This article was originally published in a journal published by Elsevier, and the attached copy is provided by Elsevier for the author’s benefit and for the benefit of the author’s institution, for non-commercial research and educational use including without limitation use in instruction at your institution, sending it to specific colleagues that you know, and providing a copy to your institution’s administrator. All other uses, reproduction and distribution, including without limitation commercial reprints, selling or licensing copies or access, or posting on open internet sites, your personal or institution’s website or repository, are prohibited. For exceptions, permission may be sought for such use through Elsevier’s permissions site at: http://www.elsevier.com/locate/permissionusematerial
Journal of
Electroanalytical Chemistry Journal of Electroanalytical Chemistry 603 (2007) 27–34 www.elsevier.com/locate/jelechem
py
Surface-enhanced Raman spectra of magnetic nanoparticles adsorbed on a silver electrode a
co
Guilherme V.M. Jacintho a, Paola Corio b, Joel C. Rubim
a,*
Laborato´rio de Materiais e Combustı´veis (LMC), Instituto de Quı´mica da Universidade de Brası´lia, CP 04478, 70919-970 Brası´lia, DF, Brazil b Instituto de Quı´mica da Universidade de Sa˜o Paulo, CP 26.077, 05513-970 Sa˜o Paulo, SP, Brazil Received 21 June 2006; received in revised form 30 January 2007; accepted 5 February 2007 Available online 22 February 2007
al
This work is dedicated to the 80th anniversary of Professor Oswaldo Sala
on
Abstract
pe
rs
Maghemite (c-Fe2O3) nanoparticles dispersed in water and forming a magnetic fluid (MF) were characterized by Raman and FT-NIR spectroscopy and were also investigated by surface-enhanced Raman spectroscopy (SERS) and cyclic voltammetry (CV). The FT-NIR results have shown that magnetite (Fe3O4) is not present on the particle as proposed by other investigators. The SER spectra and CVs results have shown that the MF nanoparticles adsorbed on the Ag electrode undergo four reduction processes. At 0.14 V, a positively charged non-stoichiometric Fe(III) oxy-hydroxide surface layer is reduced to Fe3O4. At 0.62 and 0.88 V, d-FeOOH and c-Fe2O3, respectively, are reduced to Fe3O4. Finally, at potentials more negative than 1.1 V, Fe3O4 formed in the previous processes is reduced to Fe0. In the reverse voltammetric scan the SERS spectra are quite different showing that the process is irreversible, since during oxidation of the film Fe(OH)2 and Fe(OH)3 have also been characterized. The SERS and electrochemical results lead us to conclude that the nanoparticles are composed of a c-Fe2O3 (maghemite) nuclei that present a surface layer containing d-FeOOH and a positively charged non-stoichiometric Fe(III) oxy-hydroxide surface layer. 2007 Elsevier B.V. All rights reserved.
r's
Keywords: SERS; Magnetic nanoparticles; Magnetic fluid; Silver electrode; Ferrofluid
1. Introduction
Au
th o
Magnetic iron oxide particles forming relatively stable colloids were first obtained by Elmore [1]. In our days, water suspensions of magnetic particles, e.g. magnetite (Fe3O4), having diameters in the nanometric scale are known as magnetic fluids (MFs) or ferrofluids (see Ref. [2] and references therein). In the last decade MFs have been used in several fields, especially in medicine and biosciences [3–8]. To design magnetic target carriers based on MFs it is very important to know the chemical composition of the iron oxide particles as well as their surface structure and composition.
*
Corresponding author. E-mail address:
[email protected] (J.C. Rubim).
0022-0728/$ - see front matter 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jelechem.2007.02.019
Several groups have tried to use Raman spectroscopy to investigate the chemical structure of magnetic nanoparticles and the majority of the work was done using water based magnetic fluids [9–20]. However, Raman scattering is a very weak process, with molecular cross sections ranging from 1029 to 1032 cm2 [21]. Therefore, the Raman spectra of magnetic nanoparticles give information on the chemical species present in the bulk of the nanoparticle. Raman signals from species at the nanoparticle surface are quite difficult to be obtained. The majority of the works with this goal was done using water based MF and focused the attention on the OH stretching and bending vibrations [9,10,13–15,19]. The main problem in this approach is the solvent (water) that is at a much higher concentration than the species present in the nanoparticle surface. The first report on the characterization of a chemical species (different from OH) adsorbed on the surface of water based
G.V.M. Jacintho et al. / Journal of Electroanalytical Chemistry 603 (2007) 27–34
al
co
py
ecule [26,27]. SERS is also useful to monitor the electrochemical oxidation or reduction of species at electrode/ electrolyte interfaces [28–30]. In this sense, we propose the use of electrochemical SERS, i.e., SERS on a silver electrode, to investigate the chemical composition of maghemite nanoparticles. The strategy in this case will be to record SERS spectra at different potentials applied to a SERS-active silver electrode to characterize the different chemical species present on the maghemite nanoparticle. We believe that through the SERS effect it will be possible to obtain information on the species present at the maghemite surface. This study will be conducted in 0.1 mol L1 KCl solution. Differently from maghemite, magnetite absorbs light in the near infrared (NIR) region at ca. 1400 nm [31,32]. Inaba et al. [31] have observed that a Fe3O4 magnetic fluid present an absorption near 1400 nm and assigned this absorption to the excitation of the 3d6 ! 3d54s1 orbital transition process of Fe2+ in the Fe3O4 particle while others [32] have assigned this absorption to an intervalence (Fe2+ ! Fe3+) charge transfer process. Therefore, this NIR optical absorption can be used to distinguish among magnetite and maghemite as it was already done by Tang et al. [32]. As further supports to the discussion regarding to the composition of maghemite nanoparticles we will also make use of FT-NIR spectroscopy. This experiment will help to answer the following question: do maghemite particles prepared according to the method used in the present work contain magnetite in their composition?
