J. Phys. Chem. C 2007, 111, 4027-4035
4027
Surface Structure, Hydration, and Cationic Sites of Nanohydroxyapatite: UHR-TEM, IR, and Microgravimetric Studies Luca Bertinetti,†,‡ Anna Tampieri,‡ Elena Landi,‡ Caterina Ducati,§ Paul A. Midgley,§ Salvatore Coluccia,† and Gianmario Martra*,† Dipartimento di Chimica IFM and NIS Center of Excellence, UniVersita` di Torino, Via P. Giuria 7, 10125 Torino, Italy, ISTEC-CNR, Via Granarolo 64, 48018 Faenza, Italy, and Department of Materials Science and Metallurgy, UniVersity of Cambridge, Cambridge CB2 3QZ, United Kingdom ReceiVed: September 15, 2006; In Final Form: December 11, 2006
A multi-technique study devoted to investigate the surface features of nanosized hydroxyapatite (HA) was carried out. UHR-TEM observation provided evidence that HA nanoparticles are constituted by a crystalline core, elongated in the direction of the crystallographic c-axis, coated by an amorphous layer 1-2 nm thick. By means of IR spectroscopy and microgravimetry, the amount of water and hydroxy groups on the surface was evaluated. For the as-prepared material, it was found that the first hydration layer is mainly constituted by H2O molecules interacting through a coordinative bond with Ca2+ in a 1:1 ratio, while hydroxy groups account only for ca. 20% of surface hydration species. Outgassing at increasing temperatures up to 300 °C resulted in a complete surface dehydration, accompanied by a decrease of the capability to readsorb water. Possible changes of the local structure of surface Ca2+ ions were probed by IR spectra of adsorbed CO. The combination of these data with rehydration tests suggested that a significant part of surface Ca2+ ions, once dehydrated, can undergo a relaxation inward the surface, progressively more irreversible as the outgassing temperature increases.
1. Introduction In the last three decades a significant part of the research activity in the field of biomaterials has focused on the investigation of structure-property relationships. This has allowed a finer tuning of biomaterial synthesis aiming at the optimization of their functionality and response elicited in the tissue environment. Such responses can range from the inflammation following the implant of structural prostheses, to the formation of cloths on stents and to the transformation of the implanted materials in the course of the tissue regeneration, as in the case of bioactive and/or bioresorbable materials for bone repairing. Independently of the type of response, the key role of the surface features of biomaterials has been recognized, leading to the definition, at the beginning of 2000s of the concept of “biological/biomedical” surface science.1-3 The unraveling of the ensemble of surface processes and phenomena actually occurring in vivo is still a challenge, but there is a general consensus in setting the causal sequence: (i) biomaterial surface structure, (ii) states of adsorbed water molecules, and (iii) states of adsorbed proteins, as one of the main factors ruling the fate of the interaction of the implant (then actually occurring through a hybrid synthetic/proteic interface) with cells.1-4 The recent development of the ability to prepare nanometric and/or nanostructured materials in a controlled way has encouraged the investigation of the “nano” effect (i.e., enhanced contribution of the surface against the volume) for biomaterials too. In particular, a series of studies carried out by the group of Webster put in evidence improved in vitro performances for * Corresponding autor. E-mail:
[email protected]. † Universita ` di Torino. ‡ ISTEC-CNR. § University of Cambridge.
osteoblast contacted with various types of materials intended for orthopedic/dental applications, i.e., metals,5a polymers,6 carbon fibers,7,8 ceramics,5bcd,9,10 and their composites11,12 in nanometric/nanostructured forms. Among bioceramic, hydroxyapatite-based materials play a quite relevant role, as they can be considered the synthetic version of the mineral part of bone tissue and enamel, and, indeed, the possibility to prepare them in a nanometric/ nanostructured form is considered one of the requisites to fulfill for a “biomimetic” approach to the preparation of optimized materials.13 The investigation of the surface features of hydroxyapatites (HA) has been the object of several scientific investigation since the seminal works by Neuman et al.14 and Dry and Beebe,15 also related to the wide use of HA powders as adsorbent for chromatography (another consequence of the peculiar surface feature of this material), providing insights on surface characteristics such as nature of some surface hydroxyl groups, states of adsorbed water,16 and basicity of oxygen atoms.17 As a common element, these studies essentially considered HA particles outgassed and/or treated at temperatures as high as 300 °C. More recently, a growing interest has been raised by the study of the surface of nanoHA that did not underwent any severe thermal treatment after preparation (by precipitation in aqueous medium), except drying, looking for contributions to the knowledge of both the process of formation of this biomaterial likely occurring in vivo and the surface properties of a synthetic materials closely resembling the natural model. Investigations carried out by solid-state NMR,18,19 IR, and X-ray absorption20,21 spectroscopies put in evidence the presence in such HA of highly hydrated non-apatitic domains, likely located at the surface. In particular, Ja¨ger et al. propose a structural scheme of HA nanoparticles with a disordered, non-apatitic layer
10.1021/jp066040s CCC: $37.00 © 2007 American Chemical Society Published on Web 02/20/2007
4028 J. Phys. Chem. C, Vol. 111, No. 10, 2007 1-2 nm thick surrounding an actual HA bulk, based on the assumption that nanoparticles obtained by precipitation in a liquid medium and consisting of an ordered and a disordered part should have a crystalline core coated by a disordered surface region.21 Our target has been to determine the nature of the complex structure of such particles by an ultrahigh-resolution transmission electron microscopy (UHR-TEM) study, allowing a direct observation of detailed structural features and relative location of different phases in HA nanoparticles. In addition, we combined IR spectroscopy, augmented by the use of H2O, D2O, and CO as probe molecules, and microgravimetry to obtain insights on the nature and amount of hydration species exposed at the surface of the non-apatitic layer (as a first step of a research path aimed to investigate factors affecting the interaction of nanohydroxyapatite with proteins), on physical-chemical properties of the cationic sites below, and how all these features evolve by outgassing at increasing temperature. 2. Experimental 2.1. Materials. Hydroxyapatite powder, HA (BET specific surface area ) 78 m2/g), was synthesized through an aqueous medium procedure, dropping in a Ca(OH)2 suspension a 1.3 M solution of H3PO4, to accomplish the reaction 5Ca(OH)2 + 3H3PO4 f Ca5(PO4)3OH+9H2O. The system was continuously stirred at 40 °C; a ripening time of 2 h was used. The precipitate was then washed, filtered, and dried in air at room temperature. X-ray diffractometric analysis was carried out with a Rigaku Miniflex, using a Cu KR radiation. Thermogravimetric analysis revealed the presence of carbonate species in reason of ca. 2 wt %, likely resulting from the carbonation of the Ca(OH)2 suspension. For IR measurements, high purity CO (Praxair) was employed without any additional purification except liquid nitrogen trapping, while H2O and D2O (99.9 atom % D, Aldrich) were admitted onto the samples after several freeze-pump-thaw cycles. 2.2. Methods. Observations of the powders by transmission electron microscopy (TEM) were performed with a JEOL EX4000 with acceleration potential of 400 kV and a JEOL 3010UHR with acceleration potential of 300 kV. Samples were dispersed on lacey carbon Cu grids. Image calculations for the analysis of the experimental data taken with the JEOL EX4000 instrument were made with the online version of EMS software package22 using the Bloch wave method for different values of crystal thickness and objective defocus. The values of Cs ) 1.06 mm for the spherical aberration coefficient, df ) 9 nm for the spread of focus and sc ) 0.8 mrad for the beam convergence semi-angle were used. FTIR spectra were obtained using a Bruker Vector 22 spectrometer (resolution: 4 cm-1) equipped with DTGS or MCT detector, for ATR and transmission mode, respectively. To monitor the presence of acid phosphate species, the spectra of the material in the ν4 region were collected in the ATR mode (MKII Golden gate-SPECAC equipped with a diamond crystal). Infrared spectra for the analysis of surface features were performed in transmission on HA powders pressed in selfsupporting pellets. The samples were placed in a quartz IR cell equipped with KBr windows designed to carry out spectroscopic measurements both at beam temperature (ca. 50 °C) and low temperature (ca. -170 °C, by cooling with liquid nitrogen). The cell was connected to a conventional vacuum line (residual pressure: 1 × 10-5 mbar, 1 mbar ) 102 Pa) allowing all thermal treatments and adsorption-desorption experiments to be carried
Bertinetti et al.
