 10 s. In particular, carboxylic acid groups had a key role in the proposed DNA sensor applications allowing the formation of amide bonding with the primary amine-modified DNA. On the basis of XPS results, the authors concluded that the plasma treatment introduced oxygen to defect sites disrupting the inherent structure of the MWCNTs and, along with electrochemical data, selected the optimized condition of the plasma treatment as 30 s.

Fig. 11 C 1s XPS spectra of (a) pristine, and plasma treated CNTs for (b) 10s, (c) 15s, (d) 30s and (e) 60s. Reproduced, with permission, from ref. 57.

Fig. 12 Dependence of specific chemical structure of bare pristine and plasma-treated MWCNTs on plasma treatment time determined by XPS spectra. (A) C 1s peaks of sp² hybridized graphite carbon and sp³ hybridized hydrogen-bound carbon, (B) C 1s peaks of oxygen-bound carbons. Reproduced, with permission, from ref. 58.

An extensive use of XPS in checking the stability of the sensing material is reported in a work of He and co-workers⁵⁹ who developed an electrochemical sensor for the detection of DNA. In particular they deposited propargylamine on graphene surface by plasma polymerization with the aim to exploit the high electrochemical activity of graphene and the ability of propargylamine to assist the immobilization of DNA thanks to the high density of amino group. In order to verify the possible propargylamine degradation after the exposure at different plasma power, the authors performed XPS analysis of films prepared using different plasma input power. They observed that, for higher value of plasma input power, the relative intensity of C–O/C–N component at 286.4 eV of C 1s spectra attributed to propargylamine decreased with respect the component at 283.9 and 284.6 eV attributed to C=C and C–H/C–C groups, respectively. This result was attributed to the formation of small volatile molecules such as CO, CO₂, NO, or NO₂ formed at higher input plasma power and pumped out by the instrument. Most importantly, the authors analyzed N 1s spectra of the film observing not only the C–N and N–H components at 399.3 eV of polypropargylamine, but also C=N and C–N=O components respectively at 398.8 and 400.8 eV due to the dissociation of monomer molecules into smaller species under higher plasma power. Hence C=N groups were produced by dehydrogenation of C–N whereas C–N=O resulted from the reaction between oxide species present on graphene surface and with C–N groups of propargylamine.

The integrity of sensing material also under working conditions has been successfully monitored by XPS investigation. For instance, in a recent work, Ma et al.⁶⁰ proposed the use of XPS for selecting optimal sensor working conditions in the development of a poly(1-butyl-3-[3-(N-pyrrole)propyl]imidazolium tetrafluoroborate ionic liquid) film electrode for the electrochemical detection of bisphenol A. The total film disappearance after electrode incubation with aqueous solution was observed, being ascribed to the high film hydrophilicity. The ionic liquid polymer was therefore treated with sodium dodecyl sulfate solution to exchange anions thus obtaining a hydrophobic interface. XPS analysis confirmed the efficiency of both film electropolymerization and anion replacement being B 1s at 192.9 eV, N 1s at 400.2 eV and F 1s at 684.3 eV observed on electrode after polymer deposition and S 2p at 169.4 eV (erroneously denoted as S 1s by authors) appearing after polymer treatment with sodium dodecyl sulphate solution, with the simultaneous disappearing of B 1s and F 1s signals. The as treated electrode exhibited a significantly higher long-term stability, confirming the need to use a hydrophobic electrode surface in the proposed sensing applications.

The survey of literature data evidences that such a use of XPS characterization, although able to provide useful information in sensing application, is reported in a limited number of cases. Considering the possibility of casting light on key aspects related to sensor lifetime and to possible causes of its deterioration, this certainly represents a potentially expanding field of XPS analysis in sensing technology.

5. The use of XPS for elucidating the sensing mechanism: analysis of modifications of sensing material during/after the analyte detection

The paramount importance of XPS analysis in the design of electrochemical sensors through the characterization of sensing element is certainly demonstrated by research works focused on the study of chemical modifications occurring during and after the interaction with the analyte, demonstrating the great potential of XPS analysis in investigating the analyte binding event. Interesting information about the analyte binding process were drawn by XPS analysis in the development of an electrochemical sensor for copper ion based on a composite material integrating a tripeptide, namely Gly-Gly-His, and a poly(3-thiopheneacetic acid) (PTAA) layer.⁶¹ Evidences regarding the successful attachment of Gly-Gly-His to PTAA were firstly obtained from the N1s high resolution spectrum, fitted into three components: one at 398.9 eV, attributed to the C=N nitrogen of the imidazole side chain of histidine, one at 400.1 eV assigned to the amide nitrogen while the highest binding energy component (401.5 eV) originated from the C–N nitrogen of the imidazole ring. More importantly, XPS analysis allowed investigating copper ion–tripeptide interaction. XPS spectra of Cu 2p were acquired for PTAA and Gly-Gly-His modified PTAA electrode following their immersion in copper ion solutions. Cu 2p_3/2 peak at 933.0 eV confirmed the complexation of copper ions by the tripeptide modified electrode, indicating that the surface-exposed Gly-Gly-His could effectively capture copper ions.