Au
th o
r's
pe
rs
maghemite nanoparticles was done by Souza et al. [12]. They have observed the CH stretching mode of maghemite particles modified by aspartic and glutamic acids. One way to get the Raman spectra of species at low concentration is through the use of the resonance Raman effect that can led to an enhancement of Raman intensities of 3–6 orders of magnitude when the laser excitation matches the energy of an allowed electronic transition of the target molecule. In this sense, the adsorption of methylene blue [12] and tetrasulfonated zinc phthalocyanine [17] on maghemite surfaces was characterized by resonance Raman scattering. Using normal Raman scattering, Souza et al. [11] have proposed that the surface of the maghemite nanoparticles contains Fe(III) hydroxides or oxy-hydroxides. The feature near 720 cm1 was obtained by spectral subtraction and is attributed to the presence of such species. Chourpa et al. [20] have obtained Raman spectra of different maghemite samples and claimed that none of their samples could be representative of a pure maghemite sample, since they found experimental evidences for the presence of maghemite and magnetite in all investigated samples. Their conclusion was based on the fact that the intensity ratio between the features near 662 and 713 cm1 varied from sample to sample. They did not considered that the changes in the 713/662 intensity ratio could be a result of changes in the surface concentration of Fe(III) hydroxides from sample to sample. Further, these authors claimed that they have not found any experimental evidence for the presence of Fe(III) hydroxides or oxy-hydroxides. Dubois et al. [22] have investigated the voltammetric behavior of maghemite nanoparticles dispersed in a pH 13 solution of tetramethylamonium hydroxide using a Hg working electrode. These authors have observed only two reduction waves in the first cathodic scan, at 1.45 and 1.8 V (vs SCE). They have assigned these reduction waves to the reduction of maghemite (c-Fe2O3) to magnetite (Fe3O4) and to the reduction of Fe3O4 to Fe0, respectively. According to these results only two species are expected to be present on the maghemite structure. However, their results do not permit to conclude that the maghemite nanoparticles they used were pure maghemite or a mixture of maghemite and magnetite as proposed by Chourpa et al. [20]. The different achievements of Souza et al. [11] and Choupra et al. [20] regarding to the presence of Fe(III) oxy-hydroxides (or hydroxides) on the maghemite surface as well as the above electrochemical results of Dubois et al. [22] strongly supports the idea that the investigation of the maghemite nanoparticles surface composition still deserves further experimental work. It is well known that surface-enhanced Raman scattering (SERS) is a powerful tool in the characterization of chemical species adsorbed on SERS-active substrates (mainly Ag, Cu, and Au, for reviews see Refs. [21,23– 26]). SERS combined to resonance Raman effect can provide enhancement factors of 1014 making possible the detection of molecules at trace levels or even a single mol-
on
28
2. Experimental Iron oxide nanoparticles were synthesized by the coprecipitation method as reported in the literature [33]. An acid solution of FeSO4 Æ 7H2O (0.09 mol L1) and FeCl3 Æ 6H2O (0.18 mol L1) salts in deaerated distilled water was quickly added to a basic solution of NaOH 1.5 mol L1 at 60 C and under vigorous stirring. The black precipitate (magnetite) formed was washed several times until the pH 8 was reached. The precipitate containing the nanoparaticles was then submitted to a surface treatment as described in the literature [34]: a 2 mol L1 HNO3 solution was added to the precipitate and kept at rest overnight (12 h). The supernatant liquid was removed and the precipitate was kept boiling for half an hour in a 0.5 mol L1 Fe(NO3)3 solution. Then the precipitate was washed twice with acetone and the pH adjusted around 7 (point of zero charge, PZC) with a 0.5 mol L1 (CH3)4NOH solution. At the PZC particles were coagulated and the resulting precipitate was washed several times to remove the excess of ions. Finally, the pH was adjusted to 2 with HNO3 and a stable aqueous maghemite magnetic fluid was obtained. The total iron concentration in the MF is 0.5 mol L1. This MF stock solution was used in all experiments. The Raman and SERS spectra were acquired on a Renishaw Raman System 3000 equipped with an Olympus microscope (BTH2) with a 50· objective to focus the laser
