Monolayer-protected quantum clusters (QCs) composed of a few atoms show unique properties due to their novel atomic and electronic structure. Their discrete electronic states produce well-defined luminescence in clusters, such as Au₂₅, Au₂₃, Au₂₂, Au₈, etc. 1 2 3 4 5 6 7 . They have attracted the attention of various fields such as sensors, biolabels, live cell-targeted imaging 4 , single molecule electroluminescence 8 , opto-electronics 9 , and catalysis 10 . In the case of Au₂₅, single crystal X-ray analysis has shown that it has an Au₁₃ core protected with six [Au(SR)₂] units in a core-shell-like pattern. The icosahedral Au₁₃ core has 20 triangular faces. However, only 12 facets are face-capped by the exterior 12 Au atoms which keep eight facets open 1 . These "hole" sites may be useful as active sites 11 which may participate in catalytic processes. Although the single crystal structures of the most of the clusters are yet to be solved, it is expected that many of them contain active sites. In this study, electron transfer properties of Ag₇ and Ag₈ clusters were explored using a simple reduction reaction, namely, the conversion of nitro to amino group, in different substrates. Catalysis using noble metals has caused great excitement after the initial report of Haruta 12 . It is known that reduction potential of silver nanoparticles change with size and shifts to more negative values with decrease in size, which are in good agreement with previous predictions 13 . QCs have very high negative reduction potential in comparison to bulk, and this makes them useful for the catalysis of electron transfer reactions. Silver is less expensive than other noble metals, and it was reported as an efficient catalyst for several reactions. For example, silver on alumina is a promising catalyst for selective catalytic reduction of NO by hydrocarbons from automobile exhausts 14 . It has been reported that Ag nanoparticles on hydroxyapatite in the presence of water catalyze the selective oxidation of various phenylsilanes into phenylsilanols 15 and also that the same system was found to be a highly efficient catalyst for the selective hydration of nitriles to amides 16 . It has been reported that chloroanilines were produced in large scale by the hydrogenation of chloronitrobenzenes using Ag nanoparticles on SiO₂, and the system also showed the size-dependent catalytic activity 17 . Formation of subsurface oxygen on the catalyst, enhanced in the case of supported silver clusters, plays an important role in CO oxidation 18 . Ag clusters supported on alumina are better active when compared to platinum supported on alumina for oxidant-free alcohol dehydrogenation to carbonyl compounds, and clusters less than 1 nm show structure-sensitive reactions 19 . It is reported that alumina is the best support for Ag, and it shows better selectivity for oxidation of ammonia to form nitrogen at low temperatures 20 . Silver nanoclusters on TiO₂ enhance the reduction of bis (2-dipyridyl) disulfide to 2-mercaptopyridine in the presence of water 21 . It has been reported that the selectivity for hydrogenation of crotonaldehyde was very high in the presence of Ag catalyst below 3-nm diameter; the catalyst was also structure sensitive 22 . Shimizu et al. 23 reported that clusters of silver on alumina catalyze the cross-coupling reactions of alcohols, direct amide synthesis from alcohols and amines 24 , and chemoselective reduction of a nitrostyrene with size-dependent catalytic activity 25 . The highest yield was obtained with particles of size ranging from 0.9 to 30 nm for the N-alkylation of anilines with benzyl alcohol for which silver shows high selectivity compared to other catalysts due to less stable metal-hydride bond formation 26 . Several reports are available for the reduction of nitro groups using nanoparticles 27 28 29 30 31 32 . Pal et al. reported the reduction of 4-nitrophenol (4-np) using silver nanoshells stabilized with cationic polystyrene beads 27 as well as silver deposited on silica gel 28 . It was reported that gold nanoparticles containing membranes reduce aromatic nitro compounds 29 . Ag and Au nanoparticles grown on calcium alginate gel beads are found to catalyze nitrophenol reduction 33 .