Figure 1. TEM image representative of the size and morphologly of HA particles. Original magnification: 80k×.
out in situ. The spectra of adsorbed CO are reported in absorbance, after subtraction of the spectra of the samples before adsorption as background. For the microgravimetric experiments, the pellets were transferred into the microbalance inside a quartz reactor. The system used was a Hiden Intelligent Gravimetric Analyzer (IGA). The samples were treated following the same procedures used for IR spectroscopy. 3. Results and Discussion 3.3. Morphological and Structural Investigations by UHRTEM. Figure 1 displays an image representative of the size and morphology of the HA particles. Most of them appear elongated in one direction, with length and width (in the bidimensional projection on the image plane) in the 50-150 nm and 10-70 nm range, respectively, and exhibiting rounded edges. Particles appear to be stacked on each other, mainly along their length, to form agglomerates, while in some cases they are actually joined through grain boundaries, forming larger secondary nanostructured particles. Observations at high magnification in ultrahigh resolution (UHR) conditions provided insight into the structural features of the bulk of the particles, as well as into the morphology and structure of the surface. An example is shown in Figure 2, where in the main panel is the image of a portion of a primary particle protruding from a secondary one. A regular fringe pattern extended along the entire particle can be observed, indicating its monocrystalline nature. By performing the Fourier transform of portions of the image, e.g., of the area labeled as “a,” the corresponding digital diffraction pattern was obtained (displayed in the small panel “FT of a”). By indexing the digital diffraction pattern, the zone axis was derived (in this particular case [201]), and therefore the crystallographic orientation of the particle in the image was determined. Interestingly, it was found that, for elongated particles like that in Figure 2, the main dimension runs parallel to the crystallographic c-axis, as occurs in nonnanometric HA materials.23,24 In addition, portions of the original UHR-images of such particles were simulated by a Bloch wave functions approach (EMS Online), looking for the thickness value allowing the best matching with the experimental contrast. As an example, the result obtained by such a simulation for a sub-region of the area labeled as “a” is reported in the small panel “zoom of a.” The closest similarity between the simulated subregion (central square frame) and the original image was obtained for a slice of ca. 20 nm. On average, a thickness between 15 and 25 nm was found for most particles. Such nanocrystalline nature of the material well agrees with the broadness of the peaks in the X-ray diffraction pattern, from
Nanohydroxyapatite
J. Phys. Chem. C, Vol. 111, No. 10, 2007 4029
Figure 2. High-resolution TEM image of a portion of a HA particle (main panel, left), and elaboration and details of regions a and b (right panels). Original magnification: 800k×.
Figure 3. Ultrahigh-resolution images of portions of the borders of HA particles (original magnification: 800k×).
which a size of the coherent scattering domains of ca. 30 nm was calculated (see Supporting Information, Figure 1S). Finally, observations in UHR conditions allowed to highlight peculiar features related to the surface of the particles: as shown in the zoomed view of region b, the lattice fringes running along the crystal vanish near the surface, replaced by a region 1-2 nm thick exhibiting a contrast typical of an amorphous phase. The same pattern was found along the borders of all particles inspected, and some representative images are reported in Figure 3. In the literature it has been documented that some degree of damage occurs to the hydroxyapatite structure by exposure to the electron beam with a current density of at least 1.5 A/cm2 for several minutes.25-27 In this study the TEM observations were carried out under feeble illumination conditions at all times except for the 1-5 s required to take the image. In no case were changes in shape and thickness of the amorphous surface layer observed by prolonging the time exposure. To the best of our knowledge, this result is the first experimental direct observation of a disordered non-apatitic layer, previously identified via IR, X-ray absorption, and NMR spectroscopy,18,20-21 surrounding a crystalline HA core, as in the model proposed by Ja¨ger and co-workers.19 For the sake of completeness, TEM images were taken also for the material outgassed at 300 °C, the condition corresponding to the final step of the thermal treatment procedure (presented in the following sections) carried out to assess the features of the surface hydration species and cationic sites. It was found that such treatment did not affect significantly neither the overall morphology, nor the amorphous layer covering the surface (see Figure 2S in Supporting Information).
Figure 4. IR spectra of HA: (a) in contact with H2O vapor (23.5 mbar) and then outgassed at beam temperature (ca. 50 °C) for (b) 0.5, (c) 2, (d) 15, and (e) 120 min. Curve f is the spectrum obtained after 4 cycles of H2O/D2O exchange and subsequent outgassing at beam temperature for 120 min. Insets: zoom of the 3755-3600 cm-1 (left) and 17801580 cm-1 ranges; lettering as in the main frame.