G.V.M. Jacintho et al. / Journal of Electroanalytical Chemistry 603 (2007) 27–34
on
al
co
py
190 cm1 was also observed in the Raman spectrum of polycrystalline maghemite samples as it can be seen in Fig. 1b of Ref. [35]. As pointed by Chourpa et al. [20] samples of maghemite prepared according to the procedure we have used should be a mixture of magnetite and maghemite. Since maghemite (c-Fe2O3) do not present optical absorptions in the NIR region we have measured the FT-NIR spectrum of a film of the maghemite solution in NaCl windows. For comparison purposes we also display the FT-NIR spectrum of a magnetite MF that consists of magnetite nanoparticles modified by the adsorption of oleic acid and then dispersed in cyclohexane. These spectra are presented in Fig. 2a and b, respectively. Note that, as expected, the NIR spectrum of the modified magnetite MF presents the broad absorption feature at 1410 nm (the three features near 1720 nm are due to the first overtones of CH stretching modes). However, the spectrum of the maghemite MF does not show any defined optical absorption in this region that could characterize the presence of magnetite in its composition. Therefore, we believe that the maghemite nanoparticles we are investigating do not present magnetite in their composition at least in a concentration that could be detected by FT-NIR spectroscopy. This also means that the HNO3/Fe(NO3) treatment of magnetite nanoparticles as described in Section 2 has lead to the total oxidation of magnetite to maghemite. The SER spectra of maghemite adsorbed on the silver electrode for different applied potentials are displayed in Fig. 3a. In this case, the Ag electrode was activated by performing ten oxidation reduction cycles (ORCs) at 50 mV/s in the 0.6 to 0.15 V in 0.1 mol L1 KCl solution. At open circuit, a drop of the MF stock solution was added to the SERS-cell and the potential of 0.2 V was applied to the cell. Then, SER spectra were recorded, from 0.2 to 1.2 V, in steps of 0.1 V. The total iron concentration in
pe
rs
beam on the sample. The spectra were excited by the 632.8 nm line from an air cooled He/Ne laser (Spectra Physics). The spectral resolution was 6 cm1. We have used three different laser powers measured just before the sample: (i) in the Raman spectrum of maghemite as a powder we have used 0.5 mW to avoid sample decomposition [35]; (ii) for the maghemite MF (solution) in a sealed capillary the laser power was adjusted to 5 mW – increasing or decreasing the laser power does not cause changes in the Raman spectrum; (iii) in the SERS measurements we have used 5 mW – in this case, an increase in the laser power causes changes in the spectra. The SERS substrate (working electrode) was made from a silver rod (99.999%) with 0.2 cm2 of geometrical area inserted in polytetrafluorethylene. All the potentials are referred to an Ag/AgCl reference electrode. The electrochemical system used in the SERS measurements was a PAR 263 potentiostat/galvanostat from EG&G and the one used in the cyclic voltammetry (CV) experiments was a MQPG-01 from Microquimica. In all experiments the temperature was kept constant at 22 ± 1 C. The FT-NIR spectra were obtained on a Equinox 55 spectrometer from Bruker equipped with an optical fiber accessory coupled to a Ge detector. The nominal resolution was set to 8 cm1 and each spectrum is the average of 32 interferograms. The transmission electron microscopic (TEM) images of the nanomagnetic particles were obtained on a Jeol JEM 1011 transmission electron microscope.
29
3. Results and discussion
Au
th o
r's
The Raman spectra of the maghemite MF in the solid (powder) and solution phases are presented in Fig. 1. The features at 1048 and 1624 cm1 are due to NO 3 and water, respectively, while all the other Raman features are characteristic of maghemite [11,35]. The weak feature at ca.
Fig. 1. Raman spectra of c-Fe2O3 MF in the solid (a) and in solution (b) phases.
Fig. 2. FT-NIR spectra of: (a) an aqueous film of c-Fe2O3 MF and (b) a film of Fe3O4 modified by oleic acid dispersed in cyclo-hexane. The water contribution in spectrum (a) was removed by subtracting the FT-NIR spectrum of a water film.
G.V.M. Jacintho et al. / Journal of Electroanalytical Chemistry 603 (2007) 27–34
co
py
30
Fig. 3. (a) SER spectra of c-Fe2O3 MF adsorbed on a silver electrode at the indicated potentials; (b) SER spectra of a MF film formed on the Ag electrode surface (see text for details).
