A Rational Solid-State Synthesis of Supported Au–Ni Bimetallic

Research Article www.acsami.org

A Rational Solid-State Synthesis of Supported Au−Ni Bimetallic Nanoparticles with Enhanced Activity for Gas-Phase Selective Oxidation of Alcohols Wuzhong Yi,† Wentao Yuan,‡ Ye Meng,† Shihui Zou,†,‡ Yuheng Zhou,† Wei Hong,† Jianwei Che,† Mengjia Hao,† Bin Ye,† Liping Xiao,† Yong Wang,‡ Hisayoshi Kobayashi,§ and Jie Fan*,† †

Key Lab of Applied Chemistry of Zhejiang Province, Department of Chemistry, ‡School of Materials Science and Engineering, Zhejiang University, Hangzhou 310027, China § Department of Chemistry and Materials Technology, Kyoto Institute of Technology, Matsugasaki, Sakyo-ku, Kyoto 606-8585, Japan S Supporting Information *

ABSTRACT: A facile confined solid-state seed-mediated alloying strategy is applied for the rational synthesis of supported Au−Ni bimetallic nanoparticles (BMNPs). The method sequentially deposits nickel salts and AuNP seeds into the ordered array of extra-large mesopores (EP-FDU-12 support) followed by a high-temperature annealing process. The size, structure, and composition of the AuNi BMNPs can be well tuned by varying the AuNP seeds, annealing temperature, and feeding ratio of metal precursors. Kinetic studies and DFT calculations suggest that the introduction of the Ni component can significantly prompt the O2 activation on AuNPs, which is critical for the selective alcohol oxidation using molecular O2 as the oxidant. The optimal Au−Ni BMNP catalyst showed the highest turnover frequency (TOF) (59 000 h−1, 240 °C) and highest space-time yield (STY) of benzyl aldehyde (BAD) productivity (9.23 kg·gAu−1·h−1) in the gas-phase oxidation of benzyl alcohol (BA), which is at least about 5-fold higher than that of other supported Au catalysts. KEYWORDS: solid-state synthesis, Au−Ni bimetallic nanoparticle, activation of oxygen, gas-phase selective oxidation of alcohol, bimetallic nanophase diagram

1. INTRODUCTION

Alloying gold with a second metal offers numerous opportunities for modulating their electronic structures and optimizing their catalytic performance.8 The introduction of nonprecious metal Ni to Au, both through experiment and theoretical calculations, was proven to enhance Au’s oxygen adsorption and activation.9,10 We also performed DFT calculations on two model clusters, Au50 and Au30Ni20, to elucidate the adsorption energies (ΔE) of O2 and BA (Table S2). The adsorption energy (ΔE) of O2 on Au30Ni20 (−2.286 eV) is twice of that on Au50 (−1.117 eV), while the O−O bond length is considerably stretched from 1.327 Å on Au50 to 1.389 Å on Au30Ni20, confirming that alloying Au with Ni could facilitate the activation of O2. However, Au−Ni BMNPs, especially the size in sub-10 nm range, are difficult to fabricate due to their different electronegativities and the lattice mismatch through general wet-chemical methods.11 Also, Au−Ni BMNPs obtained after huge efforts are difficult to load into the channels of porous supports, suffering from aggregation during calcination and reaction processes under high temperature. High-temperature

Oxidation of primary and secondary alcohols to the corresponding aldehydes and ketones plays a fundamental role in organic synthesis, owing to the versatility of the carbonyl group as a building block.1 Among the catalysts explored for the solvent-free aerobic oxidation of alcohols, Au-based noblemetal catalysts often exhibit high conversion and high selectivity, so they seem to be a promising candidate for industrial application (Table S1).2,3 The space-time yield (STY) ranges from a very low 0.05 (nanoporous Au) to a moderate 0.9 kg·g−1metal·h−1 (Au/SiO2). Although many efforts have been made to improve their catalytic performance, these noble-metal catalysts are still mainly used in the lab. One main limitation of gold catalysts is their low affinity to molecular oxygen and inherent difficulty in oxygen activation.4,5 The challenge of how to alter the electronic structure of tiny Au entities and how to enable them to dissociatively chemisorb oxygen in the first place, and thus enable selective-oxidation chemistry, is a persistent one and one that must be addressed experimentally and theoretically.6,7 In addition, the gold catalysts are too expensive to be widely applied in practical applications. Improving the catalytic performance and minimizing the usage are important topics for gold catalysis. © 2017 American Chemical Society

