Research ArticleChemistry

Unraveling the long-pursued Au144 structure by x-ray crystallography

See allHide authors and affiliations

Science Advances  12 Oct 2018:
Vol. 4, no. 10, eaat7259
DOI: 10.1126/sciadv.aat7259


The transition from nanocluster to nanocrystal is a central issue in nanoscience. The atomic structure determination of metal nanoparticles in the transition size range is challenging and particularly important in understanding the quantum size effect at the atomic level. On the basis of the rationale that the intra- and interparticle weak interactions play critical roles in growing high-quality single crystals of metal nanoparticles, we have reproducibly obtained ideal crystals of Au144(SR)60 and successfully solved its structure by x-ray crystallography (XRC); this structure was theoretically predicted a decade ago and has long been pursued experimentally but without success until now. Here, XRC reveals an interesting Au12 hollow icosahedron in thiolated gold nanoclusters for the first time. The Au–Au bond length, close to that of bulk gold, shows better thermal extensibility than the other Au–Au bond lengths in Au144(SR)60, providing an atomic-level perspective because metal generally shows better thermal extensibility than nonmetal materials. Thus, our work not only reveals the mysterious, long experimentally pursued structure of a transition-sized nanoparticle but also has important implications for the growth of high-quality, single-crystal nanoparticles, as well as for the understanding of the thermal extensibility of metals from the perspective of chemical bonding.


For metal nanoparticles, two regimes are of particular interest, that is, nanocrystals (typically 3 to 100 nm) and nanoclusters (ranging from subnanometer to ~2 nm) (14). The transition from nanoclusters to nanocrystals is a central issue in nanoscience (5). Therefore, the atomic structure determination of metal nanoparticles in the transition size range is of high significance for understanding the evolution from nanoclusters to nanocrystals, as well as the quantum size effect at the atomic level. However, the atomic structure determination is generally very challenging (6). Beginning in 2007, when the atomic structure of Au102(SR)44 (where SR = thiol) was resolved (7), more than a dozen metal nanoclusters have been structurally solved (815); however, the structures of metal nanoclusters in the transition region have rarely been reported (1618). In particular, the real structure of Au144(SR)60 has long remained mysterious (19). Although nanoclusters of larger or smaller sizes than Au144(SR)60 have been structurally solved by x-ray crystallography (XRC), the structure of Au144(SR)60 remains undetermined by XRC since precise determination of its formula by electrospray ionization mass spectrometry (ESI-MS) nearly a decade ago (2022). Efforts to elucidate the structure of Au144(SR)60 have continued over the past decade. In 2009, Lopez-Acevedo et al. (23) proposed a structural model based on density functional theory (DFT) simulations, which was backed up by the subsequent x-ray absorption spectroscopic analysis by MacDonald et al. (24) and Yamazoe et al. (25), nuclear magnetic resonance by Wong et al. (26), and scanning transmission electron microscopy studies by Bahena et al. (27). The proposed symmetric structure indicated that the nanoparticle should not be difficult to crystallize. Several groups acquired crystals of Au144 nanoparticles with various ligands, including phenylethanethiolate (PET; -SC2H4Ph) and para-mercaptobenzoic acid (p-MBA) (28, 29). Koivisto et al. (28) even observed strong diffraction patterns by small-angle x-ray diffraction (XRD), although they failed to do so by large-angle XRD, indicating that they obtained crystals with orientational disorders of the nanoparticles. Evidence of the disorder in the PET ligand layer of the nanoparticles was also observed in infrared (IR) spectroscopy analysis (28).