on
al
The results of Fig. 3a also show that at more negative potentials, the features characteristic of maghemite decrease in intensity and the half width at half maximum (fwhm) value of the broad signal centered at ca. 700 cm1 becomes smaller. Note also that the shoulder near 712 cm1 (at 0.4 V) has practically disappeared at 0.9 V. At this potential the main observed Raman feature is centered at 671 cm1 with a fwhm value of 60 cm1, compatible with the value expected for magnetite [35–37]. There are some Raman spectra of magnetite reported in the literature that present a feature near 180 cm1 [35–37]. However, as mentioned above, a feature near 190 cm1 is also observed in the Raman spectrum of maghemite (see Fig. 1 above and Refs. [11,20,35]). Therefore, we believe that the weak feature observed near 180 cm1 is due to maghemite. As demonstrated above (Fig. 2), the maghemite sample used in these experiments does not contain magnetite in its structure. Due to maghemite particle flocculation, we believe that maghemite was depositing on the Ag surface causing a deterioration of the signal-to-noise ratio. This hypothesis will be strengthened in the next experiment. To avoid flocculation during the experiment, we have adopted a different procedure. First, the Ag electrode was activated by performing ORCs in a 0.1 mol L1 KCl aqueous solution. Then the electrode was removed from the SERS-cell and the Ag surface was exposed to a drop of the maghemite MF solution for 60 s. After that, the electrode was thoroughly washed with double distilled water and transferred back to the SERS-cell. Afterward, SERS spectra were recorded from 0.0 to 1.2 V in steps of 0.1 V. These SERS spectra are displayed in Fig. 3b. The broad SERS features near 365 and 500 cm1 observed at the potential of 0.0 V in Fig. 3b are characteristic of maghemite. However, the two features near 700 cm1 are better defined than in the maghemite spectra of Fig. 1. At 0.0 V, the SERS feature at 712 cm1 is separated from the other broad feature that appears at 662 cm1. At 0.3 V the intensity at 712 cm1 has
Au
th o
r's
pe
rs
the SERS-cell was ca. 0.05 mol L1. It is worth to mention that at the ionic strength corresponding to 0.1 mol L1 KCl part of the particles have coagulated. However, some particles remained in solution since it becomes pale yellow. Fig. 4 displays a TEM image of the working solution containing the maghemite particles in the KCl solution. This micrograph shows that maghemite particles with diameters in the range of 10–20 nm are still present in the KCl solution. The SERS spectrum recorded at 0.2 V shows a Raman feature at 230 cm1 that is not observed in the Raman spectrum of maghemite. Since the measurements are performed in KCl 0.1 mol L1, we believe that this feature is due to the Ag–Cl stretching mode, characterizing the presence of adsorbed chloride. Indeed, this feature disappears as the potential is made more negative.
Fig. 4. TEM images of the maghemite nanoparticles dispersed in 0.1 mol L1 KCl solution.
G.V.M. Jacintho et al. / Journal of Electroanalytical Chemistry 603 (2007) 27–34
on
al
co
py
spectrum of a Pt electrode in KCl solution at open circuit condition. After that, the electrode was removed from the solution and exposed to a drop of maghemite solution for 60 s. Then, the Pt electrode was thoroughly washed with double distilled water and transferred back to the SERS-cell and another Raman spectrum was recorded. In this case we have just observed an increase in the background intensity but no Raman signal from maghemite (see supporting information). Therefore, the spectra displayed in Fig. 3b are surface-enhanced. In order to understand the changes in intensity with the applied potential and their relation to the different chemical species composing the maghemite nanoparticle we have recorded the cyclo-voltammograms (CVs) of a silver electrode exposed to a drop of the MF solution as in the experiment of Fig. 3b. These CVs are presented in Fig. 5a. The first potential scan in the direction of negative potentials shows a first reduction wave near 0.14 V, a second wave at ca. 0.88 V (IIc), and a third wave near 1.29 V (IIIc). In the reverse scan, three oxidation waves are observed at 0.93, 0.72, and 0.43 V, peaks Ia, IIa, and IIIa, respectively. Interesting to note is that in the second scan a new reduction wave is observed at 0.62 V (Ic), that was not visible in the first scan. According to Dubois et al. [22], maghemite is reduced over a Hg electrode (at pH 13) showing two reduction waves in the first cathodic scan, at 1.45 and 1.8 V (vs SCE) that were assigned to the reduction of maghemite (c-Fe2O3) to magnetite (Fe3O4) and to the reduction of Fe3O4 to Fe0, respectively. These authors have also observed that by performing successive oxidation-reduction cycles the reduction wave observed at 1.45 V shifted to 1.4 V and a new reduction wave appeared at 1.15 V (vs SCE). Considering that in our case the pH of the working solution is around 6, our observations are consistent with those of Dubois et al. [22], except for the first reduction wave at 0.14 V. As the maghemite film is very thin, the current waves observed in the CVs of Fig. 5a are very weak. Therefore we have performed another CV experiment in which a
Au
th o
r's
pe
rs
decreased and the separation between the features above and below 700 cm1 is no longer defined. For potentials more negative than 0.5 V the broad features characteristic of maghemite appearing below 600 cm1 became almost indistinguishable. It is well known that d-FeOOH presents two broad features centered at 663 and 400 cm1 with almost the same intensity [35,38]. We believe that the presence of d-FeOOH in the maghemite surface is responsible for the fact that the broad features of maghemite below 600 cm1 are not well defined in Fig. 3b. For potentials more negative, e.g. 0.9 V, the shoulder near 700 cm1 is no longer observed and a feature centered at 671 cm1 presents a fwhm value of 59 cm1. These values are compatible with those expected for Fe3O4. This result suggests that maghemite and d-FeOOH were reduced to magnetite. At 1.2 V the intensity of the 671 cm1 signal decreases considerably indicating that magnetite is reduced at this potential. It is worth to mention that the SERS intensity of the magnetite feature (at 671 cm1) at 0.9 V in Fig. 3b is almost three times as intense as in Fig. 3a. This result supports our conclusion that in the first SERS experiment, Fig. 3a, the nanoparticles were depositing on the electrode surface. The highest SERS enhancement factors are expected for the species just in contact (first layer) with the silver SERS-active sites [23,26]. If a thick film is covering the species directly in contact to the Ag surface, less exciting photons achieve the SERS-active site, causing a deterioration of the SERS signal from the first layer. This kind of behavior was already observed in the SERS study of brass passivation by benzotriazole [39]. In that experiment the passive film growing process was monitored with two different laser radiations, at 514.5 and 647.1 nm. The thicker was the passive film the more intense was the passive film normal Raman signals at 514.5 nm excitation, while the SERS intensity of the first layer of the passive film decreased at 674.1 nm excitation. One could argue, as did one of the reviewers, that in the case of Fig. 3b we are not observing the SERS spectra rather the normal Raman spectra of a maghemite thick film. To verify this hypothesis we have recorded the Raman
31
Fig. 5. Cycle voltammograms of: (a) a silver electrode that was exposed to a drop of the MF solution for 60 s and then transferred back to the electrochemical cell containing KCl 0.1 mol L1. Scan rate = 50 mV/s. (1) and (2) indicates the first and second cycle; and (b) a gold electrode in 0.1 mol L1 KCl containing maghemite MF with a total iron concentration of 0.01 mol L1.
