Colloidally Confined Crystallization of Highly ... - ACS Publications

Subscriber access provided by Kaohsiung Medical University

Functional Inorganic Materials and Devices

Colloidally Confined Crystallization of Highly Efficient Ammonium Phosphomolybdate Catalysts Alice Antonello, Cesare Benedetti, Francisco Fernando Perez-Pla, Maria Kokkinopoulou, Katrin Kirchhoff, Viktor Fischer, Katharina Landfester, Silvia Gross, and Rafael Muñoz-Espí ACS Appl. Mater. Interfaces, Just Accepted Manuscript • Publication Date (Web): 08 Jun 2018 Downloaded from http://pubs.acs.org on June 8, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Colloidally Confined Crystallization of Highly Efficient Ammonium Phosphomolybdate Catalysts Alice Antonello,† Cesare Benedetti,† Francisco F. Pérez-Pla,¶ Maria Kokkinopoulou,† Katrin Kirchhoff,† Viktor Fischer,† Katharina Landfester,† Silvia Gross,‡ Rafael Muñoz-Espí*,†,¶ †

Max Planck Institute for Polymer Research, Ackermannweg 10, 55128, Mainz, Germany.

¶

Institut de Ciència dels Materials (ICMUV), Universitat de València, c/ Catedràtic José Beltrán 2, 46980 Paterna, Spain.

‡

Dipartimento di Scienze Chimiche, Università degli Studi di Padova, via Marzolo 1, 35131 Padova, Italy.

KEYWORDS. Miniemulsion, molybdenum, catalysis, crystallization, nanoparticle

ABSTRACT: Nanodroplets in inverse miniemulsions provide a colloidal confinement for the crystallization of ammonium phosphomolybdate, influencing the resulting particle size. The effects of the space confinement are investigated by comparing the crystallization of analogous materials both in miniemulsion and in bulk solution. Both routes result in particles with rhombododecahedral habit, but the ones produced in miniemulsion have sizes between 40 and 90 nm, three orders of magnitude smaller than the ones obtained in bulk solution. The catalytic activity of the materials is studied taking the epoxidation of cis-cyclooctene as a model reaction. The miniemulsion route yields ammonium phosphomolybdate particles catalytically much more active than analogous samples produced in bulk solution, which can be explained by their higher dispersibility in organic solvents, their higher surface area, and their higher porosity. Inorganic phosphate salt precursors are compared with organic phosphate sources. Ammonium phosphomolybdate nanoparticles prepared in miniemulsion from D-glucose-6-phosphate and O-phospho-DL-serine yield a conversion in the epoxidation reaction of more than 90% after only 1 h, compared to 30% for materials prepared in bulk solution. In addition, the catalysts prepared in miniemulsion display a promising recyclability.

INTRODUCTION Ammonium phosphomolybdate (APM), (NH4)3PO4(MoO3)12, was reported for the first time by Berzelius in 1826.1 Its crystal structure remained not well understood until 1934, when Keggin determined by X-ray diffraction the structure of the phosphotungstate anion,2 analogous to the phosphomolybdate. APM was one of the first examples of a polyoxometalate (POM), which is a polyatomic ion formed by transition metal oxo-anions connected by bridging oxygen atoms to create a closed three-dimensional framework. The structure of APM, named as Keggin structure, is one of the most typical for POMs. In this framework, the phosphorus atom is at the center of a tetrahedron surrounded by four oxygen atoms. This tetrahedron is caged by 12 molybdenum oxide octahedra linked together by bridging oxygen atoms; thus, the molybdenum oxide units are almost equidistant. The precipitation of APM in acid media leads easily to the formation of rhombododecahedral particles, representing a clear example of self-assembly of the Keggin structures

through hydrogen bonding.3-5 The ammonium ions and the water molecules are probably accommodated between the anion units of the framework. There is a relatively large space between the units, so that larger ions could fit in those positions and ammonium ions can isomorphically exchange with monovalent cations in the crystal lattice. Therefore, APM may be used as an ion exchanger.6, 7 This feature can be applied, for instance, to remove cesium ions after incomplete disposal of radioactive waste from nuclear industries, nuclear weapon tests, or nuclear reactor accidents.8 Ammonium phosphomolybdate can also be employed as a catalyst, because it is a multi-electron oxidant, able to acquire up to 6 electrons without changing its Keggin structure.9 APM presents also other convenient features for heterogeneous catalysis:10 it is the salt of a solid heteropolyacid, the phosphomolybdic acid, and it is a strong Brønsted acid (almost superacid);11 it is insoluble in water and concentrated acids (solubility in water at 20 °C = 0.2 ± 0.1 g L−1) but completely soluble in 1