Despite tremendous efforts devoted to the crystal growth of Au144(SR)60 nanoparticles, it has long been questionable whether fine crystals of sufficient quality for XRC analysis could be obtained. On the basis of the previous work, we rationalize that the intra- and interparticle weak interactions should play critical roles in the crystal growth (18). On one hand, the intraparticle interactions lead to the assembled ligands on the surface of the nanoparticle and restrict the motion of the ligands; on the other hand, the interparticle interactions facilitate the ordered arrangement (assembly) of the nanoparticles and reduce the movement of the nanoparticles in the crystal. To make adequate intra- and interparticle weak interactions, the nature of the ligand, especially the rigidity, should be carefully considered. For example, too much flexibility in the ligands may lead to incompact intra- and interparticle interactions, while too much rigidity in the ligands may inhibit the formation of some intra- and interparticle interactions. This may at least partly explain why high-quality single crystals are difficult to obtain when Au144 is protected by either flexible PET or rigid p-MBA (28, 29). Therefore, we conclude that a ligand with rigidity falling between the flexible PET and the rigid p-MBA may be a resolution of the major problem that high-quality single crystals of Au144(SR)60 have been difficult to produce for a decade. In this work, using PhCH2SH as the protecting ligand, we have now successfully grown single crystals of Au144(SR)60 and determined its structure by XRC, which confirms the “compact” and “adequate” intra- and inter-nanoparticle weak interactions as described above. Notably, other factors such as the purification of the product by preparative thin-layer chromatography (PTLC) (30) and the selection of suitable conditions for crystal growth also contributed to the successful production of single crystals of Au144(SR)60.


The Au144(SR)60 nanoparticles were prepared via a reported one-step method (31). Briefly, chloroauric acid was first mixed with tetraoctylammonium bromide (TOABr) and excess thiol (PhCH2SH) in methanol to form Au(I)-SR polymers. Then, Au(I)-SR was reduced by sodium borohydride (aqueous solution) to form nanoparticles. After a size-focusing process lasting for ~4.5 hours, the precipitated product was washed with methanol and isolated using PTLC. The Au144(SCH2Ph)60 product was confirmed by ESI-MS (see fig. S1). Triangular prism-like single crystals were grown by vapor diffusion of acetonitrile into a toluene solution of the purified nanoparticles over 1 week at room temperature (fig. S2). As a close comparison, block-like crystals of PET-protected Au144(PET)60 were also grown in the same manner but were unfortunately unresolvable (fig. S3), similar to earlier reports by others (28), despite the fact that Au144(PET)60 shows a similar atomic structure as that of Au144(SCH2Ph)60 according to powder XRD (fig. S4). More experimental details are given in Materials and methods.

The ultraviolet-visible (UV-vis) absorption spectrum of the crystallized Au144(SCH2Ph)60 [redissolved in dichloromethane (DCM)] shows two prominent peaks at 355 and 715 nm, with two inconspicuous humps at 475 and 595 nm (fig. S5), indicating the nonplasmonic nature of the Au144(SCH2Ph)60 nanoparticles (5). It is worth noting that the absorption peaks of the redissolved crystals of Au144(SCH2Ph)60 are more distinct than the solution of the amorphous solids of Au144(SCH2Ph)60, probably because of the small-molecule impurities not identified by mass spectroscopy (see fig. S1). The observed red shift of the absorption peaks of Au144(SCH2Ph)60 compared with the peaks of Au144(PET)60 is attributed to the ligand effect (32). The structure of Au144(SCH2Ph)60 is solved by XRC analysis at 153 K with 0.84-Å resolution, which reveals that the crystal of Au144(SCH2Ph)60 belongs to the trigonal crystal system and adopts a space group of P-31c (R1 value of 5.98%; see tables S1 and S2). Such a highly symmetrical packing is not seen in other nanoparticle crystals, except for that of Au60S6(SCH2Ph)36, in which the nanoparticles form a 6H left-handed helical arrangement (33). Note that according to the reported results, large nanoclusters always crystallize in a triclinic or monoclinic space group (1618). Figure 1A shows the total structure of Au144(SCH2Ph)60 with the metal core diameter of ~1.7 nm and the entire nanoparticle diameter (including the surface ligand shell) of ~2.9 nm, consistent with the previous report (27). Figure 1B presents the two enantiomers of the Au144(SCH2Ph)60 nanoparticle, and the parallel hexahedron unit cell contains a racemic pair.

Fig. 1 Total atomic structure of the Au144(SCH2Ph)60 nanoparticle resolved by XRC.

(A) Dextroisomer of Au144(SCH2Ph)60. Yellow, gold; green, sulfur; gray, carbon; white, hydrogen. (B) Parallel hexahedron unit cell containing a pair of enantiomers. The left- and right-handed enantiomers are colored in magenta and yellow, respectively.