Received: June 16, 2017 Accepted: August 29, 2017 Published: August 29, 2017 31853

DOI: 10.1021/acsami.7b08691 ACS Appl. Mater. Interfaces 2017, 9, 31853−31860

Research Article

ACS Applied Materials & Interfaces

obtained PdNPs were dispersed in 100 mL of chloroform. The TEM image and size distribution are shown in Figure S2d. Synthesis of Au−Ni Bimetallic Nanoparticles. A required amount of nickel nitrate and 500 mg EP-FDU-12 were added into 5 mL ethanol, and then stirred until ethanol vapored out. The as-obtained solid powder was further dried under vacuum oven before AuNPs deposition by colloid absorption method. At last, the supported powder was calcinated for 4 h under 5% H2/Ar at a specific temperature. The loading weight of Au is 1 wt % for all samples if there is no special illustration. The order of the two deposition steps is very important. If the AuNP loading is prior to the salt deposition, organic-ligand-protected AuNPs tend to aggregate due to the presence of polar ethanol solvent. 2.2. Catalytic Measurement. The catalytic activity for gas-phase selective oxidation of BA was determined using a fixed bed vertical microreactor (h = 250 mm, d = 12 mm). In order to avoid intrareactor gradients, the catalyst powder (10 mg) was diluted in 500 mg of quartz sand. A mass-flow instrument was used to control the flow rate of oxygen and nitrogen. BA (1.8 mL·h−1) was supplied through a syringe pump and was vaporized on the reactor wall prior to the catalytic bed. The temperature of the catalysts was controlled by a furnace, and the reaction temperature was continuously monitored by a thermocouple inserted into the catalyst bed. The products were analyzed by gas chromatograph with flame ionization detectors (GC-FID). The N selectivity was calculated as NBAD × 100%, where NBAD is the number

solid-state coreduction after impregnation is an alternative way to fabricate Au−Ni BMNPs within porous supports.12 However, NPs obtained through this method often present with wide size distribution and poor morphology control. To address these synthetic challenges, herein, we report a facile confined solid-state seed-mediated alloying strategy for the rational synthesis of supported Au−Ni BMNPs. The method sequentially deposits nickel salts and AuNP seeds into EPFDU-12 support followed with a high-temperature annealing process. The size, structure, and composition of the AuNi BMNPs can be well tuned by varying the AuNP seeds, annealing temperature, and feeding ratio of metal precursors.