ACS Paragon Plus Environment

ACS Applied Materials & Interfaces 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

concentrated bases; it is stable against hydration, dehydration, and heating up to 400-500 °C; and, finally, it is microporous. These characteristics are exploitable in the refinery industry and for fine chemical production:9 for the conversion of methanol to hydrocarbons, for the desulfurization of diesel,12 for the hydration of olefins (such as propene),13 and for the oxidation of isobutane to methacrylic acid.14 Ammonium phosphomolybdate is also employed as a photocatalyst for the hydroxylation of benzophenone15 and for the photodegradation of dyes.16 In the present work, we synthesize APM starting from different phosphate sources, both organic and inorganic, keeping the same molybdenum precursor. By employing different phosphates, the effects of the counterions (in case of inorganic sources) and of the degradation kinetics of the organophosphates on the final structure of the materials can be evaluated, analyzing thereby the influence of the availability of free phosphate ions in solution. Actually, the different availability of the phosphate ions has been the subject of various investigations since the beginning of the 20th century, because it is a key point for the analytical determination of the phosphorus in solution and human fluids (blood and urine). Fiske and SubbaRow17 proposed a colorimetric method to determine the amount of phosphorus in the human fluids using the ammonium heptamolybdate in acid conditions (concentrated sulfuric acid) and the subsequent reduction of the molybdenum by aminonaphtolsulfonic acid to give a blue mixture (molybdenum blue). The inorganic phosphorus was easily measured in these conditions, while the organic one required a harsher method (i.e., higher temperature and higher acid concentration). Sumner18 took advantage of the difference between the organic and the inorganic phosphates and developed a method with ferrous sulfate as reducing agent and lower acidity, providing a greater specificity for the inorganic phosphorus, available also under these milder conditions. In accordance with these findings, polyphosphoesters require high temperature and low pH values to be hydrolyzed. In nature, the dephosphorylation (i.e., removal of the phosphate group from an organic compound by hydrolysis) is catalyzed by specific enzymes, the phosphatases.19 The organic phosphate precursors used in our work, namely phytic acid dodecasodium salt, D-glucose-6-phosphate sodium salt, and O-phospho-DL-serine have their specific phosphatase, respectively phytase, glucose 6-phosphatase, and phosphoserine phosphatase, involved in important processes for the human body.20 To the best of our knowledge, the mechanism of formation and the kinetics of precipitation of APM have been so far only studied for different counterions of ammonium salts,4 but not for different organophosphate sources. Herein, we also address the influence of the space confinement provided by colloidal systems in the crystallization of APM. More specifically, we study how APM crystallizes within the nanodroplets of water-in-oil miniemulsions. The water droplets of an inverse miniemulsion can be considered as “nanoreactors” not

Page 2 of 18

only in the context of polymer synthesis, where they have been widely used,21 but also in inorganic crystallization processes.22-28 Our working hypothesis has been that the synthesis of ammonium phosphomolybdate in the confinement of nanodroplets can provide crystals much smaller than those obtained by bulk crystallization from solution and with enhanced functional properties in the context of heterogeneous catalysis. Accordingly, in the last part of the work, the catalytic activity of the APM materials was evaluated taking the epoxidation of cis-cyclooctene as a model reaction.

EXPERIMENTAL SECTION Chemicals. Ammonium molybdate tetrahydrate (ACS reagent ≥99.0% (T), Fluka), sodium phosphate monobasic (bioXtra, ≥99.0% (RT), Sigma-Aldrich), ammonium dihydrogen phosphate (for analysis 99+%, Across Organic), phytic acid sodium salt hydrate (from rice, Sigma-Aldrich), D-Glucose 6-phosphate sodium salt (≥98%, for biochemistry, Carl Roth), O-phospho-DLserine (Sigma-Aldrich), nitric acid (puriss. p.a., reag. ISO, ≥65%, Sigma-Aldrich), cyclohexane (HiPerSolv, Chromanorm, VWR chemicals), and polyglycerol polyricinoleate (PGPR, GRINDSTED PGPR 90, Danisco) were used as received. Milli-Q water (Millipore, resistance 18.2 MΩ cm) was used to prepare aqueous solutions and suspensions. For the catalytic tests, chloroform (Spectronorm, assay 99.2% VWR chemicals), tert-butyl hydroperoxide (TBHP) 5.5 M solution in nonane (SigmaAldrich), chlorobenzene (99.9% for HPLC, Acros Organics), and cyclohexene (inhibitor-free, 99%, SigmaAldrich) were used as received; cis-cyclooctene (95%, Sigma-Aldrich, with 100-200 ppm Irganox 1076 FD as antioxidant), styrene (ReagentPlus, stabilized, ≥99%, Sigma-Aldrich), and butyl acrylate (99%, stabilized, Aldrich) were purified by column chromatography through an aluminum oxide column to remove the stabilizers. Crystallization of Ammonium Phosphomolybdate from Solution (“Bulk Solution Crystallization”). In a typical bulk solution synthesis (i.e., direct precipitation from an aqueous solution), ammonium molybdate tetrahydrate and the selected phosphate were dissolved in water to prepare two 0.1 M solutions of equal volume. The two solutions were mixed together in a round-bottom flask and placed in an oil bath at 80 °C under magnetic stirring (400 rpm). After 20 min of stirring, concentrated nitric acid was added to the system in the same volume as the ammonium molybdate solution (i.e., a large excess of HNO3). The acid started the precipitation of the product. After the selected time (generally 1 h, unless otherwise specified), the reaction was quenched in an ice-water bath and the precipitate recovered by centrifugation (13,000 rpm, 5 min). Afterwards, the product was washed three times with deionized water. The solid obtained was dried under vacuum at room temperature for 15 h and subsequently ground.