3.2. Investigation of the Surface Hydration. 3.2.1. IR and MicrograVimetric Measurements. The second aspect of the investigation dealt with the assessment, possibly quantitative, of hydroxyl groups and water molecules of the first surface hydration layer of the prepared HA. To this aim, the progressive surface dehydration of the material from the equilibrium with the water vapor pressure at room temperature, with H2O likely forming liquid-like multilayers on the surface, to the complete removal of hydroxyl groups and water molecules by outgassing at increasing temperature was monitored by combining IR spectroscopy and microgravimetry. Figure 4 reports the spectra, in the 3750-1350 cm-1 range, recorded during progressive outgassing at room temperature (curves a-e), together with the spectrum obtained at the end of an equivalent outgassing process after contact with D2O (curve f). The region at lower frequency was not observable, because of the lack of transparency due to the fundamental absorptions of lattice phosphate groups. In the range shown in the figure other absorption bands due to bulk species are present, namely the sharp peak at 3570 cm-1, due to the stretching of OH- aligned in columns (occupying the 4e position in the hexagonal lattice, hereafter referred to as “columnar” hydroxy groups), a set of weak components in the 2200-1900 cm-1 region, related to overtone and combination modes of the
4030 J. Phys. Chem. C, Vol. 111, No. 10, 2007 fundamental vibration of phosphate groups absorbing below 1350 cm-1, and bands in the 1550-1350 cm-1, with two main components at 1420 and 1454 cm-1, produced by the stretching modes of carbonate groups.28-34 Besides these signals, the spectrum of HA in equilibrium with the water vapor pressure (Figure 4, curve a) exhibits, as main component, a broad band spread over the 3700-2500 cm-1 range, with maximum at ca. 3350 cm-1 and a shoulder at ca. 3100 cm-1, that results from the superposition of the absorption due to the stretching mode of surface hydroxy groups and adsorbed water molecules; these latter are also responsible for the band at 1642 cm-1 (δH2O) and the very weak and ill-resolved component at 3695 cm-1, due to the stretching of dangling -OH of water molecules at the end of (H2O-H2O)n polymeric surface chains.35-38 By progressive outgassing at beam temperature (ca. 50 °C), the component at 3695 cm-1 disappears, and the bands in the 3600-2500 cm-1 range and at 1642 cm-1 decrease in intensity, monitoring the depletion of physisorbed H2O liquid-like multilayers (Figure 4, curves b-e). The lower intensity of the broad absorption at high frequency allows the detection of a component at 3490 cm-1. A signal in a similar position has been observed for both geological and synthetic apatites and assigned to columnar hydroxy groups interacting via H-bond with Cl-.39 As such a kind of anion is absent in the material investigated, we propose that the signal at 3490 cm-1 could be due to bulk OH- perturbed by carbonate anions, distorting the structure locally. Some change in shape and relative intensity of the bands of the carbonate groups occurs also, fully reversible by re-admission of water (not shown); such a sensitivity to hydration indicates that some of the carbonate groups are exposed at the surface. Remarkably, even after prolonged outgassing at beam temperature, a component at 1644 cm-1 is still present (Figure 4, curve e), with an integrated intensity of ca. 25% of the initial one (Figure 4, curve a), indicating that the sample retains a significant amount of water molecules (also contributing with their stretching modes to the broad 3600-2500 cm-1 band). Several studies on synthetic hydroxyapatites reported that this kind of material can contain H2O molecules occluded in the bulk.40-42 To assess the actual location of the retained water, the outgassed sample was put in contact with the vapor pressure of D2O at beam temperature, to selectively replace surface OH groups and H2O molecules with OD and D2O species, respectively; a deuteration of the subsurface layers and/or the bulk occurs at a significant extent at temperatures g150 °C.43 Several D2O adsoprtion/outgassing cycles were performed, until a constant intensity of the δH2O band was attained (Figure 4, curve f), corresponding to ca. 10% of that before isotopic exchange (Figure 4, curve e). Also the broad absorption in the 3600-2500 cm-1 range appeared decreased in intensity, in particular, on the low-frequency side, while a broad new component appears in the 2750-1800 cm-1 range, due to the superposition of the stretching bands of surface deuteroxy groups and still adsorbed D2O molecules. The band due to the deformation mode of these latter, expected at ca. 1200 cm-1, was not observed, because below the transparency cutoff of the material. The sample was then back-exchanged with H2O and outgassed at beam temperature, and the same spectrum that before contact with D2O was obtained (not shown). It can be then concluded that the overwhelming part of H2O molecules left in the material after prolonged outgassing at beam temperature is located on its surface. The resistance to outgassing suggests that these H2O molecules should be held on the surface
Bertinetti et al. through an interaction stronger that hydrogen bond, likely as the coordination to surface cations through the lone pairs of their oxygen atoms, as in the case of water adsorbed on Ti4+ ions on the (110) face of rutile.44 A second significant insight results from the lower effect of the H/D exchange on the intensity of the νOH component spread on the 3450-2500 cm-1 range with respect the δH2O band at 1642 cm-1. This behavior could be due to a higher extinction coefficient of the stretching absorptions of occluded water with respect to that on the surface, or to the presence in the bulk of hydroxy groups involved in hydrogen bonding, responsible for the broadness of the signal. The first hypothesis can be ruled out because of the position at higher frequency of the maximum of the νOH band left after H/D exchange (Figure 4, curve f), that indicates the presence of OH oscillators involved in weaker hydrogen bonds, and then with a lower νOH extinction coefficient. It can be then concluded that hydroxy groups chemically and/or structurally different from bulk columnar OHare present in the material. As these latter are characteristic of the crystalline structure of hydroxyapatite, it seems reasonable that the different, H-bonded hydroxy groups are mainly located in the non-apatitic surface layer observed by UHR-TEM. The research work carried out mainly by the Rey’s group20,21,45 and more recently by Ja¨ger et al.19 indicated the presence in this non-apatitic phase of HPO42- groups, which could well participate to H-bonding. The inspection of the spectral profile in the ν4 PO4 domain45 (see Figure 3S in Supporting Information) indicated that these species could be actually present. The material was then further dehydrated by outgassing at increasing temperature, and the related IR spectra are reported in Figure 5. Moreover, for selected outgassing steps, once recorded the IR spectrum, the sample was contacted with D2O and then outgassed at beam temperature, to separate the spectral pattern due to hydroxy groups and water molecules on the surface or in the bulk. For the sake of clarity, only the 36002700 and 1750-1550 cm-1 ranges of the spectra recorded after deuteration are reported (dotted lines). Full spectra are in Supporting Information, Figure 4S. Before proceeding to the next outgassing step, the sample was then back-exchanged with H2O and outgassed at the due temperature. The behavior of the δH2O band will be described first, as this signal is specific for water molecules, and then the evolution of the complex νOH signal in the 3700-2500 cm-1 region, related to the presence of both water and hydroxy groups, will be considered. Taking as a reference the spectra obtained for the sample outgassed at beam temperature (Figure 5, curves a and a′, the same as Figure 4, curves e and f), the integrated intensity of the overall δH2O absorption at 1644 cm-1 appeared decreased of ca. 70% after outgassing at 100 °C (Figure 5, b), while its subpart related to bulk water remained unchanged (Figure 5, curve b′). By outgassing at 130 °C, the δH2O band intensity decreased to ca. 90% of the initial one (Figure 5, curve c), thus closely corresponding to that of the signal due to water initially occluded within the material. Conversely, it disappeared by contact with D2O (Figure 5, curve c′), indicating that the treatment at such a temperature promoted a complete diffusion of occluded water from the bulk to the surface. After outgassing at 200 °C, only traces of the δH2O band were still observed (Figure 5, curve d), which disappeared by outgassing at 300 °C (Figure 5, curve e), then resulting in a complete desorption of water molecules from the material. Moving to the multicomponent band spread over the 37002500 cm-1 range, it can be observed that the broad shoulder at
Nanohydroxyapatite
J. Phys. Chem. C, Vol. 111, No. 10, 2007 4031
Figure 5. IR spectra of HA in the δH2O (section A) and νOH (section B) regions of HA outgassed for 120 min at: (a) beam temperature, (b) 100, (c) 130, (d) 200, and (e) 300 °C. Dotted spectra a′-c′ were obtained by contacting the sample outgassed at the corresponding temperatures with D2O vapor (23.5 mbar) and then outgassed 120 min at beam temperature.