2. EXPERIMENTAL SECTION 2.1. Catalyst Preparation. Synthesis of EP-FDU-12. EP-FDU-12 was synthesized according to previously published procedures.13 In a typical synthesis, 0.50 g of Pluronic F127, 1.25 g of KCl, and 50 mL of 1 M HCl were sequentially added into a 100 mL jacketed beaker, which was kept at 14 ± 0.1 °C. The mixture was stirred vigorously for 1 h before 0.7 mL of 1,3,5-trimethylbenzene was added. After 1 h, 2.08 g of tetraethylorthosilicate (TEOS) was poured into this solution. The mixture was kept stirring for 24 h to form a white suspension and transferred into an autoclave. After the hydrothermal treatment at 140 °C for 24 h, the reaction system was cooled to room temperature. The as-made product was obtained by filtration, washed with hot water, and dried in a vacuum oven at 30 °C. At last, the obtained white powder was annealed at 350 °C in air to remove the organic templates. Synthesis of AuNPs (3 nm). AuPPh3Cl (100 mg) was dissolved in 20 mL of toluene to form a clear solution, to which 400 μL of dodecanethiol was added. After the solution temperature rose to 55 °C, 84 mg of NaBH4 was then poured into the solution in one portion. The mixture turned black immediately and was stirred for 7 h at 55 °C. After the reaction system was cooled to 30 °C, the black precipitate was separated with ethanol by centrifugation and dried in a vacuum oven at 30 °C. Finally, the obtained AuNPs were dispersed in 100 mL of chloroform. The TEM image and size distribution are shown in Figure S2a. Synthesis of AuNPs (5 nm). AuPPh3Cl (100 mg) was dissolved in 20 mL of toluene to form a clear solution, to which 400 μL of dodecanethiol was added. After the solution temperature rose to 55 °C, 220 mg of borane-tert-butylamine complex (BTBC) was then poured into the solution in one portion. The mixture turned black immediately and was stirred for 7 h at 55 °C. After the reaction system was cooled to 30 °C, the black precipitate was separated with ethanol by centrifugation and dried in a vacuum oven at 30 °C. Finally, the obtained AuNPs were dispersed in 100 mL of chloroform. The TEM image and size distribution are shown in Figure S2b. Synthesis of AuNPs (8 nm). HAuCl4·4H2O (30 mg) and 10 mL of oleic amine were sequentially added into a 50 mL round-bottomed flask. The mixture was then heated to 110 °C and stirred to form a clear solution. After 30 min, the solution was cooled to 100 °C before 100 mg of BTBC was added. The reaction system turned black immediately. The black mixture was heated up to 140 °C and was stirred for 1 h. After the reaction system was cooled to 30 °C, the black precipitate was separated with ethanol by centrifugation and dried in a vacuum oven at 30 °C. Finally, the obtained AuNPs were dispersed in 100 mL of chloroform. The TEM image and size distribution are shown in Figure S2c. Synthesis of PdNPs (5 nm). PdNP is synthesized according to the reported method with a minor revision.14 In a typical synthesis, 100 mg of palladium(II) 2,4-pentanedionate and 30 mmol of oleic amine were sequentially added into a 50 mL round-bottomed flask. The mixture was then heated to 110 °C before 200 mg of BTBC was added. After that, the reaction system was heated to 140 °C immediately and was stirred for 2 h. After the reaction system was cooled to 30 °C, the black precipitate was separated with ethanol by centrifugation and dried in a vacuum oven at 30 °C. Finally, the

BA

of produced BAD molecules, and NBA is the number of reacted BA molecules. A kinetic study was conducted under a nondiffusion controlled zone, which was confirmed by the similar activities at equivalent contact times (0.011−0.023 g·s·mL−1). Oxygen and BA partial-pressure-dependence experiments were conducted at 240 °C, and N2 was used as a dilute gas to ensure the same contact time (0.013 g·s·mL−1). The gas-phase oxidation of other alcohols (4-methoxybenzyl alcohol, phenylethyl alcohol, 1-hexanol, 1-heptanol, and 1octanol) was carried out in the same fashion. 2.3. Catalyst Characterization. Wide-angle X-ray diffraction (WA-XRD) measurements were performed on a Rigaku Ultimate IV with Cu Kα radiation. Transmission-electron microscopy (TEM) analysis was conducted on a JEOL JEM-1230 operated at 80 kV. Highangle annular dark-field scanning transmission-electron microscopy (HAADF-STEM) images and elemental mapping were recorded on the FEI TITAN Cs-corrected ChemiSTEM equipped with an energydispersive X-ray spectroscope (EDS). The sample was embedded in epoxy resin and then microtomed into sub-100 nm ultrathin films at room temperature. These thin films floated on water and were collected by copper mesh for TEM and HAADF-STEM analysis. X-ray photoelectron spectroscopy (XPS) measurements were performed in a VG Scientific ESCALAB Mark II spectrometer equipped with two ultrahigh vacuum (UHV) chambers. All binding energies were referenced to the C 1s peak at 284.6 eV of the surface adventitious carbon. The Brunauer−Emmett−Teller (BET) surface area of samples was analyzed by nitrogen sorption isotherms using the Micromeritics ASAP 2020 nitrogen-adsorption apparatus. All samples were outgassed under vacuum at 200 °C prior to nitrogen-adsorption measurement. The actual contents of Au and Ni in the prepared catalysts were measured by inductively coupled plasma mass spectrometry (ICP-MS) on a Plasma-Spec-II spectrometer. 2.4. DFT Calculations. Catalysts were modeled by Au50 and Au30Ni20 clusters on which O2 or PhCH2OH was adsorbed. A few orientations of adsorbates were examined, and typical structures are shown in Figure S1. Unit-cell size was a rectangular parallelepiped of 22.66 × 20.42 × 30 Å. DFT calculations with the periodic boundary conditions were carried out using a plane-wave-based program, Castep.15,16 The Perdew−Burke−Ernzerhof (PBE) functional17 was used together with the ultrasoft-core potentials.18 The cut-off energies were 300 eV for geometry optimization and 340 eV for the post energy calculation. For the modeling of Au30Ni20 cluster, Au atoms were randomly replaced by Ni atoms, and then the structure was optimized. The electronic configurations of atoms were H: 1s1, C: 2s22p2, O: 2s22p4, Ni: 3d84s2, and Au: 5d106s1. 31854