The as-obtained structure has high symmetry with 6 C5 rotation axes, 10 C3 rotation axes, and 15 C2 rotation axes, as shown in fig. S6, similar to the Pd145 structure reported by Tran et al. (34), except for the following distinct differences: (i) the central vacancy, which reduces the metal atom count from 145 to 144, and (ii) the profound chirality in the outermost shell(s). The crystal structure matches the model proposed by Lopez-Acevedo, especially the most recent theoretical optimizations (35, 36), which demonstrate that the proposed structures have the I-symmetry (the 60 Au and 60 S-C shells), resembling the traditional “football” of Asia, although the trigonal (D3) symmetry imposed by the unit cell was not shown (37). This level of agreement between the calculations and the experiment is remarkable. The magnitude of the chirality is also very close to the calculation from the model structure by Pelayo et al. (38). It is worth noting that the theoretical modeling of this structure type, as used to account for many observed properties of Au144(SR)60 and related systems, was motivated directly by the Pd145 structure with 60 bridging carbonyl ligands (34) because of the agreement in the powder XRD patterns obtained on the samples, along with the knowledge of the precise mass and the composition (2022). Of note, the work of Lopez-Acevedo et al. (23) did not predict the structure, but rather assumed its correctness, and attempted to refine it using the DFT methods, which were finally efficient enough to optimize such a large system. Below, we discuss the details of the structure of the Au144(SR)60 nanoparticle.

As shown in Fig. 2A, the inner shell is a hollow 12-atom icosahedron. This is the first experimental observation of a hollow icosahedron (Au12) in thiolated gold nanoparticles, although a similar Ag12 icosahedron was observed before (12, 13). In contrast to the hollow structure, the center-filled M13 icosahedron is very common in gold, silver, palladium, and platinum nanoclusters (8, 9, 14, 39). A plausible speculation of the central vacancy is that the hollow structure significantly relieves the intrinsic strain associated with the Mackay icosahedral packing (23). With regard to the analogous Pd145, the palladium atom with its smaller size may accommodate the intrinsic strain. However, unexpectedly, even for the similar-sized Au133(SR)52, a center-filled Au13 icosahedron was observed (16, 17). It is known that Au144(SR)60 can be transformed to Au133(SR)52, which prompts the following question: How can a single gold atom enter the center of the hollow Au12 icosahedron (16, 17)? This needs to be clarified in future work. It is worth noting that a truncated decahedral core structure of Au144(SR)60 has been recently determined by atomic pair distribution function analysis of x-ray powder diffraction data (40), and an approximate face-centered cubic (fcc) packing of gold atoms with local variation has also been observed by aberration-corrected electron microscopy in 3-mercaptobenzoic acid–protected Au144(m-MBA)40 nanoparticles (41, 42). In addition, Vergara et al. (43) have reported the crystal structure of a Au146(p-MBA)57 nanocluster, which unexpectedly has a twinned fcc metal core. The structural differences between these comparable sizes of gold nanoclusters are not odd and are likely attributed to the differences in ligands and synthetic protocols because these factors are known to influence the kinetics and thermodynamics of the formation of the final product(s), as adequately demonstrated by our previous work (10, 44, 45), among which we even obtained a pair of structural isomers by slightly tuning the synthesis conditions (10).

Fig. 2 Dissection of the Au144(SCH2Ph)60 atomic structure.

(A) Inner icosahedral shell of 12 Au atoms (light blue). (B) Middle icosahedral shell of 42 Au atoms (pink). (C) Outer rhombicosidodecahedron shell of 60 Au atoms (orange). (D) Thirty (SR-Au-SR) staples containing 30 Au atoms (yellow) and 60 sulfur atoms (green). (E) Three-layered Au114 core with an a-b-c structure. (F) Isolated (SR-Au-SR) staple astride the parallelogram on the Au60 shell. Alkyl groups are omitted for clarity.