A similar XPS application was reported in a detailed study of Yang and his co-workers⁶² who developed gold electrodes modified with self-assembled monolayers of L-cysteine for the adsorptive stripping analysis of copper. Remarkably, XPS allowed to verify not only the successful monolayer deposition, but also to extract useful information about copper complexation process, revealing great agreement with electrochemical results. Copper XPS spectra revealed the presence of Cu(II) and Cu(I) peaks after L-cysteine modified electrode immersion in copper solution; in particular, the appearance of a new peak at 932 eV after exposure to higher copper concentration suggested the formation of a new surface species, coherently with what observed in cyclic voltammetry experiments. On L-cysteine adsorbed on gold, S 2p peak exhibited a negative shift of about 1.5 eV in comparison with the signal in the free species because of the Au⁺S⁻ thiolate binding while it did not change after coordination with copper due to the strength of sulfur–gold interaction. On the contrary, some changes were observed on N 1s spectrum upon L-cysteine interaction with copper. While on L-cysteine monolayer, N 1s peaks appeared at 401.4 eV (NH₃⁺) and 399.4 eV (NH₂) with a ratio NH₂/NH₃⁺ of 0.3, it increased to 1.15 after the coordination with copper ion suggesting that most of the ammonium groups transformed to NH₂. Moreover, the position of energy peak at 399.4 eV was close to that found for amino acid copper complex. Finally, O 1s peak and its modifications after copper coordination evidenced the involvement of oxygen bonding to the copper and allowed to propose that Cu²⁺ was complexed by coordination with acidic (COOH) and basic (NH₂) functional groups of cysteine.

A remarkable application of XPS analysis in the study of analyte-sensing element interaction was proposed by Wang et al.⁶³ who exploited surface analysis to characterize the change of chemical composition of a mannosylated polyaniline film (manno-PANI) upon its binding to the protein concanavalin A (ConA) (Fig. 13A). XPS results suggested that the ConA binding with manno-PANI triggers the switching of amine functionalities in the PANI backbone, converting them to imine forms. This was clearly revealed by N1s signal of manno-PANI before and after interactions with the protein (Fig. 13B), displaying in the former three components corresponding to amine (398.8 eV) coming from the amine group in the PANI emeraldine form, imine nitrogen (398.0 eV) as a result of amine deprotonation in the quinone structure of manno-PANI, and radical cationic nitrogen (399.5 eV) resulting from the protonation of the quinonoid imine. After interaction with ConA, N 1s peaks exhibited a shift to lower binding energy states, along with a decrease of both protonated nitrogen and radical cationic nitrogen and an increase of imine nitrogen, suggesting that most of emeraldine was converted to pernigraniline on PANI skeleton upon protein binding.

Fig. 13 (A) Deprotonation of mannosylated PANI during the ConA binding. (B) High-resolution N1s XPS core-level spectra of manno-PANI (left) and after manno-PANI interaction with ConA (right). Reproduced, with permissions, from ref. 63.

Surface analysis was extensively used in the development of an electrochemical sensor for chemical warfare agents (CWA) based on aminoferrocene derivatives.⁶⁴ A cystamine conjugate was prepared by coupling cystamine with a N-protected ferrocene amino acid derivative and assembled on a gold electrode (Fig. 14A). Upon deprotection of the amino group, its reaction with CWA mimics (namely, EtSCH₂CH₂Cl (MA) and (CN)(EtO)₂P(O) (NA)) could be electrochemically followed due to the modifications of ferrocene groups. XPS results clearly showed the attachment of cystamine conjugate as well as its reaction with CWA mimics. The binding energy of the S 2p at 161.8 eV was indicative of Au–S bond between thiolate sulfur and Au surface. After the reaction with MA, S 2p signal showed peaks at 162.2 and 163.6 eV as in the case of cystamine conjugate alone, along with two additional peaks at 163.8 and 165.0 eV, which were ascribed to sulfur species from MA not bounded to Au surface. Additional evidence for the surface reaction was provided by the N 1s spectra (Fig. 14B). The N 1s spectra of cystamine conjugate exhibited two distinct peaks at 399.3 and 399.9 eV, attributed to the N-atom of carbamide and to amide (N–C=O), respectively. The former was mostly influenced by the binding event with NA with a pronounced shift to 397.8 eV and the presence of some residual free amine at 399.3 eV. The authors attributed the 397.8 eV signal to electron withdrawing effects, which hardly can justify the claimed increase of electron density on N atom.

Fig. 14 (A) Schematic view of surface modifications of compound 2 (a) in solution, (b) on surface, (c) deprotection of 2 to 2-d, (d) exposure of 2-d with MA and (e) exposure of 2-d with NA. (B) XPS N 1s spectra for (a) 2, (b) 2-d, (c) 2-d + MA, and (d) 2-d + NA films on gold surface. Reproduced, with permission, from ref. 64.

The sensing process was extensively studied by XPS analysis in the development of a bifunctional electrochemical sensor for pentachlorophenol (PCP) and copper ions which was prepared by a layer-by-layer approach⁶⁵ alternately assembling humic acid (HA) and exfoliated Mg–Al-layered double hydroxide (LDH) nanosheets onto ITO substrates. XPS was used to compare the as-prepared (LDH/HA)₈/ITO film before and after dipping into the solution containing Cu(II) ions or PCP. After immersing (LDH/HA)₈/ITO into PCP or Cu(II) solution, the binding energy values of C 1s, O 1s and N 1s of the film were remarkably shifted, indicating an alteration of the local bonding environment of (LDH/HA) multilayer film. According to authors, it was due to PCP interaction with humic substances sites through hydrogen bonds and hydrophobic interactions. On the other hand, the high resolution XPS spectrum showed Cu 2p_3/2 peak at 931.5 eV, attributed to Cu(II) species, as confirmed by the shape of the photoelectronic peaks showing characteristic shake-up features at 942.2 eV. It was thus concluded that the uptake of Cu²⁺ was accompanied by the formation of Cu(II)–O, that is, that oxygen containing functional groups in (LDH/HA) multilayer film, such as phenolic, deprotonated carboxylic acid, and hydroxyl groups, acted as a ligand to form complexes with Cu²⁺.