DOI: 10.1021/acsami.7b08691 ACS Appl. Mater. Interfaces 2017, 9, 31853−31860

Research Article

ACS Applied Materials & Interfaces

3. RESULTS AND DISCUSSION The formation of Au−Ni bimetallic nanoparticles is monitored by XRD, HAADF-STEM, and EDS elemental mapping at different treatment temperatures. Taking Au1Ni1 as an example, as Figure 1a shows, preloaded nickel nitrate and AuNPs with an

according to the Au−Ni bimetallic phase diagram, phase separation tends to result in a gold-rich phase and a nickel-rich phase when the content of Ni is between 0.12 and 0.98. Due to the lower surface energy of Au (1130 dym·cm−1) than that of Ni (1725 dym·cm−1), surface Ni atoms tend to diffuse into the Au subsurface regions, while Au atoms segregate to the topmost surface when annealing in reductive atmospheres.20 As a result, Au dominates the particle surface to form an Au-rich shell, which is confirmed by EDS mappings (Figure 2). In addition,

Figure 2. Schematic representation of the structure evolution at 500 °C as a function of feeding ratio and their corresponding HAADFSTEM EDS mappings, HAADF-STEM pictures, and corresponding line-scan patterns. Scale bar is 2 nm. AuNPs (5.1 ± 0.2 nm) were used as seeds. Figure 1. HAADF-STEM EDS mappings of (a) Au1Ni1-RT, (b) Au1Ni1-300 °C, (c) Au1Ni1-500 °C sample; (d) XRD patterns; (e) schematic representation of the structure evolution as a function of increasing thermal-annealing temperature and their corresponding HAADF-STEM EDS mappings (scale bar is 2 nm). AuNPs (5.1 ± 0.2 nm) were used as seeds. (Note: Ni signal of first particle in 1e is very weak due to the low loading concentration and the highly dispersed nature of nickel nitrate, and we do not rule out that the Ni signal comes from background noise.)

the corresponding line-scan of a single particle exhibits a broad peak for Ni located at the center of the profile and two intensive peaks for Au on both sides, clearly indicating a Nirich@ Aurich core−shell structure. Au(111) peak shifts from 38.2° to 38.5° after reduction of the Au1Ni1-RT sample at 500 °C, indicating the formation of an Au-rich surficial alloy, in agreement with the line-scan result. The difference in the surface tension of Au and Ni can be overcome by further increasing the temperature to 700 °C. The EDS mappings of an individual Au1Ni1 NP show that Au and Ni originate in the same spatial area with a spherical shape suggesting that the NPs possess an alloy structure with a homogeneous distribution of Au and Ni elements. In addition, the lattice fringe of a typical Au1Ni1-700 °C NP displays an interplanar spacing of 0.232 nm (Figure S3), which is in good agreement with the XRD result (2θ = 38.8°), further confirming the alloy structure. The schematic representation of the structure evolution (through Ni2+ ions and AuNPs to Au−Ni alloys) as a function of increasing thermal annealing temperature is shown in Figure 1e. It was reported that Au−Ni BMNPs with different structures could be fabricated according to the Au−Ni phase-separation mechanism.21,22 In addition to adjusting the composition and structure of AuNi BMNP by the annealing temperature, they can also be tuned by varying the feeding ratio of the AuNPs and Ni salts. According to phase-separation rule, compositions of

initial molar ratio of 1:1 were uniformly dispersed throughout EP-FDU-12 before the annealing processes (denoted as Au1Ni1-RT). After thermal treatment at 300 °C under H2/Ar, part of the Ni components are concentrated on the surface of the AuNPs as revealed by EDS mapping (Figure 1b). The actual Au/Ni molar ratio collected by a number of AuNi BMNPs is about 2.1 ± 0.5. Since there is no obvious shift of the Au(111) diffraction peak (Figure 1d), we believe that the Ni component just deposits onto the surface of AuNPs (denoted as Au@Ni), which played the role of the seeds for the reductive growth of most Ni components. Also, Au NPs can facilitate the reduction of Ni2+ ions.19 Meanwhile, some monometallic Ni clusters formed through self-nucleation and growth (denoted as Ni-mc, orange circle in Figure 1b) since there were no AuNPs nearby. As the temperature further increases to 500 °C, 31855