In this article, we studied the catalysis of supported QCs of Ag_7,8 using 4-np as the model system. Ag₇ and Ag₈ are new QCs prepared efficiently by the interfacial method 34 . Similar silver clusters are also made by other routes 35 . The reduction reaction occurs with a rate constant of 8.23 × 10^-3 at 35°C, and the TOF measured was 1.87 s^-1 per cluster. Performance of various supports has been evaluated.

The as-synthesized QCs of Ag were characterized by optical absorption studies. The corresponding absorption spectra are shown in Figure 1a, and the peak around 550 nm is due to HOMO-LUMO transitions from the 4d valence band to the 5sp conduction band-derived states of QCs 34 38 39 . These were observed after 48 h of etching process 34 and the absence of nanoparticles was confirmed using high resolution transmission electron microscopy of the etched materials. This confirmed that the peak around 550 nm was due to clusters. Inset of Figure 1a shows the photoluminescence spectrum of Ag_7,8 QCs having an excitation at 675 nm and emission at 772 nm; these data were collected at 273 K, and the photograph in the inset corresponds to red emission from the as-prepared crude mixture of the QC solutions under UV light irradiation. Figure 1b shows that the emission intensity of the solution decreased with the addition of alumina powder. As QCs got coated on alumina, the concentration of QCs in the solution decreases, and finally, the solution turned colorless as shown in the inset of Figure 1b.

Polyacrylamide gel electrophoresis of the as-prepared clusters showed two bands. This indicated the presence of two clusters, Ag₇ and Ag₈ in the crude solution, which were separated and dissolved in water 34 . The solutions exhibited strong emission. Except for these purified clusters, the excitation and emission are different from the crude clusters without separation (shown in the inset of Figure 1a). The clusters have the molecular formulae, Ag₈MSA₈ and Ag₇MSA₇, but are described merely as Ag₈ and Ag₇. The crude cluster is a 80:20 mixture of Ag₈ and Ag₇ which were used for this study. Therefore, we refer to the clusters as Ag_7,8.

Additional file 1, Figure S1A shows the HRTEM image of the QCs supported on alumina. It was observed that silver QCs were uniformly coated on alumina. They were highly sensitive to the electron beam, and started to fuse and became nanoparticles upon continuous exposure. The lattice fringes with an interplanar spacing of 2.4 Å correspond to Ag (111) (Figure S1B in Additional file 1), indicating the formation of nanoparticles. Corresponding EDAX shows the presence of silver on alumina, as shown in Additional file 1, Figure S1C-E.

SEM image and EDAX spectrum of Al₂O₃@Ag_7,8 are shown in Figure 2. Elemental maps using appropriate lines are also shown. The spectrum and images confirmed the presence of silver on alumina. It is clear that silver is uniformly coated on alumina. Elemental mapping confirmed the presence of other elements such as sulphur, oxygen, and carbon quantitatively on the alumina matrix. Nearly, 1:1 ratio of Ag:S is observed in the sample. It is confirmed from the data that Ag QCs are protected with MSA.

Further confirmation of the presence of Ag QCs in the supported material was available from LDI-MS as shown in Figure 3. It gives characteristic features of Ag_nS_m ^-. Laser irradiation at 337 nm cleaves AgS-C, bond and Ag_nS_m clusters alone are observed in the gas phase. The intact chemical composition of the QC was not observed, as typical of the spectra of thiolated clusters 34 38 . Ag_nS_m ^- species observed in the gas phase are dissociation products as well as gas phase reaction products. Intact clusters, along with the monolayers, are seen only in MALDI-MS and ESI-MS of the free clusters 34 . One of the interesting aspects of silver is its isotope distribution, which helps us to unambiguously assign the ions. To illustrate this, the experimental isotope pattern of one cluster fragment is compared with its theoretical pattern in the inset of Figure 3. It may be noted that the clusters do not fragment upon adsorption on the alumina surface, as properties of the clusters such as luminescence are retained on the oxide surface.