The sensing mechanism occurring in a non-enzymatic Pt NP/MWCNTs paste-based mediated glucose sensor was hypothesized by Xie et al.⁶⁶ on the basis of XPS data. According to authors, the high sensitivity of the sensor could be ascribed, at least in part, to the enhanced electron transfer. With the aim to verify such hypothesis, they performed XPS analysis of the electrode before and after electrochemical detection of glucose. In particular they observed that before glucose testing two peaks at 288.6 and 287 eV were present, which were attributed, respectively, to C=O from carboxylic acid group and hydroxide group formed during the fabrication of the sensor, which disappeared after glucose sensing. The authors hypothesized that, under the positive potential applied during glucose sensing, some oxygen atoms bonded to MWCNTs were released and used as oxygen source for Pt NP to form platinum oxide. This hypothesis was supported by XPS analysis of Pt 4f region showing a small shift (∼0.2 eV) at higher binding energy after glucose sensing, suggesting the formation of oxidized species on platinum surface. Moreover, the absence of C=O groups on Pt NPs was considered as indicative of the proximity between nanoparticles and MWCNTs surface, promoting the electron transfer and consequently increasing sensor sensitivity. However it must be underlined that the claimed 0.2 eV chemical shift is very close to the limit of energetic accuracy.

The usefulness of XPS analysis in elucidating the sensing mechanism is demonstrated also in a work of Hua and co-workers⁶⁷ who developed a MWCNTs/polybenzimidazole (PBI) electrode for the analytical determination of H₂O₂ by exploiting the possibility of imine structures of PBI to be chemically oxidized in the presence of peracetic acid (a mixture of acetic acid (AcOH) and H₂O₂) to form N-oxides. In order to support the proposed mechanism, the authors performed XPS analysis of PBI before and after exposure to peracetic acid (Fig. 15). In particular by fitting N 1s peak they found two peaks at 398.4 eV and 400.2 eV ascribable, respectively, to imine and amine groups. By quantitative analysis they estimated a ratio of 1 between the two components. After the exposure of the polymer to peracetic acid, N 1s spectrum showed a new peak at 402.0 eV indicating the presence of N-oxide. In this case, the ratio of the peaks areas N-imine : N-amine : N-oxide was found to be 0.86 : 1 : 0.1 suggesting that ∼14% of the reactive sites of the imine of PBI were oxidized by peracetic acid. Moreover, increasing the molar ratio H₂O₂ : AcOH caused a further increase in the amount of oxidized reactive sites of the imine, suggesting the effective possibility to use the system as a sensor for H₂O₂ detection.

Fig. 15 N 1s XPS spectra of (a) PBI and of solutions with molar ratios of PBI : H₂O₂ : AcOH of (b) 1 : 0.5 : 1, (c) 1 : 1 : 1 and (d) 1 : 2 : 1. Reproduced, with permission, from ref. 67.

Also in the design of biosensor systems, the possibility of gaining useful information about the binding event by XPS analysis has been demonstrated during the last years, especially in the DNA biosensor field. An example of DNA hybridization monitoring by XPS analysis was reported by Dharuman et al.⁶⁸ who assembled a composite monolayer by sequential adsorption of thiol modified single stranded DNA (HS-ssDNA), 6-mercaptohexanol (MCH) and 3-mercaptopropionic acid (MPA) (abbreviated as HS-ssDNA/MPA/MCH), applied for label free DNA hybridization electrochemical sensing (Fig. 16). From XPS analysis, the elemental composition for pure DNA layers before and after (denoted as HS-dsDNA) hybridization suggested poor hybridization efficiency (calculated from the ratio of atomic percentage of the hybridized to the un-hybridized surfaces, Fig. 16). Similar observations were done after the MCH and the MPA treatment to form binary mixed monolayer suggesting also in this case reduced target hybridization. On the contrary, the ternary layer showed increased ratio values for N and P percentages, evidencing the enhanced hybridization efficiency. According to authors, this was further corroborated from the increased N 1s peak area ratio, being 1.28 in the case of ternary HS-ssDNA/MPA/MCH layer and 1.15 and 1.22, respectively, on the binary layers HS-ssDNA/MCH and HS-ssDNA/MPA. Nevertheless, it should be pointed out that the reported differences are too low to be considered highly significant.

Fig. 16 (Left) Scheme of ternary HS-ssDNA/MPA/MCH monolayer for simultaneous DNA orientation control and surface passivation for electrochemical label free DNA hybridization discrimination. 6-Mercapto-1-hexane–ssDNA (HS-ssDNA); 3-mercapto-propionic acid (MPA); 6-mercapto-1-hexanol. (Right) Elemental compositional data for the hybridized and un-hybridized layers of pure and mixed layers constructed on the Au surfaces. Reproduced, with permission, from ref. 68.