DOI: 10.1021/acsami.7b08691 ACS Appl. Mater. Interfaces 2017, 9, 31853−31860

Research Article

ACS Applied Materials & Interfaces

500 °C in two-phase region, Au−Ni BMNPs with the core− shell structure with varying Au-rich shells could be obtained. Also, through adjusting the annealing temperature, the Au−Ni nanostructure form, ranging from immiscible Au@Ni to the partially miscible Ni-rich@Au-rich core−shell and then to the Au−Ni alloy structure, could be tuned. Particle size is another important factor of catalytic activity. In addition to the facile size control through the confining effect of the mesoporous EP-FDU-12 support,25 the final size of the Au−Ni BMNPs can also be well tuned by the size of the AuNP seeds, further confirming the seeding-growth mechanism. By using 5.1 ± 0.2 nm Au NPs as seeds, Au1Ni1-500 °C BMNPs were highly dispersed throughout the EP-FDU-12 supports and were of uniform size with narrow distribution, 5.3 ± 0.4 nm (Figure 4a−c), which is in agreement with the XRD

the two separate portions will lie on opposite ends of the tie line, and their relative content varies with the ratio of the two elements.23 In this case, the content of the two separate phases (Au-rich shell and Ni-rich core) varies with the feeding ratio of Au/Ni. Namely, the shell thickness of Au−Ni core−shell NPs could be tuned via changing the initial ratio of Au and Ni in the two-phase area. To confirm this, we prepared a series of Au−Ni BMNPs with different Au/Ni ratios and collected their HAADF-STEM EDS mappings, HAADF-STEM pictures, and corresponding line-scan patterns for comparison (Figure 2). From EDS mappings, we can clearly find that in comparison to Au1Ni1-500 °C, Au2Ni1-500 °C owned a thicker Au-rich shell, while Au1Ni2-500 °C owned a thinner Au-rich shell. For Au2Ni1, Au1Ni1, and Au1Ni2, line scanning exhibited a broad peak for Ni located at the center of the profile and two intensive peaks for Au on both sides, clearly indicating the core−shell structure. Also, the reduced peak width of Au signal further confirms the shrinking of the gold shell during the structure evolution. Alternatively, as the initial ratio of Au and Ni increased from 2:1 to 3:1, Au−Ni BMNPs moved from the two-phase region to the single-phase region at 500 °C, with Au3Ni1-500 °C showing an alloy structure. The schematic representation of the structure evolution of the Au−Ni BMNPs as the Ni/Au ratio increased at 500 °C is shown in the third row of Figure 2. More ESD mappings and the corresponding XRD patterns of these particles are presented in Figure S4 and Figure S5. Since smaller-sized particles favor the formation of singlephase alloy, their nanophase diagrams usually differ from those of the bulk phase.23 According to the Au−Ni bimetallic bulkphase diagram,24 Au1Ni1-700 °C and Au3Ni1-500 °C, the actual Au/Ni ratio of which was 1.2 ± 0.2 and 4.2 ± 0.3 according to EDS elemental analysis, respectively, were located in two-phase region. EDS mapping results, however, demonstrated homogeneous alloy structures, indicating that the Au−Ni nanophase boundary curve shifts to the lower-temperature region as compared to the corresponding bulk-phase diagram. The Au@ Ni core−shell structure of Au1Ni1-300 °C indicates that Au and Ni are immiscible at 300 °C though only in the nanophase. Coupled with the fact that phase separation happened to Au2Ni1-500 °C (actual Au/Ni ratio is 3.5 ± 0.3), we drew an approximate nanophase boundary curve (Figure 3), and we are certain that a more precise curve could be achieved once we acquired enough critical points. A good knowledge of nanophase diagrams is helpful for us to design and fabricate BMNPs. For example, by simply changing the feeding ratio at