32
G.V.M. Jacintho et al. / Journal of Electroanalytical Chemistry 603 (2007) 27–34
1.45 V and pH 13. At 0.5 mol L1 HCl (pH 0), maghemite microparticles reduce at 0.51 V (vs Ag/AgCl) [44]. Further, it was also observed that the reduction potential of the Fe(III) oxide microparticles follows the linear relationship: 0.5 to 0.05 pH and that it also depends on the complexing capability of the supporting electrolyte anion [44]. Therefore, the differences in the reduction potentials of maghemite in relation to those presented in the literature [22,44] are due to differences in pH and type of supporting electrolyte. The reduction waves at 1.38 and 1.29 V for Au and Ag, respectively, corresponds to the reduction of magnetite to Fe0 according to the following reaction:
py
Au electrode was used instead of Ag, and maghemite nanoparticles were present in solution. Fig. 5b shows the CVs of a gold electrode in 0.1 mol L1 KCl solution prior and after the addition of maghemite MF to the working solution. The first reduction wave at 0.12 V is similar to the one observed for the silver electrode at 0.14 V. Therefore, we believe that at the silver electrode this reduction wave has two sources: (i) the reduction of an Ag–Cl layer and (ii) the reduction of a Fe(III) species to an Fe(II) containing species. We suggest that this first reduction process is due to the following reaction:
co
3FeOx (OH)32x (x < 1) + Hþ + e ! Fe3 O4 + (53x)H2 O ð1Þ
ð3Þ
on
al
This reaction was observed by Dubois et al. [22] at 1.8 V. The reduction of magnetite at these potentials is evidenced by the decrease in the intensity of the 671 cm1 (fwhm = 56 cm1) SERS feature at the potential of 1.2 V. As mentioned above, the cathodic peak Ic, observed in Fig. 5a and b at 0.62 and 0.68 V, respectively, was also observed by Dubois et al. [22] at 1.15 V. However, this cathodic wave is not apparent in the first scan of Fig. 5a nor in the first scan in the experiment of Dubois et al. [22] with the Hg electrode. However, this reduction peak is well defined in the CV of Au in the first scan. We tentatively assign this cathodic current peak to the reduction of another oxy-hydroxide species present on the nanoparticle surface, probably the d-FeOOH, according to the reaction:
Au
th o
r's
pe
rs
The non-stoichiometric oxy-hydroxide is then characterized by the SERS feature at 712 cm1. We do not believe that this species is reduced to Fe(OH)2 since the SERS spectrum at 0.3 does not show the broad Raman feature characteristic of Fe(OH)2 at ca. 535 cm1 [40–43]. Further, when we fit this spectrum to a Lorentzian function, we find among others a Raman signal at 671 cm1 with a fwhm of ca. 60 cm1 that is characteristic of magnetite (Fe3O4). It is worth to mention that Cepria´ et al. [44] have investigated the electrochemical behavior of maghemite microparticles immobilized in paraffin covered pyrolytic graphite electrodes. Their voltammetric results were obtained in 0.5 mol L1 HCl solution and show reduction processes at ca.0.03 and 0.51 V vs Ag/AgCl/KClsat. Cepria´ et al. [44] have interpreted theses processes as due to the reduction of a Fe(III) species do Fe(II) and to the reduction of c-Fe2O3 to FeO, respectively. We agree that the first process does involve an Fe(III) species as discussed above. However, we have not found evidences for the formation of FeO that presents a Raman signal at 652 cm1 [35]. One could argue that this reduction process could be due to Fe(III) hydrated chloride species. As observed by Domenech et al. [45] such species present a well-defined reduction peak at +0.25 V for a glassy carbon electrode in a 5 · 104 mol L1 solution of FeCl3 Æ 6 H2O in 0.01 mol L1 HCl and 0.10 mol L1 NaCl. Therefore, we can conclude that such species is not present in the electrode surface. Probably, the non-stoichiometric oxy-hydroxide is protonated, i.e., positively charged, and interacts with the silver surface through the adsorbed chloride, characterized by the SERS feature near 225 cm1. The CV of Fig. 5b shows other three reduction waves, at 0.69, 0.86, and 1.38 V, in a sequence similar to the three cathodic peaks observed in Fig. 5a for the silver electrode. The reduction wave at 0.86 V on Au corresponds to the one observed at 0.88 V for the Ag electrode and the cathodic current at this potential is attributed to the reaction:
Fe3 O4 + 8Hþ + 8e ! 3Fe0 + H2 O
3c-Fe2 O3 + 2Hþ + e ! 2Fe3 O4 + H2 O
ð2Þ
This reduction process is similar to the one observed by Dubois et al. [22] for the Hg electrode at the potential of
3d-FeOOH + Hþ + e ! Fe3 O4 + 2H2 O
ð4Þ
In order to understand the origin of this reduction process we have investigated the SERS spectra during the anodic scan of the maghemite film experiment corresponding to the Ag CV displayed in Fig. 5a. In the anodic scan, from 1.2 to 0.0 V, we have observed three oxidation waves and the corresponding SERS spectra are displayed in Fig. 6a. Note that at 0.8 V the feature at ca. 669 cm1 broadens and a broad feature in the 400–550 cm1 range emerges. At less negative potentials the peaks become more defined and at 0.4 V we see a feature centered near 664 cm1 (fwhm = 100 cm1) and another broad feature at 534 cm1. At these potentials the Fe0 formed upon reduction of the adsorbed magnetic particles was oxidized to Fe(OH)2 and d-FeOOH, characterized by the features at 534 and 664 cm1, respectively [38–41]. The presence of magnetite can not be disregarded since the following reaction can occur [46]: Fe(OH)2 + 2d-FeOOH ! Fe3 O4 + 2H2 O
ð5Þ
As the potential is made less negative (e.g. 0.0 V) the 664 cm1 feature increases in intensity in relation to the feature at 534 cm1, indicating that the Fe(OH)2 is further oxidized to d-FeOOH. After the spectrum at 0.0 V of Fig. 6a was recorded the electrode potential was scanned from 0.0 to 0.15 V and back to 0.2 V at 50 mV/s. Then, SERS spectra were
33
co
py
G.V.M. Jacintho et al. / Journal of Electroanalytical Chemistry 603 (2007) 27–34
Au
th o
r's
on
pe
rs
recorded at different potentials, from 0.2 to 1.2 V. These spectra are displayed in Fig. 6b. The feature at ca. 224 cm1 is due to the Ag–Cl vibration. The broad feature at 534 cm1 is due to Fe(OH)2 and the signal at 664 cm1 is assigned to d-FeOOH. The shoulder near 705 cm1 can not be assigned to the non-stoichiometric oxi-hydroxide, since the reduction wave characteristic of this species, at 0.14 V, is not observed in the second scan of the CV in Fig. 5a. The SERS spectrum of Fig. 6b recorded at 0.2 V was submitted to a curve fitting procedure (see Supporting information, Fig. 1S). The fitted spectrum shows a feature at 705 cm1 that is assigned to Fe(OH)3, two broad features at 663 and 420 cm1 that are assigned to dFeOOH, and another broad feature at 534 cm1 that is due to Fe(OH)2. As the potential is made more negative, Fe(OH)3 and d-FeOOH are reduced to Fe(OH)2 and Fe3O4 [38]. Note that at the potential of 0.6 V the Fe(OH)2 characteristic Raman signal at 534 cm1 presents its highest intensity. Therefore, we attribute the cathodic peak Ic, at 0.62 V to the reduction of d-FeOOH. Since Fe(OH)3 has a broad Raman signal near 700 cm1 [47], this species may also be present. We can not rule out the presence of these species on the maghemite nanoparticles surface since this cathodic wave was also observed in the CV of Au with maghemite MF in solution as in Fig. 5b. This cathodic wave is not seen in the first scan of the Ag CV of Fig. 5a due to the very low surface concentration of dFeOOH. This can also be the reason why Dubois et al. [22] have also not seen this cathodic wave in their experiment at the first scan. The results of Fig. 6b also show that as the potential is made more negative, the Fe(OH)2 peak disappears remaining a Raman feature at ca. 669 cm1 with a fwhm of ca. 60 cm1, characterizing the formation of Fe3O4 that can be formed as a reduction product of Fe(III) hydroxides or oxy-hydroxide and due to reaction (5). It is worth to mention that the SERS spectra of the film at potentials less negative than 0.4 V (Fig. 6a and b) are
al
Fig. 6. SER spectra of maghemite nanoparticles adsorbed on a silver electrode at the indicated potentials. The SERS spectra were recorded: (a) in the positive scan of electrode potentials and (b) in the negative scan, after a previous SERS activation procedure corresponding to varying the applied potential up to 0.15 V (silver oxidation) and then returning to 0.2 V.