An interesting example of XPS application to the study of biorecognition process was proposed by Martić et al.⁶⁹ who followed the process of kinase-catalyzed phosphorylation of peptides by XPS analysis. In this work, an electrochemical sensor was fabricated exploiting the ability of particular protein kinases to specifically transfer a redox-labeled phosphoryl group to peptides attached to a gold substrate (Fig. 17). In particular, the process could be electrochemically monitored as a ferrocenoyl-ATP (Fc-ATP) was used as a co-substrate for the phosphorylation of peptides. After phosphorylation the results of high-resolution XPS analysis of Fe 2p region showed the presence of Fe 2p_3/2 at 708.3 eV which was ascribed to Fe(II) of ferrocene group. Importantly, in the control experiment performed in the presence of Fc-ATP and in the absence of protein kinase, no signals related to iron were observed by XPS.

Fig. 17 Schematic illustration of the stepwise assembly of the electrochemical biosensor surface for detection of protein kinase-catalyzed Fc phosphorylation using Fc-ATP. (A) Initial formation of a thin film of NHS-ester on Au surfaces is followed by (B) peptide incubation resulting in the formation of a film containing target peptides chemically bound to the transducer surface. (C) After blocking with ethanolamine, which reacts with the excess NHS ester on the surface, and back-filling with dodecanethiol, the protein kinase-catalyzed Fc-phosphorylation reaction was carried out in the presence of kinase and Fc-ATP co-substrate. Reproduced, with permission, from ref. 69.

In another work⁷⁰ the development of a DNA-based biosensor for Vibrio cholerae is reported and the study of hybridization process by XPS analysis is proposed. The fabrication of the device was performed in successive steps: firstly, nanostructured magnesium oxide (nMgO) grafted carboxyl functionalized multi-walled carbon nanotubes (nMgO–cMWCNTs) were deposited electrophoretically onto indium tin oxide (ITO) coated glass electrode (nMgO–cMWCNTs/ITO). After then, covalent immobilization of amine-terminated single stranded DNA, specific for Vibrio cholerae (NH₂–ssDNA), onto nMgO–cMWCNTs/ITO surface was achieved. XPS analysis was reported for cMWCNTs/ITO, nMgO–cMWCNTs/ITO, NH₂–ssDNA/nMgO–cMWCNTs/ITO systems and also after hybridization with fragmented target DNA (dsDNA/nMgO–cMWCNTs/ITO). In survey spectra, peaks related to C 1s and O 1s were observed at 284.5 eV and 529 eV, for all the samples; moreover, Mg 2p at 49.6 eV was observed only for the nMgO modified electrodes. A peak related to N 1s observed at 397 eV onto NH₂–ssDNA/nMgO–cMWCNTs/ITO surface confirmed that DNA was covalently functionalized. Fitting of C 1s region for cMWCNTs/ITO, nMgO–cMWCNTs/ITO and dsDNA/nMgO–cMWCNTs/ITO films resulted in peaks at 284.5 eV and at 285.6 eV, assigned to graphitic carbon (sp² C=C) and to the defects on the nanotube structure (sp³ C–C), respectively. The other peaks at 286.6, 287.8 and 289.4 eV were attributed to the C–O, C=O (carbonyl) and O–C=O (carboxyl) groups respectively, and reported as a result of different oxygenated moieties during chemical functionalization. In case of dsDNA/nMgO–cMWCNTs/ITO sample, an additional broad peak found at 287.9 eV was attributed to amide bond (N–C=O) formation between carboxyl group present at nMgO–cMWCNTs/ITO electrode surface and the aminated DNA. The carboxyl group percentage of 17.6% in nMgO–cMWCNTs/ITO sample decreased to 7.3% in dsDNA/nMgO–cMWCNTs/ITO sample, indicating that most carboxyl groups were involved in DNA functionalization via amide bond formation. Nevertheless, some criticisms in the drawn conclusions should be evidenced. First of all, the appearance of a novel component at 287.9 eV corresponding to amidic groups is particularly difficult to discriminate from the component at 287.8 eV, as clearly shown by C 1s reported spectra. For the scope of the work, preliminary XPS analysis of amine-terminated single stranded DNA alone could have been helpful for better evaluating the possible modification of C 1s signal after DNA functionalization, as well as the use of angle-resolved XPS for analyzing near surface regions differently affected by the presence of DNA. Finally, for unequivocally demonstrating DNA binding process, XPS signals before and after hybridization should be compared.

The role of XPS as a tool for identifying the species involved in sensing mechanism has been demonstrated in our laboratory carrying out some works on platinum electrodes modified with tellurium species, as micro and nano-structures, to develop electrochemical sensors for glucose and hydrogen peroxide detection.^54,71,72 Reported XPS characterization helped to highlight that tellurium species were spontaneously adsorbed on platinum surface and confirmed that the tellurium species involved in electrochemical processes were substantially Te(0) with a peak at 573.3 eV (Te 3d_5/2) and Te(IV) at 576.6 eV. Instead, the absence of the peak component at 577.8 eV related to Te(VI) after the electrochemical treatment evidenced that this species was not stable at used pH conditions.⁷³ In addition, the component at 575.6 eV was attributed to Te(II)–Pt adsorbed species.^71,74,75

6. Insight into prospective applications of XPS in sensors

XPS technology has enormously evolved since its first application in sensor field. Development of technique has been reviewed even recently.^76–78 Several exciting novelties have been proposed and applied to characterization of new materials, but new developments have not found yet application in sensor investigation, so that the application to this field seems still in its infancy.