2


Page 3 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Crystallization of Ammonium Phosphomolybdate in Miniemulsion. In a typical miniemulsion synthesis, ammonium molybdate tetrahydrate and the selected phosphate were dissolved in Milli-Q water, both at a concentration of 0.1 M. These aqueous solutions were mixed with the organic phase (i.e., 1 wt% solution of PGPR in cyclohexane), in a weight ratio of 1:4. The aqueous and the organic phases were stirred together to promote a pre-emulsification. The mixture was then ultrasonified with a Branson digital sonifier W-450D (½” tip, sonication time of 2 min, 70% amplitude, pulse of 1 s, pause of 0.1 s). Immediately after the ultrasonication, an excess (i.e., 12 times the moles of Mo) of HNO3 (65 wt%, 14.4 M) was added to the freshly prepared stable miniemulsion, and this mixture was ultrasonified again with the same parameters to trigger the penetration of HNO3 inside the droplets and to start the precipitation of the product within that confined space. The miniemulsion was immediately placed at 80 °C under stirring (400 rpm) for 1 h. After the reaction time, the suspension was centrifuged (13,000 rpm, 5 min) to collect the solid product. Subsequently, this product was washed once with a mixture of acetone and ethanol (1:1 v/v) and three times with deionized water. The obtained solid was dried under vacuum at room temperature for 15 h and eventually ground. Characterization Methods. Scanning electron microscopy (SEM) was performed in a Leo Gemini Zeiss 1530 microscope with an extractor voltage of 700 V. Samples for SEM characterization were prepared either by depositing dry powder on carbon tape (in case of the bulk solution samples) or by drop-casting of diluted dispersions on silicon wafers (in case of miniemulsion samples). A FEI Tecnai F20 microscope operated at an acceleration voltage of 200 kV was used for transmission electron microscopy (TEM) measurements. The micrographs were recorded on a 2k CCD camera (Gatan Ultrascan 1000). The sample was embedded in a matrix of ionic liquid 1-ethyl-3-methylimidazolium tetrafluoroborate. Further details of the preparation method of the TEM samples are reported elsewhere.29 Average particle sizes were calculated measuring the dimensions of a statistical number of particles from SEM and TEM images by using the software Fiji-ImageJ.30 Focused ion beam (FIB, FEI Nova600Nanolab) was employed to cut a lamella from one of the ammonium phosphomolybdate bulk samples (specifically, sample APM-glu). More details about the sample preparation are provided in the Supporting Information. High-resolution micrographs of the lamella were collected with by TEM. X-ray diffraction (XRD) of washed and vacuum-dried powders was conducted with a Philips PW 1820 diffractometer with monochromatic Cu Kα radiation (λ = 1.5418 Å, 40 kV, 30 mA, Δ2θ = 0.02°, 5 s per step). The size of the coherently scattering domains (simplistically referred to as crystallite sizes) was calculated with the Scherrer formula31 on the most intense reflection of the diffractograms, namely (222).