TABLE 1: Weight Loss (by Microgravimetry) and Decrease in Intensity of the δH2O IR Band through the Various Outgassing Steps of the HA Materiala col.
1
2
3
4
5
6
entry
T (°C), P (Mbar) conditions
weight loss (mg) per 1 g of sampleb
∆w (%) with respect to the total weight loss
∆I δ(H2O) (%) with respect to the initial intensity
∆w (%) due to removal of molecular water
∆w (%) due to removal of hydroxy groups
1 2 3
T ) 50°; P from 23.5 to 1 × 10-5 T from 50° to 130°; P ) 1 × 10-5 T from 130° to 300°; P ) 1 × 10-5 total
36 ( 1.8 13 ( 0.7 12 ( 0.6 61 ( 3
59 ( 3 21 ( 1 20 ( 1 100
75 ( 2 22.5 ( 0.5 2.5 ( 0.1 100
59 ( 3 17 ( 1 2 ( 0.4 78
0 4 ( 1.4 18 ( 1.1 22
a Data are affected by an error of ca. ( 5% (microgravimetric) and ca. ( 3% (δH2O integrated intensity). bThese values result in 3.6, 1.3, 1.2, and 6.1 weight loss %, in the order.
ca. 3100 cm-1 seemed to follow the same behavior of the δH2O band, loosing most of its intensity rising the outgassing temperature up to 130 °C (Figure 5, cruves a-c), and then it should be mainly ascribed to the νH2O modes. Differently, the other broad component with maximum at 3350 cm-1 exhibited a trend uncorrelated with that of the δH2O band, in particular by increasing the outgassing temperature from 130 up to 300 °C (Figure 5, curves b-e), and then it can be mainly attributed to hydrogen-bonded hydroxy groups, both in the bulk and on the surface. As in the case of water molecules, these two locations can be distinguished on the basis of the effect of the contact with D2O. As such a treatment affected in a quite minor part the absorption left in the 3400-2500 cm-1 range after outgassing at 130 °C (Figure 5, curves c and c′), it can be concluded that such absorption is mainly due to H-bonded hydroxy groups (likely belonging to HPO42- groups) within the material. At higher frequency, the main narrow peak at 3570 cm-1 and its weaker partner at 3490 cm-1 were essentially unaffected by the outgassing temperature up to 200 °C (Figure 5, curves a-d), and appeared only slightly decreased in intensity after outgassing at 300 °C (Figure 5, curve e). Conversely, after outgassing at 130 °C a new, very weak component appeared at ca. 3660 cm-1, assignable to vibrationally free surface POH16,43,46 previously interacting with water molecules (removed from most part by outgassing at 130 °C), and then contributing to the broad feature at lower frequency. Accordingly, the broad absorption spread over the 3400-2500 cm-1 range progressively decreased in intensity as the outgassing temperature was raised up (Figure 5, curve c-e). In addition, the decrease in intensity
of this broad signal could be also due to some consumption of acid phosphates (likely giving origin to pyrophosphates). This statement is supported by the changes in the ν4 PO4 domain (see Supporting Information, Figure 3S). Finally, bulk phosphate and carbonate groups, responsible for the sets of signals in the 2200-1900 and 1600-1350 cm-1 ranges, respectively, appeared essentially stable up to outgassing at 300 °C (Figure 5, curves a-e). In contrast, such bands and the signals related to columnar OH- (3570 cm-1 and 3490 cm-1) progressively decreased in intensity and changed in shape by raising the outgassing temperature up to 800 °C, indicating the occurrence of a transformation of the bulk structure (Supporting Information, Figure 5S). 3.2.2. Quantification of H2O Interacting with Ca2+ Surface Cations. To attain an evaluation as much quantitative as possible of the hydration features of the surface, the outgassing of the material was monitored by microgravimetry. These measurements were carried out in conditions equivalent to those employed for IR studies, allowing then a transfer of the quantitative data to the spectroscopic results. The losses of weight per gram of material occurred from the initial equilibrium with a water vapor pressure of 23.5 mbar at a sample temperature of 50 °C to outgassing until constant weight at (i) 50, (ii) 130 and (iii) 300 °C are listed in Table 1. The removal of physisorbed water molecules (outgassing at 50 °C) produced a weight loss of 3.61% that corresponds to 59% ( 3% of the total weight loss (6.1%), while that in the IR data resulted in a decrease of 75% ( 2% of the integrated intensity of the δH2O band (IδH2O) in the corresponding IR spectra (Figure 1, curves a and e). The ratio between the decrease of
4032 J. Phys. Chem. C, Vol. 111, No. 10, 2007
Bertinetti et al.