Another interesting finding is that the average Au–Au bond length in the hollow M12 core is only 2.748 Å (see fig. S7), which is much smaller than the normal Au–Au bond length found in gold nanoparticles (including nanoclusters and nanocrystals) (1618). The second layer is composed of 42 gold atoms, forming a 54-atom Mackay icosahedral structure with the hollow Au12 core (54-MI) and exhibiting 20 triangular (111) faces (Fig. 2B), rather than the 55-MI in the reported Pd145 and Au133 structures. Then, 60 gold atoms enclose the 54-MI and assemble into a rhombicosidodecahedron (Fig. 2C), which was previously observed in the Pd145 nanoparticle (34). The surface of this polyhedron consists of 12 pentagons and 20 equilateral triangles. The centers of the 12 pentagons correspond to the vertices of 54-MI, and a triangular (111) face covers each face of 54-MI in a hexagonal close-packed (hcp) packing mode. The shared vertexes of the pentagons and the triangles on the rhombicosidodecahedron form 30 parallelograms [not squares as predicted (23)], on which 30 (RS-Au-SR) staples ride astride, giving rise to the complete Au144(SR)60 structure (Fig. 2, D and F). The anatomy of the Au114 kernel is shown in Fig. 2E, indicating an a-b-c layer stacking. In previous reports, the Au–Au bonds in and between each shell were believed to be equivalent (24). However, detailed analyses (fig. S7) demonstrate that multiple kinds of Au–Au bond lengths exist in Au144(SCH2Ph)60, implying a transition state of Au144(SCH2Ph)60 from nanoclusters to nanocrystals in view of the chemical bonding. For example, the average Au–Au bond lengths in the pentagons and triangles of the c layer are 3.217 and 2.884 Å, respectively (the shorter bonds of the triangle’s edges presumably arise from the steric strain, which was also found in Pd145); the Au–Au bond length in the b layer is 2.914 Å. The bond length of 2.884 Å (one type of bond in layer c) is close to that of bulk gold, indicating that the Au144 nanoparticle has some metal character, while the other bond lengths indicate that the nanoparticle has some nonmetal character. Note that, although the shell-by-shell (a-b-c, 12-42-60) construction and chiral assembly of 30 monomeric (S-Au-S) staple motifs on the surface of Au114 have been correctly proposed theoretically (23), the Au–Au bond lengths in the first and second shells predicted by the calculation have deviations from the experimental results [first shell, 2.845 (theory)/2.748 (experiment) Å; second shell, 3.010 (theory)/2.914 (experiment) Å], and unfortunately, the detailed Au–Au bond length in the third shell was not provided in the theoretical work. In addition, the staple motifs in the theoretically proposed structure were oriented in an arbitrary manner such that the total structural symmetry was undefined. A detailed comparison between the present solved structure and the structure proposed by Lopez-Acevedo et al. is shown in fig. S8. The range of Au–Au distances in the kernel of Au144 is similar to the one reported by Azubel et al. (46) for Au68(3-MBA)32. In that work, an unexpected optical gap of 2500 cm−1 was shown in the IR spectrum of Au68(3-MBA)32. The Au144(SCH2Ph)60 seems to show a gap at ~750 cm−1, as seen in fig. S9.

To stabilize the large Au114 core, 30 monomeric (S-Au-S) staple motifs are distributed on the surface of the Au114 core with highly ordered patterns. From the side view shown in Fig. 3A, the staples self-assemble into an annulus pattern, which can be divided into three parts (marked with different colors): The 5-membered rings with 5 staple motifs near the poles of the Au114 kernel are marked in red; the 10-membered ring consisting of 10 staple motifs on the equator is marked in yellow, and the remaining two 5-membered rings located in the midlatitude regions of the Au114 core are marked in blue. From the top view (Fig. 3B), the five-membered ring looks similar to a pentagonal ripple, and the red and blue rings are arranged in opposite rotary directions. All five rings surround a fivefold axis across the two opposite pentagons. Note that the Au60 rhombicosidodecahedron has 12 pentagons and six fivefold axes (27, 35). Thus, each fivefold axis can be encompassed by five rings (Fig. 3C). Moreover, the 10-membered ring on the equator can be separated into two alternate sets: One set consists of the five parallel staples along the fivefold axis, and the other deviates from the equator. Meanwhile, the five parallel staples can be regarded as the vertices of a penta-star (Fig. 3D). In another view, the 30 staples can also be grouped into six penta-stars. The reverse orientations of the five rings give rise to the chirality of the Au144(SCH2Ph)60 nanoparticle (Fig. 3E), and a pair of enantiomers are presented in Fig. 3F. In contrast, the monomeric (S-Au-S) staple motifs in Au144 are replaced by triethylphosphine ligands directly connected to the outermost 30 palladium atoms, and the assumed 60 CO groups (disordered in the crystal) bridge the palladium atoms on the edges of Pd145 (34). Thus, Pd145 does not exhibit the chirality found in Au144. To sum up, the Au144 structure shares some similarities with Pd145, but there are distinct differences as well.