The interfacial tension values were measured with a spinning drop tensiometer (Dataphysics SVT 20N) equipped with a thermostatic bath to reach 80 °C. Inductively coupled plasma–mass spectrometry (ICP–MS, Agilent 7900) was carried out to determine the molar ratio between molybdenum and phosphorus. Elemental analysis was conducted in a CE Instruments EA1110 CHNS spectrometer to quantify the amount of carbon present in the samples and to calculate the molar ratio between molybdenum and nitrogen. Specific surface areas and porosity were determined by measuring the nitrogen physisorption at the critical temperature with an Autosorb MP1 instrument (Quantachrome). The specific surface area was calculated according to the Brunauer–Emmett–Teller (BET) theory in the range of Ps/P0 0–1.0. Before the measurements, the products were outgassed at 150 °C for 24 h under vacuum to ensure the complete removal of any previously adsorbed material. Thermogravimetric analysis (TGA) was carried out with a SETARAM Setsys thermobalance under a nitrogen atmosphere (from 35 to 700 °C with a heating rate of 10 °C min−1). Fourier transform infrared spectroscopy (FTIR) was performed with a Spectrum BX spectrometer from PerkinElmer and the absorption of KBr pellets of the samples was measured between 4000 and 400 cm−1. Light Transmission Measurements. Precipitation kinetics was studied monitoring the desupersaturation curves by turbidity measurements. Light transmission measurements were conducted in the dark with a selfmade setup comprised of a He–Ne red laser source (JDSU, model 1145P, 633 nm, 25 mW), a filter (Schott) that reduces the intensity of the transmitted light by a factor of 20, a collimator that aligns and centers the beam and a photodiode detector (photosensitive area =10×10 mm). The detected current was amplified by a transimpedance amplifier, coupled with a multimeter (Keithley 2700) connected to a computer for data acquisition, as in previous works.32, 33 Typical precipitation experiments were conducted in a glass jacketed reactor at 80 °C controlled by a thermostat (Lauda Ecoline Gold RE415) under magnetic stirring. The laser light passed through the reactor containing the initially transparent precursor solution. After the stabilization of the transmitted light signal, the precipitating agent (nitric acid) was added and this event established the time zero for the experiments. The values of intensity of transmitted light were normalized to the initial value. Catalytic Epoxidation of cis-cyclooctene. The catalytic activity of the ammonium phosphomolybdates was tested for the oxidation of cis-cyclooctene with tert-butyl hydroperoxide (TBHP) at 80 °C. The catalyst (11 mg), chloroform (9.7 mmol), chlorobenzene (4.6 mmol), and purified cis-cyclooctene (4.6 mmol) were placed in a vial, and the reaction was started by addition of TBHP (9.2 mmol). The closed vial was placed in a thermoshaker (HLC, MKR 23) at 80 °C and 500 rpm. Sampling volumes of 50 μL were collected at different 3



Page 4 of 18

times and diluted with 2.45 mL of chloroform to quench the reaction. To assure the separation of the catalyst from the reaction solution, the catalyst was let to sediment for 2 min before sampling. 50 μL of the chloroform solution were withdrawn and poured in a vial with 950 μL of dichloromethane to reach the concentration range suitable for the subsequent analysis. Chlorobenzene was employed as an internal standard. The samples were analyzed by a gas chromatograph coupled with a mass spectrometer (GC-MS, Shimadzu, GCMS-QP2010 Ultra) to follow the oxidation and to identify the products. The catalyst was recovered after the reaction and the powder was washed three times with chloroform. The oxidation of cis-cyclooctene was carried out for four additional cycles using the recovered powder to test the reusability of the catalysts. A blank reaction in the absence of the catalyst was carried out to demonstrate the catalytic activity of ammonium phosphomolybdate.

The experimental values of conversion were compared with those provided by the model ones by means of unweighted nonlinear least-squares fitting. Calculations were performed using the above-mentioned values of c and r so that the model depended only on the variable k, whose value was obtained by minimizing the leastsquares function. Once the value of k was known, the second-order rate coefficient k2 was estimated from eq. 4. The goodness of the fit was assessed from the nonlinear correlation coefficient (R) and the lack-of-fit index (LOF), expressed in %. This later index is defined as ∑4$01 − 0 21 &3 (6) ,-. = × 566 3 ∑4 0 1

Analogous catalytic experiments were carried out for selected catalysts with cyclohexene (4.6 mmol), styrene (4.6 mmol), and butyl methacrylate (4.6 mmol). The protocol used was identical to the one used with ciscyclooctene.

RESULTS AND DISCUSSION

Analysis of Kinetic Data. Reaction kinetics was followed recording the GC-MS peak areas of cis-cyclooctene (CyO) using chlorobenzene as an internal standard. The experimental conversion for

cis-cyclooctene = 1 − the expression:

= 1−

r, ,

was determined from

(1)

in which r and r, are the relative areas of cis-cyclooctene with respect to chlorobenzene at time t and t = 0. The rate law of epoxidation in the presence of tert-butyl hydroperoxide (TBHP) was determined by trial and error. Good results were obtained when analyzing the data with the following differential equation with a global second order: CyO (2) − = CyOTBHP The integration of eq. 2 over time in terms of leads to expression: # = 1− (3) $1 + #&' ( − 1 where k and c correspond to: (4) = )#CyO TBHP (5) # = ) *+ −1 CyO The ratio of the stoichiometric coefficient of TBHP with respect to cis-cyclooctene, r, was found to be equal to 1 from calibration of peak areas of cis-cyclooctene and its oxide. Consequently, the value of c was taken equal to 1, because all kinetic runs were conducted with a ratio [TBHP]t=0:[CyO]t=0 ratio of 2:1.

where 7 and 87 are the experimental and model calculated conversions. All calculations were carried out using an ad hoc routine wrote in Julia programing language.34

I. Bulk Crystallization from Solution: Effect of the Phosphate Source The precipitation of ammonium phosphomolybdate starts from an aqueous solution containing ammonium heptamolybdate and a phosphate source, after the addition of a strong acid (i.e., HNO3). In this highly acidic environment (pH < 0, see Table S1 in Supporting Information for pH measurements of all solutions involved), phosphomolybdic acid H3PO4(MoO3)12 is formed and the yellow ammonium salt hydrate (NH4)3PO4(MoO3)12∙4H2O precipitates from the solution. Differently from previous studies,4, 35 in the present work we investigate the use of different phosphate sources, both organic and inorganic. Five phosphate compounds (Scheme 1) were selected to evaluate any potential templating effect of organophosphates with respect to inorganic sources of phosphate, as well as the consequences of the phosphate precursor on the kinetics of precipitation and on the final features of the formed materials.