A list by no means exhaustive of improvements/innovations could include:

- Potential dependent measurements.

- Improved spatial and time resolution.

- High-brightness sources.

- High ambient sample pressure.

- In liquid measurements.

- In working conditions measurements.

- Expert systems.

- New data analysis.

Present XPS capabilities are well illustrated in Fig. 18 from a recent review.⁷⁹

Fig. 18 Illustration of a typical experimental configuration for X-ray photoelectron spectroscopy experiments, together with the various types of measurements possible, including (a) simple spectra or energy distribution curves, (b) core-level photoelectron diffraction, (c) valence-band mapping or binding energy vs. k plots, (d) spin-resolved spectra, (e) exciting with incident X-rays such that there is total reflection and/or a standing wave in the sample, (f) using much higher photon energies than have been typical in the past, (g) taking advantage of space and/or time resolution, and (h) surrounding the sample with high ambient sample pressures of several torr. Reproduced, with permission, from ref. 79.

Among them, two emerging applications seem particularly interesting in relation to study of sensor mechanisms. In a first example, Suzer⁷⁶ reported an XPS investigation of a Si-diode performed during its operation under bias. The XP spectrometer was modified in order to perform the analysis of the sample subject to an external voltage.

In particular, the diode was grounded from the n-side and the potential was applied to p-side. XPS analysis performed on the junction region (ca. 100 μm spot size) when no bias was applied, showed one Si⁰ peak and one Si oxide peak (Fig. 19). The Si peak contained both p-Si and n-Si components. When +8 V reverse bias was applied, only the p-component (Si⁰ and oxide) was +8 V up-shifted in the binding energy scale, as expected. This work suggests, in combination with the possibility to collect spectra at high pressure, that XPS experiments can be executed under working conditions for entire devices (e.g. sensors). On this basis, application to investigation of FET sensors exposed to gas analytes can be forecast.

Fig. 19 The Si 2p region of the p–n junction of a Si-diode recorded under: no bias and +8 V reverse bias. The inset schematically displays the electrical connection. Adapted from ref. 76.

Even more exciting seems the possibility to perform XPS experiments in liquid under potential application, simulating working conditions for electrochemical sensors. It represents in some way the combination of several developments sited. Very recently, a new method to investigate by XPS solid–liquid interface was reported by Axnanda et al.⁷⁷ A new AP-XPS system (ambient pressure XPS), including a three-electrode electrochemical cell in the analysis chamber and employing a “tender” X-ray synchrotron source, was used for the experiments. This apparatus was used to create a stable nanometer-thick thin liquid film on a platinum electrode by a “dip & pull” method. Moreover, the authors demonstrated that this system can be used for electrochemistry studies by probing Pt oxidation under working conditions. Detailed schemes of three-electrode electrochemical cell and of the “dip & pull” procedure, able to form a thin electrolyte layer on the employed Pt working electrode (WE), are shown in Fig. 20. At first, the authors reported a XPS study to demonstrate the formation of a thin electrolyte layer on the Pt electrode surface after the “dip & pull” procedure. In particular, they showed the change of Pt 4f spectra under changing experimental conditions (under vacuum condition, water vapor pressures between 16 Torr and 20 Torr and different potentials with CV treatment). In this way, they revealed the photoelectron attenuation from electrolyte layer. Moreover, the authors calculated the thickness of the electrolyte layer (∼13 nm) through the analysis of the Pt 4f intensity attenuation. This thin film was stable when water vapor pressure was constant. Secondly, the authors evaluated changes of XPS Pt 4f signal (relevant to Pt working electrode) and O 1s and K 2p (relevant electrolyte layer) upon variation of the applied electrochemical potential. The experiments showed no BE shift on Pt 4f, as expected, due to the fact that the Pt WE was grounded along with the XPS analyzer. As far as the O 1s spectrum was concerned, two peaks located at BE of 534 eV and BE of 537 eV were ascribed to electrolyte liquid phase and water vapor phase, respectively. The “vapor” peak was shifted in respect to “liquid” one due to final state effects.⁸⁰ More interesting was to compare the BE positions of bulk liquid peaks taken at −0.8 V and −0.4 V (holding potential). In fact, the shift of O 1s signal (0.4 eV) nicely corresponds to the electrochemical potential change (0.4 V) applied to the working electrode, demonstrating the conductive nature of the thin electrolyte film formed, able to preserve the electrochemical potential difference across the solid–liquid interface. The K 2p spectra showed similar trend in BE shift as the liquid phase O 1s spectra. To further demonstrate the capability of proposed approach, the authors conducted an in situ Pt oxidation of the Pt WE. This analysis was conducted changing the applied voltage to 0.0 V and 1.2 V. At 0.0 V holding potential, Pt 4f showed a single peak (metallic peak) located at BE of 71.2 eV, whereas at 1.2 V holding potential, additional peaks appeared at higher binding energy. Therefore, two oxidized components (BE 72.6 eV (Pt²⁺) and 74.1 eV (Pt⁴⁺)) were introduced in the fit of Pt 4f_7/2 signal. This demonstrated a clear oxidation state change of the Pt WE and that electrochemistry under working conditions can be investigated by this novel strategy.