Figure 6. Section A: IR spectra of HA outgassed at 130 °C for 120 min (a) and then contacted with increasing amount of water vapor (b-i) up to an equilibrium pressure of 1 mbar. Section B: plot of the integrated intensities of the νOH vs δH2O, as obtained after subtraction of spectrum a as a reference.
the δH2O band intensity and the weight loss % due to the removal of water is then of ca. 1.3. By assuming that the extinction coefficient of the δH2O mode, scarcely sensitive to the states of water,47 is the same for H2O molecules present in the material in different forms, the same ratio should hold for the IR and microgravimetric data related to both water coordinated to surface cations and occluded in the bulk. These molecules account for 25% ( 0.5% of the overall IδH2O, and then the complete removal by outgassing at 300 °C should account for a 19% ( 1% of the total weight loss determined in the microgravimetric measurements, stopped at that temperature. As such a loss was of 61 ( 3 mg per gram of material, H2O molecules coordinated to surface cations and occluded in the bulk should correspond to a mass of 11.6 ( 0.5 mg. Isotopic H2O/D2O exchange experiment indicated that H2O molecules in the bulk represent ca. 10% of water still present after outgassing at 50 °C, and then the amount of H2O adsorbed to surface cations should be of ca. 10.5 ( 0.5 mg per gram of material. Taking into account the value of SSABET ) 78 ( 8 m2‚g-1, such amount corresponds to 4.5 ( 0.5 H2O molecules per nm2 left coordinated to surface cations. Interestingly, Kukura et al.48 determined the amount of Ca2+ exposed at the surface of hydroxyapatite particles to be of ca. 4.3 per nm2, and then the obtained value indicates that each coordinated water molecule should seat on one surface cation. Moreover, the amount of water molecules per nm2 we found is in good agreement with the result of ca. 5.4 obtained by Park et al. for the first water layer in direct contact with the surface of fluoroapatite.49 3.2.3. Quantification of Surface Hydroxyl Groups. Having determined the contribution of molecular water (Table 1, column 5) to the weight loss % recorded at each step of ougtassing (Table 1, column 3), it is straightforward to derive, by difference, the contribution due to the removal of hydroxy groups, (Table 1, column 6), accounting for a ca. 22% ( 3% of the total (4% ( 1.4% by outgassing from 50 to 130 °C, and 18% ( 1.1% by outgassing from 130 to 300 °C). The following step, aimed to extract from the combination of these data with IR results the amount of hydroxyl groups exposed at the surface of the material, is rendered more difficult by the overlap in the νOH region of contributions from such groups and water molecules, that, in addition, are present both in the bulk and on the surface. Furthermore, the vibrational features of the OH groups acting as proton donors in H-bonding are very sensitive to the H-bond (stronger this last, lower the
νOH frequency and higher the extinction coefficient related to the νOH absorption50). However, at least an evaluation can be proposed by taking into consideration that: (a) the components assigned to columnar OH- (peak at 3570 cm-1 and the component at 3490 cm-1), appeared quite insensitive to the outgassing up to 300 °C, while this treatment significantly affected the broad absoprtion due to H-bonded hydroxy groups; (b) as reported when commenting spectra c and c′ in Figure 5, the overwhelming part of such band remaining after outgassing at 130 °C is due to H-bonded hydroxy groups in the bulk; (c) as a consequence, most part of surface H-bonded hydroxy groups have been then removed by increasing the outgassing temperature from 50 to 130 °C; (d) the removal of hydroxy groups during such outgassing step should correspond to 4% ( 1.4% of the total weigh loss of 61 ( 3 mg per grams of material (Table 1), and then to ca. 2.2 ( 0.7 mg per grams of material. Hence, by assuming that all hydroxy groups removed in this outgassing step are on the surface, such an amount should correspond to 1 ( 0.4 surface hydroxyl groups per nm2 (overestimated, as actually also bulk H-bonded hydroxy groups should have been removed by outgassing at 130 °C). This value is significantly lower than that obtained for water molecules coordinated to surface cations (4.5 ( 0.5 per nm2), which then appear to represent the overwhelming part of the first hydration layer of the synthetic hydroxyapatite studied. 3.2.4. ReVersibility of the Surface Hydration in Dependence on the Outgassing Temperature. The investigated material was produced at mild temperature and in equilibrium with an aqueous medium, and then combined dehydration and energetic inputs experienced by the material during outgassing at higher temperature could have modified its surface features. In this regard, the reversibility of dehydration attained was assessed in a subsequent series of experiments by re-contacting the sample with water vapor after each outgassing step. The first aspect considered has been the nature of readsorption, if molecular or dissociative. Figure 6A shows a series of IR spectra of the material pre-outgassed at 300 °C and then contacted with small doses of H2O, resulting in the developing of absorptions in the νOH and δH2O regions. Differences with respect the initial spectrum were performed, and the integrated intensity (I) of the absorptions in the two regions calculated. The results are reported in Figure 6B, where the values of IνOH
Nanohydroxyapatite
Figure 7. Plot of the surface hydration level attained after: (b) readmission of water vapor (25.3 mbar) and (9) subsequent outgassing at beam temperature on HA outgassed at increasing temperature, with respect to the untreated material (100%). The evaluation was carried out on the basis of the integrated intensities of the δH2O band in the IR spectra.
versus IδH2O are plotted. It can be observed that the they exhibit a highly linear relationship, indicating that νOH and δH2O signals are due to the same oscillators in the whole data range, i.e., water is essentially readsorbed in a molecular way. Similar results where obtained for the materials pre-outgassed at 100, 130, and 200 °C. Actually, the occurrence, even partial, of dissociative absorption should have depressed the intensity of the δH2O band (related only to water molecules) in favor of the νOH band (in such a case related also to newly formed hydroxyl groups) in the spectra obtained after the admission of the first doses of water, where H2O molecules have a higher probability to interact with sites (e.g., Lewis acid-base pairs) strong enough to promote their dissociation.51,52 Because of the molecular character of water readsorption, the reversibility of the dehydration was evaluated by calculating the percent value of the integrated intensity exhibited by the δH2O band obtained by rehydration after each outgassing step, with respect to the IδH2O recorded for the original, thermally untreated, material (data in Figure 7). In particular, for each case two hydration conditions were considered: i) sample preoutgassed at the due temperature in contact with water vapor pressure at room temperature (when both water molecules coordinated to surface cations and physisorbed in upper layers are present, circle symbol), and (ii) after subsequent outgassing at beam temperature (when only water molecules coordinated to surface Ca2+ are left, square symbol). It can be observed that already after outgassing at 100 °C the material lost ca. 10% of the capability to adsorb water in the overall physisorbed + coordinative and coordinative only states. Outgassing at higher temperatures resulted in a progressive increase in such a loss, more severe for the amount of water coordinated to surface cations. Actually, the amount of surface water re-adsorbed in this state after pre-outgassing at 300 °C was only ca. 40% of that found for untreated material. As no significant changes in SSABET occurred by outgassing up to 300 °C, the observed trend can be interpreted in terms of changes in the surface structure, likely with a downward relaxation of initially exposed Ca2+ ions, then not longer accessible to H2O molecules. A similar behavior was observed for Ti4+ centers exposed at the surface of a commercial titania powder.44 The decreased extension of this type of surface “primary” hydration might also limit the subsequent adsorption of additional water overlayers. It is worth to notice that changes in the surface structure by outgassing at 300 °C are not related to a transformation of the
J. Phys. Chem. C, Vol. 111, No. 10, 2007 4033
Figure 8. IR spectra of CO adsorbed at ca. 100K on HA outgassed at 130 °C (section A) and 300 °C (section B). Lettering is in the sense of decreasing CO pressure, from (a) 25 mbar to (o) outgassing for 1 min. Spectra are reported after subtraction of the spectrum of HA outgassed at the corresponding temperatures before CO adsorption. In the insets are the original spectra in the 3680-3625 cm-1 region of: (a′) HA pre-outgassed at 130 °C (section A) and 300 °C (section B) and (b′) in contact with 25 mbar CO (both sections).