Figure 4. (a,b) High-resolution transmission-electron microscopy scanning transmission-electron microscopy (HRTEM-STEM) image and (c) size distribution of Au1Ni1-500 °C by using 5.1 ± 0.2 nm AuNPs as seeds; (d) XRD pattern and corresponding size of Au1Ni1500 °C by using 3.2 ± 0.3 nm (black line), 5.1 ± 0.2 nm (red line), and 7.8 ± 0.5 nm (blue line) AuNPs as seeds.

results (5.6 nm). When the size of Au seeds changed to 3.2 ± 0.3 nm or 7.8 ± 0.5 nm, BMNPs of 3.8 and 8.1 nm were obtained, respectively (Figure 4d). The actual metal-loading weight was examined by ICP-MS (Table S3), and the content of Ni and Au was very close to that of the feeding, indicating a precise control of our synthesis strategy. The similar nitrogen sorption curve and surface area data for all catalysts suggested that the deposition of Ni and Au did not obviously change the textural structure of the mesoporous support (Figure S6, Table S4). Nitrogen sorption measurement also suggests that as-prepared catalysts are fully accessible to gas molecules, which is necessary for a heterogeneous catalytic process to proceed smoothly. X-ray photoelectron spectroscopy (XPS) was then used to reveal the electronic interaction between Au and Ni (Figure S7). The binding energy of Au 4f7/2 gradually shifted from 83.6 to 84.0 eV as the Ni/Au ratio increased from 0 (Au) to 2 (Au1Ni2), indicating the generation of surface Auδ+ species, which was reported to benefit the oxidative reactions.26 Since superficial Ni tends to be oxidized under an oxygen atmosphere, the XPS

Figure 3. Au−Ni bimetallic phase diagram from bulk scale to nanoscale. 31856

DOI: 10.1021/acsami.7b08691 ACS Appl. Mater. Interfaces 2017, 9, 31853−31860

Research Article

ACS Applied Materials & Interfaces

Table 1. Structure Parameters of Au−Ni Catalysts and Their Corresponding Catalytic Performance in BA Oxidation Using O2 as Oxidanta entry

catalystb

sizec (nm)

Au/Ni ratiod

n(O2)e

n(BA)f

Ea (kJ·mol−1)

conversion (%)

TOFg h−1/1000

STYh kgBAD·gAu−1·h−1

1 2 3 4 5 6 7 8 9 10 11

Ni Au (Ni+Au)i Au/NiO Au Au3Ni1 Au2Ni1 Au1Ni1 Au1Ni2 Au1Ni1 Au1Ni1

5.8 5.8 5.9 4.0 5.5 5.6 5.6 5.8 3.8 8.1

± ± ± ± -

0.68 0.54 0.51 0.48 0.40 -

0.09 0.23 0.30 0.34 0.53 -

71.7 67.3 72.2 73.5 75.6 -

99% in all tests; carbon balance was >99% in all tests. Reaction conditions: 30 mL·min−1 of O2, 1.8 mL·h−1 of BA, 10 mg of catalyst diluted with 500 mg of quartz; N2 was used as diluted gas in partial pressure dependence tests. Reaction temperature was 240 °C. a

spectra suggested only a Ni2+ oxidation state in Au−Ni BMNPs samples. Ni 2p3/2 XPS peaks are best fitted by the presence of Ni(OH)2 (856.4 eV) and NiO (853.6 eV),27 with their relative intensity varying for different samples. The content of Ni(OH)2 was 100% for Au3Ni1-500 °C, 77% for Au2Ni1-500 °C, 34% for Au1Ni1-500 °C, and 40% for Au1Ni2-500 °C. Transition-metal oxides or hydroxides on the surface of noble-metal NPs not only increases density of the metal/oxide interface but also alerts the adsorption and activation of reactant molecules, leading to unique catalytic activity.28,29 The gas-phase oxidation of BA was used as a probe reaction to evaluate the catalytic properties of the as-synthesized Au−Ni BMNPs catalysts. The structure parameters of Au−Ni catalysts and their corresponding catalytic performance are presented in Table 1, and detailed catalytic data is shown in Figure S7. It is obvious that the Au−Ni BMNPs catalysts exhibit much higher catalytic activity than monometallic Au catalysts. Due to the low loading weight of Ni (0.3 wt %, the same with that in Au1Ni1 sample), low catalyst mass (10 mg), and high reactant flow (1.8 mL·h−1), only a trace reaction (conversion