Fig. 20 Schematics of three-electrode electrochemistry setup in the AP-XPS chamber. (a) Positions of electrodes before immersion and corresponding representative Pt 4f, O 1s spectra and electrochemical profile. (b) Electrodes are immersed in the electrolyte, where any electrochemical treatment can be performed within the AP-XPS chamber. Shown is a representative Pt foil CV in 6 M KF aqueous electrolyte. (c) Electrodes are placed at the AP-XPS measurement position, and corresponding representative Pt 4f and O 1s of the partially removed electrodes are overlaying the representative vapor exposed electrode spectra. (d) is an image of the 3-electrode apparatus that has been “dip & pulled” from the electrolyte in the beaker and placed into XPS position while ensuring all three electrodes are in contact with the electrolyte within the beaker. Reproduced from ref. 77.

Finally, achievements have also been made in data analysis. A very promising tool is described in the work by Baltrusaitis et al.⁷⁸ They proposed to interpret complex spectra arising from several species of the same element (Mo) replacing traditional XPS spectra fitting procedures using purely synthetic spectral components with experimental spectra components obtained investigating suitable standards. To this aim, mixed species systems originated from degradation of MoO₂ and MoO₃ were preliminarily studied and spectral components relevant to Mo(IV), Mo(V) and Mo(VI) were extracted employing principal component analysis (PCA). Then those components were successfully applied to rationalization of Mo spectra from electrodeposited amorphous molybdenum oxide samples. A very interesting feature was represented by the use of spectral components combining peaks of different elements from the same species (e.g. Mo, O, etc.). This approach allows to introduce stoichiometry constrains, as determined through experimental spectra of standards, in complex spectra fitting. In addition, the proposed strategy avoids large errors which can occur in traditional fitting procedures^81,82

7. Conclusions

The wide reviewed literature demonstrates XPS has been of fundamental importance in the development of chemical sensors, employing many potentialities of this technique. In the same manner, it can be easily forecast that the exciting development expected (see e.g. ref. 79) for XPS in the direction of improved spatial and time resolution, high-brightness sources, high ambient sample pressure (several torr), etc. will be quickly applied to the sensor field to enlighten more subtle details in measurement conditions closer to working ones, offering a powerful tool in improvement of new chemical sensors.

Acknowledgements

This activity was partially funded by the Italian Ministry of University and Research (MIUR) Futuro in Ricerca (FIR) programme under Grant no. RBFR122KL1_003 (SENS4BIO).