Scheme 1. Phosphate precursors used for the synthesis of ammonium phosphomolybdates. 4


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Two inorganic phosphate sources (i.e., NaH2PO4 and NH4H2PO4) and three organic ones (i.e., phytic acid dodecasodium salt hydrate, D-glucose-6-phosphate sodium salt, and O-phospho-DL-serine) were used. In the following, the ammonium phosphomolybdates obtained from the different precursors are labeled as APM-Na (from sodium dihydrogenphosphate), APM-NH4 (from ammonium dihydrogenphosphate), APM-phy (from phytic acid dodecasodium salt), APM-glu (from Dglucose-6-phosphate sodium salt), and APM-ser (from O-phospho-DL-serine). Differences in the precipitation rates were already macroscopically observable. APM-Na and APM-NH4 started to precipitate immediately upon acid addition and the reaction medium became deeply yellow. The reaction yield was high in both cases, as reported in Table 1. APM-phy precipitated also fast, but the intensity of the yellow color of the medium was slightly less pronounced and the amount of precipitate formed was much lower. In contrast, the reactions for APM-glu and APM-ser under the same conditions were clearly slower and the yield was still very low after 1 h of reaction. To quantify the precipitation rates, light transmission was measured by using a homemade setup registering the transmitted intensity of a laser beam passing through the reactor in which the crystallization took place. As shown in Figure 1, different induction times (i.e., the time between the mixing of the two solutions and the formation of a crystal that can be detected) are observed for the different systems, following the same trend observed macroscopically. The values of induction time, directly read from the curves and reported in Table 1, are very low for the systems containing inorganic phosphate sources (2.7 s for AMP Na and 0.6 s for APM-NH4) and one order of magnitude larger for the APM-glu and APM-ser systems (51.5 s and 61.5 s, respectively). Interestingly, when phytic acid sodium salt is used as a phosphate source (APM-phy system), the induction time is very similar to the inorganic phosphates (1.1 s). After the induction time, precipitation takes place and the transmitted intensity decreases rapidly until the medium becomes so turbid that the intensity value approaches zero. After subtraction of the induction time from the abscissa values, the desupersaturation curves of Figure 1 can be well fitted to simple exponential functions of the form:

9 = 9: + ' *(

(7)

where the preexponential factor A is fixed to 1 (because the intensity values in the curves are normalized with respect to the value before precipitation starts), 9: takes values close to zero, and k is the exponential decay constant(Table 1). The values of the desupersaturation constant k differ at least in one order of magnitude when comparing the inorganic sources with the organic APM_glu and APM-ser systems, corresponding to a significantly slower desupersaturation rate in the latter cases. Also in this case, the system APM-phy behaves differently from the other organophosphates, with a

desupersaturation constant k more than double than for the inorganic system APM-NH4. The different induction times and desupersaturation rates can be ascribed to the different availability of the phosphate groups before and at the moment of the precipitation. NaH2PO4 and NH4H2PO4 are inorganic salts that dissociate in water to give phosphate anions, while the organophosphates are phosphoesters in which the phosphate is covalently bound to the rest of the organic part through a C–O bond. In the organophosphates, phosphate groups become available only after hydrolysis of the C–O bond, as mentioned in the introduction.19, 20 Normalized transmitted intensity / a. u.

Page 5 of 18

1,0

APM-Na APM-NH4

0,8

APM-phy APM-glu APM-ser

0,6

0,4

0,2

0,0 0

5

50

100

150

Time / s

Figure 1. Bulk precipitation of the ammonium phosphomolybdate with different phosphate precursors monitored by turbidity measurements: NaH2PO4, NH4H2PO4, phytic acid dodecasodium salt hydrate, D-glucose-6-phosphate, O-phospho-DL-serine.

Under the reaction conditions, the dephosphorylation of D-glucose-6-phosphate sodium salt and O-phospho-DLserine requires a certain time to occur, which explains the induction times observed in the turbidity measurements. When the phosphate concentration reaches the required threshold level to form nuclei of a critical size, precipitation starts and the registered light intensity drops rapidly. The precipitation decreases the phosphate concentration in solution and shifts the hydrolysis equilibrium further to the product side. However, as judged from the low yield of the precipitation reaction, the hydrolysis takes place only up to a limited extent. In contrast, at pH 5 (pH value of the mixture ammonium heptamolybdate and phosphate) and 80 °C, phytic acid is hydrolyzed very fast.36 Since there are six phosphate groups (compared to the one single group of D-glucose-6phosphate sodium salt and O-phospho-DL-serine), the hydrolysis of the C–O bond of at least one group can easily occur, making phosphate groups available in the reaction medium. The yield of the precipitation is significantly higher for APM-phy than for APM-glu and APM-ser but remains far from the values of the inorganic salts (cf. Table 1), in which the availability of the phosphate is mainly determined by the solubility equilibrium of salt and not by the hydrolysis of the precursor. 5