surface layer in a more ordered structure, as evidenced by UHRTEM (see Supporting Information, Figure 2S, section B). 3.3. Monitoring of Surface Ca2+ Sites by IR Spectroscopy of CO Adsorbed at 100 °C. The data discussed in the previous section indicated that outgassing of HA at temperature higher than ca. 50 °C promoted the desorption of water molecules from surface Ca2+ ions, then allowing the adsorption of other molecules able to probe additional features of these sites. In this regards, significant information can be provided by the study of the IR spectra of adsorbed CO, as well-known in the field of surface science applied to heterogeneous catalysts.53 As reported above, pre-outgassing at 130 °C promoted the desorption of water molecule from ca. 85% of surface Ca2+ centers, with a modification of the surface structure limited to 20% of these sites. Such pre-outgassing appeared then the activation condition allowing a good compromise between number of cationic surface sites accessible by CO and extension of surface modification. Figure 8A reports the spectra in the νCO region obtained for decreasing coverage of CO adsorbed at ca. 100 K on HA preoutgassed at 130 °C. In the presence of 20 Torr CO, a quite intense band with maximum at 2170 cm-1, slightly asymmetric on the lowfrequency side, was observed (Figure 8A, curve a). Higher CO pressures (or lower temperatures) simply resulted in the growth of an additional component at ca. 2140 cm-1, due to CO aspecifically adsorbed in a liquid-like state. By decreasing the CO coverage the band intensity progressively decreased and the maximum shifted to higher frequency (Figure 8A, curves b-g). Finally, in the last stages, the band profile appeared quite broadened because of the contribution of a component at ca. 2180 cm-1 (Figure 8, curves h-o), previously overcome by the main absorption at lower frequency. Such a component appeared slightly more resistant to the outgassing and moved and/or broadened toward high frequency in an extent large enough to cross the high frequency onset of the spectral profile at higher CO coverage. At the various coverages, the νCO band appeared located in the region where CO molecules in interaction with Ca2+ ions absorb, ranging from centers with a strong Lewis acid strength, as Ca2+ in supercages of Y zeolite (νCO ) 2197 cm-1)54 to weak Lewis acid sites as Ca2+ penta-coordinated at the surface of
4034 J. Phys. Chem. C, Vol. 111, No. 10, 2007 CaO (νCO ) 2155 cm-1).55 However, beside calcium cations, surface hydroxy groups can also act as adsorbing centers toward CO, which, when adsorbed on isolated P-OH, was observed to produce an IR band at ca. 2170 cm-1.56 Actually, P-OH are exposed at the surface of the HA material studied, and their νOH signal at ca. 3660 cm-1 was the only one in the highfrequency region to be perturbed (downshifted) when CO was admitted on the sample (Figure 8A, inset), indicating their interaction with probe molecules. Nevertheless, an evaluation based on literature data56 of the ratio between the integrated intensity of the νOH band of unperturbed P-OH and that expected for the stretching band of CO adsorbed on them, suggested that in the present case this latter should contribute to the νCO signal for a negligible part (displayed in Figure 8A as the simulated shadowed component). The νCO band observed must be then essentially attributed to CO molecules adsorbed on surface Ca2+ sites by electrostatic/ polarization interaction53 that results in an upward shift of the stretching frequency of adsorbed CO, increasing as the Lewis acid strength of the adsorbing center increases. The two components at 2170 and 2180 cm-1 monitor the presence of two families of Lewis acid centers, the stronger ones constituting a minor fraction. Surface Ca2+ ions with Lewis acid features similar to these latter were observed for Ca-doped Al2O3.57 The upward shift exhibited by both components as the CO coverage decreased should be ascribed to static adsorbateadsorbate through-solid interactions of inductive nature, as widely observed on insulating materials, such as oxides58 and salts.59-61 For the sake of completeness, IR spectra of CO adsorbed on HA pre-outgassed at 300 °C were also recorded, as this treatment resulted in the complete desorption of water molecules from the surface accompanied by a severe modification of the surface itself. Surprisingly, the spectra of adsorbed CO at decreasing coverage (Figure 8B) were not so markedly different form those obtained for the sample pre-outgassed at 130 °C: the component at high frequency appeared more intense and slightly up-shifted, but the overall intensity of the spectral pattern at high coverage was essentially equivalent to the previous case. By assuming that the molar extinction coefficient of adsorbed CO species is independent of frequency changes in the 2185-2165 cm-1 interval,62,63 such spectral pattern indicated that the amount of surface Ca2+ probed by CO was essentially unaffected by the pre-outgassing at higher temperature, while a change in the local structure of a significant part of them occurred, resulting in an increase of their Lewis acid strength. There is then a relevant discrepancy between the indication on the effect of outgassing at 300 °C on the surface structure provided by the rehydration test, i.e., by H2O as probe molecules, and the adsorption of CO. However, it can be proposed that a part of surface Ca2+ dehydrated by outgassing at 130 °C relaxed inward the surface, in an extent large enough to render them unable to adsorb a weak probe molecule as CO, that then can monitor the Ca2+ ions remained more exposed at the surface. However, the relaxation can be reversed by interaction with H2O, definitely stronger as ligand, and then in rehydration tests water molecules probed the presence of a larger amount of surface Ca2+ sites. By increasing the outgassing temperature up to 300 °C, a more severe modifications of the surface structure could have occurred, resulting in an inward relaxation, irreversible by rehydration, of ca. 60% of surface cations, that then cannot be longer probed neither by CO nor H2O. Furthermore, also the local structure of a relevant part of