References

1. D. Briggs and M. Seah, Practical Surface Analysis, John Wiley and Sons, New York, USA, 2nd edn, 1990.
2. J. E. Castle, Surf. Interface Anal., 1984, 6, 302.
3. H. W. Nesbitt, G. M. Bancroft, R. Davidson, N. S. McIntyre and A. R. Pratt, Am. Mineral., 2004, 89, 878–882.
4. C. Malitesta and E. Margapoti, in Inorganic Micro- and Nanomaterials. Synthesis and Characterization, ed. A. Dibenedetto and M. Aresta, De Gruyter, Berlin, 2013, pp. 47–56.
5. C. Malitesta, G. Morea, L. Sabbatini and P. G. Zambonin, in Surface Characterization of Advanced Polymers, ed. L. Sabbatini and P. G. Zambonin, VCH Verlagsgesellschaft mbH, Weinheim, New York, Basel, Cambridge, Tokyo, 1993, pp. 181–216.
6. W. Prissanaroon-Ouajai, P. J. Pigram, R. Jones and A. Sirivat, Sens. Actuators, B, 2008, 135, 366–374.
7. P. Chandra, N. X. Son, H.-B. Noh, R. N. Goyal and Y.-B. Shim, Biosens. Bioelectron., 2013, 39, 139–144.
8. M. Martí, G. Fabregat, F. Estrany, C. Alemán and E. Armelin, J. Mater. Chem., 2010, 20, 10652.
9. L. Wang, Z. Feng and H. Cai, J. Electroanal. Chem., 2009, 636, 36–39.
10. D. Micić, B. Šljukić, Z. Zujovic, J. Travas-Sejdic and G. Ćirić-Marjanović, Electrochim. Acta, 2014, 120, 147–158.
11. A. Micoli, A. Turco, E. Araujo-Palomo, A. Encinas, M. Quintana and M. Prato, Chem.–Eur. J., 2014, 20, 5397–5402.
12. F. M. Toma, A. Sartorel, M. Iurlo, M. Carraro, P. Parisse, C. Maccato, S. Rapino, B. R. Gonzalez, H. Amenitsch and T. Da Ros, Nat. Chem., 2010, 2, 826–831.
13. X. X. Wang, J. N. Wang, H. Chang and Y. F. Zhang, Adv. Funct. Mater., 2007, 17, 3613–3618.
14. A. Fabbro, A. Sucapane, F. M. Toma, E. Calura, L. Rizzetto, C. Carrieri, P. Roncaglia, V. Martinelli, D. Scaini, L. Masten, A. Turco, S. Gustincich, M. Prato and L. Ballerini, PLoS One, 2013, 8, e73621.
15. V. Martinelli, G. Cellot, F. M. Toma, C. S. Long, J. H. Caldwell, L. Zentilin, M. Giacca, A. Turco, M. Prato and L. Ballerini, ACS Nano, 2013, 7, 5746–5756.
16. J. Yuan, X. Jin, N. Li, J. Chen, J. Miao, Q. Zhang, L. Niu and J. Song, Electrochim. Acta, 2011, 56, 10069–10076.
17. Y. Shan, G. Yang, Y. Jia, J. Gong, Z. Su and L. Qu, Electrochem. Commun., 2007, 9, 2224–2228.
18. S. Liu, D. G. Kurth, H. Mohwald and D. Volkmer, Adv. Mater., 2002, 14, 225–228.
19. H.-W. Chang, Y.-C. Tsai, C.-W. Cheng, C.-Y. Lin and P.-H. Wu, Sens. Actuators, B, 2013, 183, 34–39.
20. Y. Zhu, S. Murali, W. Cai, X. Li, J. W. Suk, J. R. Potts and R. S. Ruoff, Adv. Mater., 2010, 22, 3906–3924.
21. Q. Wang, I. Kaminska, J. Niedziolka-Jonsson, M. Opallo, M. Li, R. Boukherroub and S. Szunerits, Biosens. Bioelectron., 2013, 50, 331–337.
22. L. Cao, Y. Liu, B. Zhang and L. Lu, ACS Appl. Mater. Interfaces, 2010, 2, 2339–2346.
23. P. Wang, Z.-G. Liu, X. Chen, F.-L. Meng, J.-H. Liu and X.-J. Huang, J. Mater. Chem. A, 2013, 1, 9189.
24. G. E. De Benedetto, C. Malitesta and C. G. Zambonin, J. Chem. Soc., Faraday Trans., 1994, 90, 1495.
25. S. Cosnier and M. Holzinger, Chem. Soc. Rev., 2011, 40, 2146–2156.
26. H. Wang and S. Mu, Sens. Actuators, B, 1999, 56, 22–30.
27. C. Malitesta, F. Palmisano, L. Torsi and P. G. Zambonin, Anal. Chem., 1990, 62, 2735–2740.
28. C. Malitesta and M. R. Guascito, Biosens. Bioelectron., 2005, 20, 1643–1647.
29. M. R. Guascito, D. Chirizzi, C. Malitesta and E. Mazzotta, Analyst, 2011, 136, 164–173.
30. D. Chirizzi and C. Malitesta, Sens. Actuators, B, 2011, 157, 211–215.
31. M. R. Guascito, D. Chirizzi, C. Malitesta, L. Giotta, D. Mastrogiacomo, L. Valli and L. Stabili, Biopolymers, 2014, 101, 461–470.
32. J. Wang and J. A. Carlisle, Diamond Relat. Mater., 2006, 15, 279–284.
33. A. A. Abdelwahab, W. C. A. Koh, H.-B. Noh and Y.-B. Shim, Biosens. Bioelectron., 2010, 26, 1080–1086.
34. M. Coates and T. Nyokong, J. Electroanal. Chem., 2012, 687, 111–116.
35. H. C. Kolb, M. G. Finn and K. B. Sharpless, Angew. Chem., Int. Ed., 2001, 40, 2004–2021.
36. Z. Üstündağ and A. O. Solak, Electrochim. Acta, 2009, 54, 6426–6432.
37. S. Ciampi, B. Guan, N. A. Darwish, Y. Zhu, P. J. Reece and J. J. Gooding, Phys. Chem. Chem. Phys., 2012, 14, 16433–16439.
38. S. Ciampi, P. K. Eggers, G. le Saux, M. James, J. B. Harper and J. J. Gooding, Langmuir, 2009, 25, 2530–2539.
39. Y. Chen and J. Huang, Diamond Relat. Mater., 2009, 18, 516–519.
40. M. L. Yola and N. Atar, Electrochim. Acta, 2014, 119, 24–31.