Page 6 of 18

Table 1. Yield (after 60 min), induction time, and empiric parameters of the exponential fitting (cf. eq. 7 crystallization from solution of ammonium phosphomolybdate samples prepared from different phosphate sources (cf. Scheme 1). Exponential fitting of desupersaturation curves

Sample

Yield 60 min (%)

Induction time / s

APM-Na

87

2.7

(1.2 ± 0.2) × 10

APM-NH4

92

0.6

(2.9 ± 0.7) × 10

APM-phy APM-glu APM-ser

28 3 4

1.1 51.5 61.5

9:

−1

k/s −3

1.038 ± 0.005

−4

7.333 ± 0.005

−4

15.035 ± 0.002

−3

0.10893 ± 0.00012

−3

0.0806 ± 0.0012

(6.4 ± 0.2) × 10 (5.5 ± 0.2) × 10

(7.6 ± 0.2) × 10

SEM images of the different samples, presented Figure 2, indicate that the crystals produced have in all cases a rhombododecahedral morphology, but the particle sizes are remarkably different when comparing inorganic and organic phosphate sources (see more detailed discussion on the particles size in the next subsection, Table 2, and in Table S2 in Supporting Information). Inorganic phosphate sources lead to particle sizes one order of magnitude larger (average sizes ≥ 10 µm) than organophosphates (average sizes < 3 µm). The smaller sizes are consistent with the lower amount of phosphate in the medium (proved by the lower yields of the organophosphates). Phosphate acts as a limiting reagent and determines the growth of the ammonium phosphomolybdate crystals.

Fourier transform infrared (FTIR) spectra were also almost identical for all samples and presented the typical bands of the Keggin structure37 (Figure S2). The vibrational bands of P–O, Mo–O, and Mo–O–Mo are positioned at 1063, 963, and 866 cm−1, the P–O deformation vibration at 506 cm−1; and the NH4+ ions vibration band is assigned to the peak at 1407 cm−1.38

In the SEM images of the samples prepared from organophosphates, we sporadically observed holes that could point out to the formation of hollow or at least not completely solid structures. To rule out or to confirm this possibility, we cut a lamella of a crystalline representative sample (APM-glu) under the focused ion beam (FIB) and the corresponding micrograph is shown in Figure 3. The materials appeared to be solid and we did not observe any hint of hollow structures. High-resolution TEM micrographs of the lamella prepared under the FIB indicate clearly the presence of small nanocrystalline domains surrounded by a less crystalline matrix. At the border of the micrometric rhombododecahedrons, the density of the nanocrystalline domains appears to be higher than in the inner part, where the less crystalline matrix becomes more obvious (Figure 3(b)–(c)). Regardless of the phosphate source used, all the products presented the same X-ray diffraction (XRD) pattern, as shown in Figure 4. The reflections match the reference pattern of cubic (NH4)3PO4(MoO3)12∙4H2O (ICDD card no. 00-009-0412). All samples are highly crystalline already in very early stages (see as an example the diffractograms of samples recorded at different reaction times in Figure S1). Despite the chemical nature of the used phosphate precursors, the system evolves to isostructural phases, although the results cannot unequivocally confirm or exclude that all materials have exactly the same crystalline structure.

Figure 2. SEM micrographs of the ammonium phosphomolybdate particles after one hour of synthesis. The sample names are reported in the figures.

6


Page 7 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. (a) Focused ion beam micrograph of a lamella of sample APM-glu. (b) High-resolution TEM micrograph of a region close to a border of one of the observed rhombododecahedral microcrystals. (c) High-resolution TEM micrograph of an inner part of the same microcrystal.

7


ACS Applied Materials & Interfaces APM-Na

APM-NH4

APM-phy

I / a.u.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 18

APM-glu

APM-ser

10

20

30

2θ / °

40

50

60

Figure 4. X-ray diffractograms of the ammonium phosphomolybdate synthesized in bulk for 60 min with different phosphate precursors: the diffraction patterns are perfectly overlapping and matching with the reference pattern ICDD card no. 00-0090412 (red lines), corresponding to (NH4)3PO4(MoO3)12∙4H2O.