Bertinetti et al. Ca2+ remained exposed could have been modified, with a consequent increase of their Lewis acidity. 4. Conclusions As indicated by TEM, the adopted preparation method resulted in the production of actual nanosized hydroxyapatite particles that are elongated in the direction of the crystallographic c-axis, similarly to biological nanoapatites and synthetic larger HA grains. Primary particles can be joined through grain boundaries forming larger secondary nanostructured particles. Significantly, TEM observations in Ultra Highresolution conditions provided evidence of the presence of a disordered surface layer, 1-2 nm thick, surrounding a crystalline core, in agreement with previous solid-state NMR, IR and X-ray absorption studies by other groups. The combined use of IR spectroscopy and microgravimetry and the control of the atmosphere surrounding the samples in both kind of experiments was useful for a quantitative estimation of the amount of hydroxyl groups and water constituting the first surface hydration layer. Such a layer, for the as-prepared material, appeared mainly constituted by H2O molecules coordinated to surface Ca2+ ions, approximately in a 1:1 ratio. This interaction is ruled by the Lewis acid strength of the exposed cations, evaluated by the use of CO as vibrational probe. Furthermore, IR spectra of adsorbed CO provided indication of the presence of at least two families of surface Ca2+ differing in Lewis acid strength. Removal of water by outgassing at increasing temperature produced a progressive modification of the surface structure, with an increased level of irreversibility as the temperature increased, resulting in the loss of sites able to coordinate H2O molecules by subsequent rehydration. In turn, such changes in the first hydration layer resulted in the decrease of the extension of water overlayers. This result indicated that it can be possible to modify in a controlled way the behavior toward water of synthetic HA materials, one of the main aspects ruling their interaction with the biological media. Studies based on the methods adopted for this work, devoted to the investigation of the modification of the surface features of synthetic HA nanoparticles by introduction of cations different from Ca2+ (e.g., Mg2+) and/or aging treatments, are in progress. Acknowledgment. The authors wish to thank Regione Piemonte (Italy), research project “Biocompatible Nanostructured Materials for Biomedical Applications” (fall 2004, project code D33) and Royal Society for financial support; Prof. Piero Ugliengo (Dipartimento di Chimica IFM, University of Torino) is acknowledged for fruitful discussions. L.B. is obliged to Prof. Sir J. M. Thomas for having provided access to electron microscopy facilities in Cambridge. Supporting Information Available: XRD pattern, TEM images of HA particles outgassed at 300 °C, IR-ATR spectra of HA untreated, outgassed at 300 °C, and references to highly crystalline HA in the 500-620 cm-1 range, IR spectra of HA: (1) outgassed between 100 and 300 °C, contacted with D2O and re-outgassed at r.t., (2) outgassed between 300 and 800 °C. This material is available free of charge via the Internet at http:// pubs.acs.org. References and Notes (1) Castner, D. G.; Ratner, B. D. Surf. Sci. 2002, 500, 28-60. (2) Kasemo, B. Surf. Sci. 2002, 500, 656-677. (3) Tirrell, M.; Kokkoli, E.; Biesalski, M. Surf. Sci. 2002, 500, 6183.
Nanohydroxyapatite (4) Wilson, C. J.; Clegg, R. E.; Leavesley, D. I.; Pearcy, M. J. Tissue Eng. 2005, 11, 1-18. (5) a) Webster, T. J.; Ejiofor, J. U. Biomaterials 2004, 25, 4731-4739. b) Webster, T. J.; Ergun, C.; Doremus, R. H.; Siegel, R. W.; Bizios, R. Biomaterials 2001, 22, 1327-1333. c) Webster, T. J.; Ergun, C.; Doremus, R. H.; Siegel, R. W.; Bizios, R. Biomaterials 2000, 21, 1803-1810. d) Webster, T. J.; Siegel, R. W.; Bizios, R. Biomaterials 1999, 20, 12211227. (6) Thapa, A.; Miller, D. C.; Webster, T. J.; Haberstroh, K. M. Biomaterials 2003, 24, 2915-2926. (7) Elias, K. L.; Price, R. L.; Webster, T. J. Biomaterials 2002, 23, 3279-3287. (8) Price, R. L.; Waid, M. C.; Haberstroh, K. M.; Webster, T. J. Biomaterials 2003, 24, 1877-1887. (9) Balasundaram, G.; Sato, M.; Webster, T. J. Biomaterials 2006, 27, 2798-2805. (10) Gutwein, L. G.; Webster, T. J. Biomaterials 2004, 25, 4175-4183. (11) Savaiano, J. K.; Webster, T. J. Biomaterials 2004, 25, 1205-1213. (12) Sato, M.; Slamovich, E. B.; Webster, T. J. Biomaterials 2005, 26, 1349-1357. (13) a) Landi, E.; Tampieri, A.; Celotti, G.; Langenati, R.; Sandri, M.; Sprio, S. Biomaterials 2005, 26, 2835-2845. b) Tampieri, A.; Celotti, G.; Landi, E. Anal. Bioanal. Chem. 2005, 381, 568-576. c) Celotti, G.; Landi, E.; Sandri, M.; Roveri, N.; Falini, G. J. Biomed. Mater. Res. Part A 2003, 67A, 618-625. (14) Neuman, W. F.; Toribara, T. Y.; Mulryan, B. J. J. Am. Chem. Soc. 1953, 75, 4239-4242. (15) Dry, M. E.; Beebe, R. A. J. Phys. Chem. 1960, 64, 1300-1304. (16) Ishikawa, T.; Wakamura, M.; Kondo, S. Langmuir 1989, 5, 140144. (17) Cheng, Z. H.; Yasukawa, A.; Kandori, K.; Ishikawa, T. Langmuir 1998, 14, 6681-6686. (18) Isobe, T.; Nakamura, S.; Nemoto, R.; Senna, M.; Sfihi, H. J. Phys. Chem B 2002, 106, 5169-5176. (19) Ja¨ger, C.; Welzel, T.; Meyer-Zaika, W.; Epple, M. Magn. Reson. Chem. 2006, 44, 573-580. (20) Cazalbou, S.; Combes, C.; Eichert, D.; Rey, C. J. Mater. Chem. 2004, 14, 2148-2153. (21) Eichert, D.; Salome, M.; Banu, M.; Susini, J.; Rey, C. Spectroc. Acta Pt. B-Atom. Spectr. 2005, 60, 850-858. (22) Stadelmann, P. A. Ultramicroscopy 1987, 21, 131-145. (23) Suvorova, E. I.; Buffat, P. A. J. Microsc. 1999, 196, 46-58. (24) Tamai, M.; Nakamura, M.; Isshiki, T.; Nishio, K.; Endoh, H.; Nakahira, A. J. Mater. Sci.-Mater. Med. 2003, 14, 617-622. (25) Meldrum, A.; Wang, L. M.; Ewing, R. C. Am. Miner. 1997, 82, 858-869. (26) Wang, L. M.; Wang, S. X.; Ewing, R. C.; Meldrum, A.; Birtcher, R. C.; Provencio, P. N.; Weber, W. J.; Matzke, H. Mater. Sci. Eng. A-Struct. Mater. Prop. Microstruct. Process. 2000, 286, 72-80. (27) Vallet-Regi, M.; Gutierrezrios, M. T.; Alonso, M. P.; Defrutos, M. I.; Nicolopoulos, S. J. Solid State Chem. 1994, 112, 58-64. (28) Busca, G.; Lorenzelli, V. Materals Chemistry 1982, 7, 89-126. (29) Koutsopoulos, S. J. Biomed. Mater. Res. 2002, 62, 600-612. (30) Berry, E. E.; Baddiel, C. B. Spectrochim. Acta A. 1967, 23, 17811792. (31) LeGeros, R. Z.; LeGeros, J. P.; Trautz, O. R. DeV Appl Spectros 1970, 7B, 3-12. (32) Rheman, I.; Bonfield, W. J. Mater. Sci. Mater. Med. 1997, 8, 1-4. (33) Tadic, D.; Peters, F.; Epple, M. Biomaterials 2002, 23, 2553-2559. (34) Landi, E.; Celotti, G.; Logroscino, G.; Tampieri, A. J. European Ceram. Soc. 2003, 23, 2931-2937.