41. K. N. Han, C. A. Li, M.-P. N. Bui, X.-H. Pham, B. S. Kim, Y. H. Choa and G. H. Seong, Sens. Actuators, B, 2012, 174, 406–413.
42. H. Liu, S. Li, D. Sun, Y. Chen, Y. Zhou and T. Lu, J. Mater. Chem. B, 2014, 2, 2212–2219.
43. S. M. Khor, G. Liu, C. Fairman, S. G. Iyengar and J. J. Gooding, Biosens. Bioelectron., 2011, 26, 2038–2044.
44. Y. Zhu, P. Chandra, K.-M. Song, C. Ban and Y.-B. Shim, Biosens. Bioelectron., 2012, 36, 29–34.
45. C. Malitesta, E. Mazzotta, R. A. Picca, A. Poma, I. Chianella and S. A. Piletsky, Anal. Bioanal. Chem., 2012, 402, 1827–1846.
46. A. Turco, S. Corvaglia and E. Mazzotta, Biosens. Bioelectron., 2015, 63, 240–247.
47. E. Mazzotta, C. Malitesta, M. Díaz-Álvarez and A. Martin-Esteban, Thin Solid Films, 2012, 520, 1938–1943.
48. E. Mazzotta and C. Malitesta, Sens. Actuators, B, 2010, 148, 186–194.
49. S. Reisberg, D. F. Acevedo, A. Korovitch, B. Piro, V. Noel, I. Buchet, L. D. Tran, C. A. Barbero and M. C. Pham, Talanta, 2010, 80, 1318–1325.
50. M. A. Booth, J. Leveneur, A. S. Costa, J. Kennedy and J. Travas-Sejdic, J. Phys. Chem. C, 2012, 116, 8236–8242.
51. Y. Li, Q. Lu, S. Wu, L. Wang and X. Shi, Biosens. Bioelectron., 2013, 41, 576–581.
52. X. Chen, H. Pan, H. Liu and M. Du, Electrochim. Acta, 2010, 56, 636–643.
53. J. Radnik, C. Mohr and P. Claus, Phys. Chem. Chem. Phys., 2002, 5, 172–177.
54. M. R. Guascito, D. Chirizzi, R. A. Picca, E. Mazzotta and C. Malitesta, Mater. Sci. Eng., C, 2011, 31, 606–611.
55. G.-T. Liu, H.-F. Chen, G.-M. Lin, P. Ye, X.-P. Wang, Y.-Z. Jiao, X.-Y. Guo, Y. Wen and H.-F. Yang, Biosens. Bioelectron., 2014, 56, 26–32.
56. P. Wang, M. Du, M. Zhang, H. Zhu, S. Bao, M. Zou and T. Yang, Chem. Eng. J., 2014, 248, 307–314.
57. E. J. Park, J.-Y. Lee, J. H. Kim, C. J. Lee, H. S. Kim and N. K. Min, Talanta, 2010, 82, 904–911.
58. E. J. Park, J.-H. Jin, J. H. Kim and N. K. Min, Microchim. Acta, 2011, 174, 231–238.
59. L. He, Y. Zhang, S. Liu, S. Fang and Z. Zhang, Microchim. Acta, 2014, 181, 1981–1989.
60. M. Ma, X. Tu, G. Zhan, C. Li and S. Zhang, Microchim. Acta, 2014, 181, 565–572.
61. M. Lin, M. S. Cho, W. S. Choe and Y. Lee, Biosens. Bioelectron., 2009, 25, 28–33.
62. W. Yang, J. Justin Gooding and D. Brynn Hibbert, J. Electroanal. Chem., 2001, 516, 10–16.
63. Z. Wang, C. Sun, G. Vegesna, H. Liu, Y. Liu, J. Li and X. Zeng, Biosens. Bioelectron., 2013, 46, 183–189.
64. M. A. K. Khan, Y.-T. Long, G. Schatte and H.-B. Kraatz, Anal. Chem., 2007, 79, 2877–2884.
65. S. Yuan, D. Peng, X. Hu and J. Gong, Anal. Chim. Acta, 2013, 785, 34–42.
66. J. Xie, S. Wang, L. Aryasomayajula and V. K. Varadan, Nanotechnology, 2007, 18, 1–9.
67. M.-Y. Hua, H.-C. Chen, R.-Y. Tsai and C.-S. Lai, Talanta, 2011, 85, 631–637.
68. V. Dharuman, B.-Y. Chang, S.-M. Park and J. H. Hahn, Biosens. Bioelectron., 2010, 25, 2129–2134.
69. S. Martić, M. Labib and H.-B. Kraatz, Analyst, 2011, 136, 107–112.
70. M. K. Patel, M. A. Ali, S. Srivastava, V. V. Agrawal, S. G. Ansari and B. D. Malhotra, Biosens. Bioelectron., 2013, 50, 406–413.
71. M. R. Guascito, D. Chirizzi, C. Malitesta, M. Siciliano, T. Siciliano and A. Tepore, Electrochem. Commun., 2012, 22, 45–48.
72. M. R. Guascito, D. Chirizzi, C. Malitesta, T. Siciliano and A. Tepore, Talanta, 2013, 115, 863–869.
73. J. A. Ricketts and L. W. Tresselt, J. Am. Chem. Soc., 1959, 81, 2305–2307.
74. M. R. Guascito, D. Chirizzi, C. Malitesta, E. Mazzotta, M. Siciliano, T. Siciliano, A. Tepore and A. Turco, Biosens. Bioelectron., 2011, 26, 3562–3569.
75. W. Zhou, L. Kibler and D. Kolb, Electrochim. Acta, 2002, 47, 4501–4510.
76. S. Suzer, Anal. Methods, 2012, 4, 3527.
77. S. Axnanda, E. J. Crumlin, B. Mao, S. Rani, R. Chang, P. G. Karlsson, M. O. M. Edwards, M. Lundqvist, R. Moberg, P. Ross, Z. Hussain and Z. Liu, Sci. Rep., 2015, 5, 9788.
78. J. Baltrusaitis, B. Mendoza-Sanchez, V. Fernandez, R. Veenstra, N. Dukstiene, A. Roberts and N. Fairley, Appl. Surf. Sci., 2015, 326, 151–161.
79. C. S. Fadley, J. Electron Spectrosc. Relat. Phenom., 2010, 178–179, 2–32.
80. S. Yamamoto, H. Bluhm, K. Andersson, G. Ketteler, H. Ogasawara, M. Salmeron and A. Nilsson, J. Phys.: Condens. Matter, 2008, 20, 184025.
81. J. M. Conny and C. J. Powell, Surf. Interface Anal., 2000, 29, 444–459.
82. J. M. Conny, C. J. Powell and L. A. Currie, Surf. Interface Anal., 1998, 26, 939–956.

---