Aqueous solution of 4-np shows characteristic absorption maximum at 317 nm due to the n → π* transition (Figure 4) 40 41 . Upon addition of freshly prepared ice-cold aqueous NaBH₄ solution, the peak position of 4-np red shifted to 400 nm. This indicates the formation of 4-nitrophenolate ion in alkaline solution. The color of the solution deepened (from pale yellow to deep yellow). Without the addition of clusters, reduction was not observed as seen from the retention of the color. Even for several days, the peak at 400 nm due to 4-nitrophenolate ion remained unaltered.

With the addition of QCs supported on alumina, fading and ultimate leaching of the dark yellow color due to phenolate ions occurred, and brown color of 4-ap appeared. 4-nitrophenolate ion peak at 400 nm got reduced and within 10 min, a new peak around 295 nm appeared due to 4-ap 42 43 . The spectrum of 4-ap was verified with that of a standard sample. Reduction can be visualized with the color change, and it was almost complete which was authenticated by the optical absorbance value of 4-ap. Excess amount of reductant NaBH₄ was used, and therefore, a pseudo-first-order rate equation may be considered. As a result of adsorption of 4-np and BH₄ ^- on the cluster surface, electron transfer from donor BH₄ ^- to the acceptor 4-nitrophenolate ion is facilitated. The reduction was carried out at three different temperatures; 15, 25, and 35°C (Figure 4a,b,c). It was observed that at lower temperatures, the time required was high. Isobestic point observed during the transformation is shown in Additional file 2, Figure S2.

Figure 5A shows the variation of concentration with time of 4-nitrophenolate ion at different temperatures, 15, 25, and 35°C. The rate constants of the reaction (Additional file 3, Table 1) are plotted against 1/T in Figure 5B. Corresponding increase in product concentration is shown in Additional file 4, Figure S3A and B gives the ln(k) versus 1/T plot for the product formed.

The plot of ln(k) versus 1/T (Figure 5b) yields a straight line. The activation energy was found to be 55.6 kJ mol^-1 at 298 K. The TOF was 1.87 s^-1 per cluster. Although the calculated TOF is comparable to that reported for Ag nanoparticles (which is 1-2) 32 , we note that certain number of atoms of the clusters are not accessible in the catalytic process as they are used for surface binding. This aspect reduces the available number of surface atoms and increases the TOF per cluster.

Catalysts remain active at the end of the reaction, and these were separated from the product. Again, a fresh batch of 4-np was added to the used catalyst. Fresh reducing agent was not needed till the completion of four cycles, but subsequent cycles required fresh BH₄ ^-. The reaction followed the same kinetics as mentioned above. In this way, four consecutive fresh batches of 4-np were reduced with the same batch of catalyst. Reusability of the same batch of catalyst for the reduction cycles was tested; it remained active for ten cycles. The data for the second, third, fourth, and fifth reduction cycles are shown in Additional file 5, Figure S4. The product obtained was characterized with positive ion ESI-MS as shown in Additional file 6, Figure S5. Formation of 4-ap was confirmed with a peak at m/z 109, and the precursor 4-np peak at m/z 139 also got disappeared. Although the 4-np peak due to protonated 4-np (m/z 110) was not detected (which is the major peak in the pure sample), the molecular ion (m/z 109) is seen in the product. The peak at m/z 109 in the parent compound, 4-np, is due to the loss of NO 44 . This peak in the product spectrum is not due to the presence of unreacted substrate 4-np as its molecular ion signature at m/z 139 disappeared completely.