8


Page 9 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


II. Confined Crystallization within Nanodroplets and Comparison with Bulk Crystallization The inverse miniemulsion route was employed to study the effect of the space confinement introduced by water nanodroplets on the crystallization of ammonium phosphomolybdate. An inverse miniemulsion is composed of two phases: a dispersed (aqueous) phase and a continuous (organic) phase in which a lipophilic surfactant is dissolved. In our synthesis, commercial polyglycerol polyricinoleate (PGPR) was dissolved in cyclohexane and used as a continuous phase. PGPR was selected as a lipophilic surfactant because it has a low hydrophilic-lipophilic balance (HLB) (around 1.5),39 stabilizing efficiently inverse miniemulsions. The disperse phase contained aqueous solutions of our precursors. Ultrasound was applied to produce high shear forces to drive the miniemulsification of the two-phase system. As in the bulk solution systems, an aqueous solution of nitric acid was added to the miniemulsion to trigger the precipitation within the water droplets. The system was subsequently ultrasonified a second time to enhance the penetration of the acid in the water droplets. The different phosphate sources shown in Scheme 1 were also used in the miniemulsion approach. The starting miniemulsion of the salts was milky white in all cases after the first emulsification (as common when a transparent solution is miniemulsified). After the addition of nitric acid and the second sonication, the miniemulsion turned immediately intensively yellow for APM-Na-ME and APM-NH4-ME, paler yellow for APMphy-ME and lightly yellowish white for APM-glu-ME and APM-ser-ME (the label “ME” is used to denote the miniemulsion systems, in contrast to the bulk solution systems). APM-Na-ME = 68%; APM-NH4-ME = 72%; APM phy-ME = 32%; APM-glu-ME = 8%; APM-ser-ME = 8%). As in the bulk samples, although with a much smaller particle size, a rhombododecahedral morphology was obtained also in miniemulsion for the inorganic systems APM-Na-ME and APM-NH4-ME. This morphology became less evident when organophosphates were employed, as seen in the SEM and TEM micrographs presented in Figure 5 (additional HR-TEM micrographs in Figure S3, Supporting Information). Also in this case, with all the precursors, the resulting materials were crystalline and the typical reflections of the ammonium phosphomolybdate (ICDD card No. 00-009-0412) can be found in the corresponding XRD patterns (Figure 6). However, differently from the bulk solution, in the case of APM-glu-ME and APM-ser-ME, additional reflections ascribable to MoO3 hydrate (ICDD card no. 00-048-0399) and sodium phosphate Na2H2(PO3)4 (ICDD card no. 00-009-0100) are present. The appearance of these phases can be explained by the increased pressure within the nanodroplets.

Figure 5. SEM and TEM images (left and right, respectively) of the ammonium phosphomolybdate samples synthesized in miniemulsion with different phosphate precursors.

9


ACS Applied Materials & Interfaces (a) APM-Na-ME

(b) APM-NH4-ME

(c) APM-phy-ME

I / a.u.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 18

(d) APM-glu-ME

(e) APM-ser-ME

10

20

30

40

50

60

2θ / ° Figure 6. X-ray diffractograms of the ammonium phosphomolybdate synthesized with different phosphate precursors in miniemulsion at a reaction time of 60 min. The reflections match with the reference phase (NH4)3PO4(MoO3)12∙4H2O (ICDD card no. 00-009-0412, red continuous drop lines); (d) and (e) present reflections of MoO3 hydrate (ICDD card no. 00-048-0399, green dashed lines) and Na2H2(PO3)4 (ICDD card no. 00-009-0100, blue dashed lines).

The crystallization of inorganic materials in the confined space of miniemulsion droplets can occur at milder conditions (lower temperature and pressure) than typically required in bulk, as reported in previous works from our group.24-26 Generally, the Laplace pressure (i.e., the pressure difference between the inner and outer part of a curved surface caused by the surface tension at the liquid/liquid interface) is higher for small droplets than for bigger ones, which results in their disappearance in favor of the bigger ones (i.e., Ostwald ripening). This molecular diffusion is counteracted in miniemulsion by the presence of an osmotic pressure agent (i.e., a substance soluble in the disperse but not —or to a very limited amount— in the continuous phase, in our case the inorganic salts).22, 40 The interfacial tension between the two liquid phases (water with the salts and cyclohexane with PGPR) was experimentally determined

at 80 °C (i.e., the reaction temperature) with the spinning drop method and a value of 5 mN m–1 was obtained (see Table S3). The Laplace pressure in the droplets is given by the Young–Laplace equation: ∆?@>AB = DB )FMG?@BI

(8)

For model spherical droplets of 100 nm of diameter, the resulting value of ∆PLaplace is about 2∙105 Pa (2 bar), doubling the atmospheric pressure, which combined with the presence of the organophosphates can explain the formation of molybdenum oxide in the miniemulsion systems APM-glu-ME and APM-ser-ME. The average sizes determined by statistical treatment of electron micrographs, reported in Table 2, point out that the bulk samples prepared from organophosphates are significantly smaller (about one order of magnitude) than 10