Fig. 7. Schematic picture of the maghemite particle adsorbed at the Ag electrode surface.
quite different from those observed in Fig. 3b at the same potentials. These results clearly show that the reduction process of nanoparticles adsorbed on the silver surface forming the film is not reversible. At the beginning, i.e., before reduction, the film consists of nanoparticles composed of c-Fe2O3 and a surface layer composed of dFeOOH, probably Fe(OH)3, and a non-stoichiometric oxy-hydroxide positively charged as depicted in the schematic picture of Fig. 7. At 0.14 V the non-stoichiometric oxy-hydroxide is reduced to magnetite. For potentials more negative than 0.2 V, d-FeOOH and Fe(OH)3 are reduced to magnetite. Finally, for potentials more negative than 1.2 V the magnetite film is reduced to Fe0 and its further oxidation leads to the formation of Fe3O4, dFeOOH, Fe(OH)3 and Fe(OH)2 but no evidence is found for the formation of c-Fe2O3. 4. Conclusion The results presented in this work have clearly shown that the nanoparticles of maghemite based MF prepared according to the method described in Ref. [34] do not contain magnetite (Fe3O4) in their composition. We have also
34
G.V.M. Jacintho et al. / Journal of Electroanalytical Chemistry 603 (2007) 27–34
co
al
rs
Acknowledgements
on
(i) A nuclei of c-Fe2O3, characterized by a broad SERS features near 700 (composed by a peak at 662 and a shoulder at 710 cm1), 500, and 365 cm1. The cFe2O3 is reduced to Fe3O4 near 0.9 V. (ii) A surface layer containing a non-stoichiometric oxy-hydroxide characterized by a SERS feature at 712 cm1, that is reduced to Fe3O4 at 0.14 V. The surface layer also contain d-FeOOH, characterized by two broad (fwhm 100 cm1) features at 663 and 410–420 cm1 and may also contain Fe(OH)3. These latter two species are reduced to Fe3O4 near 0.6 V. The presence of magnetite was detected only upon reduction of the Fe(III) species present on the maghemite nanoparticle. Its characteristic SERS signal was observed at ca. 671 cm1 (fwhm 60 cm1) and its reduction to Fe0 was observed for potentials more negative than 1.0 V.
[9] P.C. Morais, S.W. da Silva, M.A.G. Soler, M.H. Sousa, F.A. Tourinho, J. Magn. Magn. Mater. 201 (1999) 105. [10] M.H. Sousa, F.A. Tourinho, J.C. Rubim, J. Raman Spectrosc. 31 (2000) 185. [11] M.H. Sousa, J.C. Rubim, P.G. Sobrinho, F.A. Tourinho, J. Magn. Magn. Mater. 225 (2001) 67. [12] J.C. Rubim, M.H. Sousa, J.C.O. Silva, F.A. Tourinho, Braz. J. Phys. 31 (2001) 402. [13] P.C. Morais, S.W. da Silva, M.A.G. Soler, N. Buske, Biomol. Eng. 17 (2001) 41. [14] J.E. Weber, A.R. Gon˜i, C. Thomsen, J. Magn. Magn. Mater. 277 (2004) 96. [15] M.C. Bautista, O. Bomati-Miguel, X. Zhao, M.P. Morales, S.W. da Silva, M.A.G. Soler, C. Gansau, N. Buske, P.C. Morais, J. Magn. Magn. Mater. 226–230 (2001) 1890. [16] J.C.O. Silva, M.H. Sousa, F.A. Tourinho, J.C. Rubim, Langmuir 18 (2002) 5511. [17] S.W. da Silva, T.F.O. Melo, M.A.G. Soler, E.C.D. Lima, M.F. da Silva, P.C. Morais, IEEE Trans. Magn. 39 (2003) 2645. [18] M.A.G. Soler, T.F.O. Melo, S.W. da Silva, E.C.D. Lima, A.C.M. Pimenta, V.K. Garg, A.C. Oliveira, P.C. Morais, J. Magn. Magn. Mater. 