J. Phys. Chem. C, Vol. 111, No. 10, 2007 4035 (35) Buch, V.; Devlin, J. P. J. Chem. Phys. 1991, 94, 4091-4092. (36) Jiang, J. C.; Chang, J. C.; Wang, B. C.; Lin, S. H.; Lee, Y. T.; Chang, H. C. Chem. Phys. Lett. 1998, 289, 373-382. (37) Rowland, B.; Fisher, M.; Devlin, J. P. J. Chem. Phys. 1991, 95, 1378-1384. (38) Miyazaki, M.; Fujii, A.; Ebata, T.; Mikami, N. Science 2004, 304, 1134-1137. (39) Fowler, B. O. Inorg. Chem. 1974, 13, 194-207. (40) Eunsung, P.; Condrate, S. R.; Lee, D.; Gallagher, K.; Gallagher, P. K. J. Mater. Sci.-Mater. Med. 2002, 13, 211-218. (41) Fu¨redi-Milhofer, H.; Hlady, V.; Baker, F. S.; Beebe, R. A.; Wikholm, N. W.; Kittelberger, J. S. J. Colloid Interface Sci 1979, 70, 1-9. (42) LeGeros, R. Z.; Bonel, G.; Legros, R. Calcif. Tiss. Res. 1978, 26, 111-118. (43) Ishikawa, T.; Teramachi, A.; Tanaka, H.; Yasukawa, A.; Kandori, K. Langmuir 2000, 16, 10221-10226. (44) Martra, G. Appl. Catal. A-Gen. 2000, 200, 275-285. (45) Rey, C.; Shimizu, M.; Collins, B.; Glimcher, M. J. Calcif. Tiss. Int. 1990, 46, 384-394. (46) Ishikawa, T. In Adsorption on New and Modified Inorganic Sorbents; Dabrowski, A., Tertykh, V. A., Eds.; Elsevier: Amsterdam, 1996; pp 301-318. (47) Burneau, A.; Gallas, J. P. In The Surface Properties of Silicas; Legrand, A. P., Ed.; Wiley: New York, 1998, p 196. (48) Kukura, M.; Bell, L. C.; Posner, A. M.; Quirk, J. P. J. Phys. Chem. 1972, 76, 900. (49) Park, C.; Fenter, P.; Zhang, Z.; Cheng, L.; Sturchio, N. C. Am. Miner. 2004, 89, 1647-1655. (50) George, W. O.; Lewis, R. In Handbook of Vibrational Spectroscopy; Chalmers, J. M., Griffiths, P. R., Eds.; Wiley: Chichester, U.K., 2002, p 1992. (51) Zecchina, A.; Lofthouse, M. G.; Stone, F. S. J. Chem. Soc. Faraday Trans. 1 1975, 71, 1476-1490. (52) Coluccia, S.; Tench, A. J.; Segall, R. L. J. Chem. Soc. Faraday Trans. 1 1979, 75, 1769-1779. (53) Hadjiivanov, K. I.; Vayssilov, G. N. AdV. Catal. 2002, 47, 307511. (54) Bolis, V.; Fubini, B.; Garrone, E.; Giamello, E.; Morterra, C. In Structure and ReactiVity of Surfaces; Morterra, C., Zecchina, A., Costa, G., Eds.; Elsevier: Amsterdam, 1989, p 159. (55) Babaeva, M. A.; Bystrov, D. S.; Kovalgin, A. Y.; Tsyganenko, A. A. J. Catal. 1990, 123, 396-416. (56) Gianotti, E.; Oliveira, E. C.; Coluccia, S.; Pastore, H. O.; Marchese, L. Stud. Surf. Sci. Catal. 2004, 154, 1498-1504. (57) Morterra, C.; Bolis, V.; Magnacca, G.; Cerrato, G. J. Electron Spectrosc. Relat. Phenom. 1993, 64-5, 235-240. (58) Hadjiivanov, K. I.; Vayssilov, G. N. Characterization of Oxide Surfaces and Zeolites by Carbon Monoxide as an IR Probe Molecule; Academic Press: San Diego, 2002, Vol. 47; pp 307-511. (59) Zecchina, A.; Scarano, D.; Garrone, E. Surf. Sci. 1985, 160, 492508. (60) Scarano, D.; Zecchina, A. J. Chem. Soc. Faraday Trans. 1. 1986, 82, 3611-3624. (61) Chang, H. C.; Ewing, G. E. Phys. ReV. Lett. 1990, 65, 2125. (62) Platero, E. E.; Scarano, D.; Spoto, G.; Zecchina, A. Faraday Discuss. Chem. Soc 1985, 80, 183-193. (63) Hush, N. S.; Williams, M. L. J. Mol. Spectrosc. 1974, 50, 349368.