XPS investigation of the catalyst was carried out before and after the reaction. Survey spectra of Al₂O₃@Ag_7,8 before and after the reaction are shown in Additional file 7, Figure S6, and the expanded regions are shown in Figure 6. Spectral shift due to charging was corrected with respect to C 1s at 285.0 eV. Expanded spectra in the Ag 3d region show binding energies of 367.9 and 374.0 eV due to Ag 3d_5/2 and Ag 3d_3/2, respectively of Ag (0). S 2p shows a peak at 161.7 eV, Al 2p shows a peak at 74.6 eV, and O 1s appears at 530.8 eV. All the data correspond to the fresh catalyst. The O 1s position indicates hydroxyl groups at the surface, as expected. The C 1s region shows two peaks at 285.0 and 288.3 eV, corresponding to the CH/CH₂ and -COO^- groups. After three cycles, Ag 3d shows peaks at 367.9 and 374.1 eV; corresponding to Ag (0). Al 2p, O 1s, and S 2p did not change significantly. The C 1s region shows a reduction in the peak intensity of the -COO^- feature. Reduction in the intensities of sulfur and carbon is noticed. This indicated a slight desorption of the MSA monolayer.

The same experiment was performed with Ag_7,8 QCs (0.5 mg/0.5 ml) alone (unsupported) in the presence of NaBH₄, and complete reduction happened within 3 min. The same reduction reaction was repeated with Ag@citrate nanoparticles of 40 nm core diameter supported on alumina. For the supported Ag@citrate nanoparticles (10% loading, 50 mg), the reduction time increased thrice when compared to QCs. It appears that the electron transfer reaction depends upon the surface area of the catalyst, besides the electronic effect. This follows pseudo-first-order kinetics as shown in Additional file 8, Figure S7. The reduction time was reduced with the QCs compared with the nanoparticles.

Other nitro compounds such as 3-na, 4-na, and 3-np were tested with supported QCs (Additional file 9, Figure S8). The peaks at 225 and 358 nm indicate the progress of the reduction of 3-na. The reduction in the peak at 380 nm of 4-na indicates the progress in the reduction with the appearance of the new peaks at 240 and 305 nm. The peaks at 330 and 270 nm reduced with the appearance of the peak at 290 nm during the reduction 3-np.

QCs, Ag₇, and Ag₈ were supported on various substrates to prepare catalysts, such as Al₂O₃@Ag_7,8, SiO₂@Ag_7,8, TiO₂@Ag_7,8, and Fe₂O₃@Ag_7,8. Such catalysts show enhanced catalytic activity for the reduction of several nitro compounds. Detailed studies were performed with Al₂O₃@Ag_7,8. The pseudo-first-order rate constant was found to be twice larger than the supported silver of 3.29% loading on an anion exchange resin 27 . The rate constant was found to be 8.23 × 10^-3 s^-1, and the activation energy was 55.6 kJ mol^-1 at 298 K. Other nitro aromatics such as 3-np, 3-na, and 4-na were also investigated. The results suggest that the cluster system is a better catalyst for the reactions investigated.

ITO: indium tin oxide; MSA: mercaptosuccinic acid; QCs: quantum clusters; SEM: scanning electron microscopy.

The authors declare that they have no competing interests.

AL conducted the experiments and drafted the manuscript. TUB synthesized the quantum clusters. TP conceived the study, and participated in its design and coordination. All authors read and approved the final manuscript.

HRTEM-EDAX spectrum and images of Al₂O₃@Ag_7,8, isobestic point in the UV-vis spectra of the reduction of 4-np, increase in concentration of 4-ap during 4-np reduction reaction with corresponding plot of ln(k) versus 1/T, reusability of the supported Al₂O₃@Ag_7,8 for the reduction of 4-np, positive ion ESI-MS of the product, XPS survey spectra of Al₂O₃@Ag_7,8 after reactions (first and third cycles), UV-vis spectra of the reduction reaction using Ag@citrate nanoparticles with corresponding plot of concentration versus time, variation of the spectral intensities of other nitro aromatics during reduction and UV-vis spectra for the reduction of 4-np with SiO₂@Ag_7,8, TiO₂@Ag_7,8 and Fe₂O₃ @Ag_7,8. Supplementary data pertaining to with this article can be found, in the online version, in Additional files 1, 2, 3, 4, 5, 6, 7, 8, 9 and 10.