272–276 (2004) 2357. [19] K.S.K. Varadwaj, M.K. Panigrahi, J. Ghose, J. Sol. State Chem. 177 (2004) 4286. [20] I. Chourpa, L. Douziech-Eyrolles, L. Ngaboni-Okassa, J.F. Fouquenet, S. Cohen-Jonathan, M. Souce, H. Marchais, P. Dubois, Analyst 130 (2005) 1395. [21] G.A. Baker, D.S. Moore, Anal. Bioanal. Chem. 382 (2005) 1751. [22] E. Dubois, J. Chevalet, R. Massart, J. Mol. Liq. 83 (1999) 243. [23] A. Otto, I. Mrozek, H. Grabhorn, W. Akemann, J. Phys. Condens. Matter 4 (1992) 1143. [24] Z.Q. Tian, B. Ren, Annu. Rev. Phys. Chem. 55 (2004) 197. [25] R.F. Aroca, R.A. Alvarez-Puebla, N. Pieczonka, S. Sanchez-Cortez, J.V. Garcia-Ramos, Adv. Coll. Interf. Sci. 116 (2005) 116. [26] K. Kneipp, H. Kneipp, J. Kneipp, Acc. Chem. Res. 39 (2006) 443. [27] F.J.P. Gouveia, I.G.R. Gutz, J.C. Rubim, J. Electroanal. Chem. 371 (1994) 113. [28] J.C. Rubim, J. Electroanal. Chem. 220 (1987) 339. [29] Z.Q. Tian, J. Raman Spectrosc. 36 (2005) 466. [30] S.H.A. Nicolai, P.R.P. Rodrigues, S.M.L. Agostinho, J.C. Rubim, J. Electroanal. Chem. 527 (2002) 103. [31] N. Inaba, H. Miyajuma, H. Takahashi, S. Taketomi, S. Chikazumi, IEEE Trans. Magn. 25 (1989) 3866. [32] J. Tang, M. Myers, K.A. Bosnick, L.E. Brus, J. Phys. Chem. B 107 (2003) 7501. [33] N.M. Gribanov, E.E. Bibik, O.V. Buzunov, V.N. Naumov, J. Magn. Magn. Mater. 85 (1990) 7. [34] F.A. Tourinho, R. Franck, R. Massart, J. Mater. Sci. 25 (1990) 3249. [35] D.L.A. de Faria, S.V. Silva, M.T. de Oliveira, J. Raman Spectrosc. 28 (1997) 873. [36] L.V. Gasparov, D.B. Tanner, D.B. Romero, H. Berger, G. Margaritondo, L. Forro, Phys. Rev. B 62 (2000) 7939. [37] O.N. Shebanova, P. Lazor, J. Sol. State Chem. 174 (2003) 424. [38] N. Boucherit, A.H.-L. Goff, S. Joiret, Corros. Sci. 32 (1991) 497. [39] J.C. Rubim, Chem. Phys. Lett. 167 (1990) 209. [40] J.C. Rubim, J. Du¨nnwald, J. Electroanal. Chem. 258 (1989) 327. [41] I.C.G. Thanos, Electrochim. Acta 31 (1986) 811. [42] J. Gui, T.M. Devine, Corros. Sci. 37 (1995) 1177. [43] L.J. Simpson, C.A. Melendres, Electrochim. Acta 41 (1996) 727. [44] G. Cepria´, A. Uso´n, J. Pe´rez-Arantegui, J.R. Castillo, Anal. Chim. Acta 477 (2003) 157. [45] A. Dome´nech, J. Pe´rez-Ramı´rez, A. Ribera, G. Mul, F. Kapteijn, I.W.C.E. Arends, J. Electroanal. Chem. 519 (2002) 72. [46] M. Stratmann, K. Bohnenkamp, H.J. Engell, Corros. Sci. 23 (1983) 969. [47] J. Du¨nnwald, A. Otto, Corros. Sci. 29 (1989) 1167.
py
shown that the SERS effect on electrodes can be used to characterize the different species present on the maghemite surface. The SERS and electrochemical measurements present evidences that the magnetic nanoparticles investigated in this work are composed by:
pe
The authors thank CNPq, FAPDF, FINATEC, and FAPESP for research grants. G.V.M.J. thanks CAPES and P.C., PAZS, and J.C.R. thank CNPq for research fellowships. Appendix A. Supplementary data
r's
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jelechem. 2007.02.019.
th o
References
Au
[1] W.C. Elmore, Phys. Rev. 54 (1938) 309. [2] F.A. Tourinho, J. Depeyrot, G.J. da Silva, M.C.L. Lara, Braz. J. Phys. 28 (1998) 413. [3] T. Neuberger, B. Schopf, H. Hofmann, M. Hofmann, B. von Rechenberg, J. Magn. Magn. Mater. 293 (2005) 483. [4] A.S. Lubbe, C. Bergemann, H. Riess, F. Schriever, P. Reichardt, K. Possinger, M. Matthias, B. Dorken, F. Herrmann, R. Gurtler, P. Hohenberger, N. Haas, R. Sohr, B. Sander, A.J. Lemke, D. Ohlendorf, W. Huhnt, D. Huhn, Cancer Res. 56 (1996) 4686. [5] M.C. Bautista, O. Bomati-Miguel, X. Zhao, M.P. Morales, T. Gonza´lez-Carren˜o, R. Pe´rez de Alejo, J. Ruiz-Cabello, S. Veintemillas-Verdaguer, Nanotechnology 15 (2004) S154. [6] F.Y. Cheng, C.H. Su, Y.S. Yang, C.S. Yeh, C.Y. Tsai, C.L. Wu, M.T. Wu, D.B. Shieh, Biomaterials 26 (2005) 729. [7] D.K. Kim, Y. Zhang, J. Kehr, T. Klason, B. Bjelke, M. Muhammed, J. Magn. Magn. Mater. 225 (2001) 256. [8] D.C.F. Chan, D.B. Kirpotin, P.A. Bunn, J. Magn. Magn. Mater. 122 (1993) 374.