Page 11 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the ones prepared from the inorganic phosphates, as explained above, while in the miniemulsion systems there is only a slight decrease in size. On the other hand, there is only a small difference in crystallite sizes between organic and inorganic phosphate sources, both in bulk and miniemulsion systems. Interestingly, by comparing crystallite sizes and particles sizes determined by electron microscopy, it can be concluded that the bulk particles are composed by several crystallites, whereas the size of the miniemulsion particles is almost comparable with the corresponding crystallite sizes. Table 2. Average particle size statistically measured from electron micrographs and crystallite size (L) calculated by applying the Scherrer formula for the XRD reflection (222) at 26.5° (reaction time of 60 min). Bulk solution synthesis

Miniemulsion

System

Average size / µm

L/ nm

Average size / nm

L/ nm

APMNa

15 ± 2

76

90 ± 20

44

APMNH4

10 ± 2

78

90 ± 20

43

APMphy

2.5 ± 0.6

67

70 ± 30

36

APMglu

0.9 ± 0.2

68

43 ± 5

39

APMser

1.4 ± 0.3

69

60 ± 20

34

Inductively coupled plasma mass spectrometry (ICP-MS) and elemental analysis were carried out to determine the molybdenum/phosphorus and molybdenum/nitrogen molar ratios for bulk solution and miniemulsion systems. Thermogravimetric analysis (TGA) was also performed to estimate the number of molecules of the hydration water present in the resulting products (see Supporting Information, Figures S4–5). The stoichiometry of the ammonium phosphomolybdate corresponding to the reference to which the XRD patterns have been ascribed is (NH4)3PO4(MoO3)12∙4H2O, which should correspond to a Mo/P ratio of 12, a Mo/N ratio of 4, and a hydration of 4 water molecules per formula unit. Table 3 presents the experimental values of Mo/P and Mo/N ratios and the hydration stoichiometries for the different samples. For all samples, the amount of molybdenum is higher than the expected stoichiometric values of the formula (NH4)3PO4(MoO3)12∙4H2O. Furthermore, the miniemulsion systems show a higher amount of molybdenum compared to the bulk solution systems. From the observation of the TEM micrographs in Figure 3, we had argued the presence of a less crystalline phase surrounding nanocrystals. This argument may explain, at least partially, the higher molybdenum content when compared to the theoretical stoichiometric one of the crystalline phase. It is also worth pointing out that samples produced from organophosphates may crystallize

under conditions of phosphate deficiency, because of the partial hydrolysis of the precursors. The non-negligible amount of carbon in the samples produced in miniemulsion from organophosphates is ascribable to the surfactant and to the unreacted organic precursors trapped in the materials during crystallization within the droplets. The hydration stoichiometries determined by TGA for bulk samples follow a different trend from those prepared in miniemulsion. In samples prepared in bulk from organophosphates, the calculated number of crystallization water molecules per formula unit is approximately 2, half than the expected stoichiometry (4), while, for the samples prepared from the inorganic phosphates, the value is slightly higher than 4. For samples prepared in miniemulsion, the crystallization water molecules are close to 5 for the inorganic systems and close to 4 for the sample prepared from phytic acid. For the other two organophosphates, the presence of additional phases prevents a reliable calculation of the stoichiometry. The prepared materials are exploitable as heterogeneous catalysts and as ion-exchanger for separation and purification. For these applications, the specific surface area and the porosity are key factors. Table 4 lists the specific surface area, the average pore size, and total pore volume, calculated according to the physisorption of nitrogen at the critical temperature following the Brunauer–Emmett–Teller (BET) theory. All samples are mesoporous, with pore sizes between 2 and 50 nm (Figure S6–7).41 The specific surface area is, in general, higher for the inorganic precursors, although the value of the system APM-phy-ME is also comparable to the inorganic systems. III. Catalytic Experiments The oxidation of cis-cyclooctene to cyclooctene oxide by tert-butyl hydroperoxide in chloroform was taken as a representative reaction to evaluate the catalytic activity of the synthesized ammonium phosphomolybdate particles. Figure 7 shows experimental and calculated conversions, and Table 5 lists the relevant fitting results. The statistical correlation coefficient (R) and the lack-of-fit index (LOF) indicate a very good correlation for most of the systems. It can be therefore accepted that the model predicts correctly the activity of the catalysts. The visual inspection of Figure 7 clearly reveals two different behaviors. The catalysts synthesized in miniemulsion were considerably more active than the ones prepared in bulk. If the ratio k2,ME/k2,bulk calculated for miniemulsion and bulk samples during the first epoxidation cycle is taken as an indicator of the activity of the fresh catalyst, the miniemulsion synthesis yielded catalysts with an activity enhancement of about 10–30 times, with the exception of the APM-NH4 system, which exhibited similar performances regardless of the preparation method. The enhanced performances of the miniemulsion samples can be explained by two main reasons: (i) their better dispersibility in organic solvents, 11



Page 12 of 18

Table 3. Stoichiometric analysis of the samples. Bulk solution synthesis

Miniemulsion

System Hydration [a] stoichiometry

Mo / P

Mo /N

APMNa

4.8

16.4 ± 0.3

5.3 ± 0.2

Colloidally Confined Crystallization of Highly ... - ACS Publications

Recommend Documents