Fig. 3 Self-assembly mode of surface (SR-Au-SR) staples.

(A) Side view, (B) top view, and (C) rotated view of the ring-like assembled staples on the surface; different groups of gold atoms are colored in red, blue, and yellow, respectively. (D) Exhibition of the five paralleled staples at the vertex peaks of a penta-star. (E) Side view and (F) top view of the chiral ring-like assembled staples. The white arrows indicate the orientation of each ring. Alkyl groups are omitted for clarity.

As mentioned in Introduction, the intra- and internanoparticle interactions were proposed to play critical roles in the crystal growth (18). Here, we found the weak interactions in the crystal. As shown in Fig. 4A, the H⋅⋅⋅H interaction exists between the hydrogen atoms of two adjacent benzenes in the same staple, evidenced by their very close distance (2.545 and 2.837 Å, respectively; see fig. S10A) and the similar inward orientation of the two adjacent benzenes. In addition to the H⋅⋅⋅H interaction, the C-H⋅⋅⋅π intraparticle interaction also exists, and the H⋅⋅⋅π distance is short (average, 2.80 Å; see Fig. 4B and fig. S10B) (18). These interactions lead to the reverse spiral arrangements of the two sets of benzene rings in the staple ring (inside and outside, five benzenes for each set), shown in fig. S10C, which was not previously reported. As a consequence, these interactions enable every single nanoparticle to be compacted and ordered with reduced intraparticle vibration, contributing to the formation of high-quality single crystals.

Fig. 4 Self-assembly mode analysis of surface ligands.

(A) Different direction views of monomeric (SR-Au-SR) staple. The white arrows point in the orientation of the benzene ring. Yellow, gold; green, sulfur; gray, carbon; white, hydrogen. (B) Weak interactions between the interacted 10 ligands. Red dashed lines, C-H⋅⋅⋅π interaction; blue dashed lines, H⋅⋅⋅H interaction.

In addition to the intraparticle interactions, interparticle interactions are also found at the interface of neighboring Au144(SCH2Ph)60 nanoparticles. As shown in Fig. 5 (A and B), the enantiomers are packed into different layers and organized into a square superlattice in the (100) and (010) planes. A hexagonal pattern can be observed in the (001) plane (Fig. 5C), with each Au144(SR)60 nanoparticle being wrapped by six equivalent anti-chiral nanoparticles (Fig. 5D). When the interface of two adjacent nanoparticles is magnified, six thiolates are found enwound by both C-H⋅⋅⋅π and H⋅⋅⋅H interactions, which act as a “clip” to fix the interface ligands and the neighboring nanoparticles, finally leading to the formation of an ordered superlattice (Fig. 5, E and F, and fig. S11) (18).

Fig. 5 Self-assembly mode analysis of packing nanoparticles in a crystal.

(A) View of the superlattice from the x axis, (B) y axis, and (C) z axis. (D) Packing of a dextroisomer with the six nearest levoisomers. Magenta and yellow, gold atoms in left- and right-handed enantiomers; green, sulfur; gray, carbon; white, hydrogen. (E and F) Weak interactions between ligands in neighboring nanoparticles. Magenta and yellow, carbon atoms in left- and right-handed enantiomers; green, sulfur; white, hydrogen. Red dashed lines, C-H⋅⋅⋅π interaction; blue dashed lines, H⋅⋅⋅H interaction.

We reproduced the structure by analyzing another single crystal obtained in another synthesized batch, demonstrating the reliability of our experimental results. The reliable structure provides a prerequisite to investigating the thermodynamic influence on the structure of the Au144(SCH2Ph)60 nanoparticles. As shown in fig. S12 and tables S3 to S6, the structural framework is essentially unchanged as the temperature increases from 100 K (0.84-Å resolution) to 243 K (0.83-Å resolution). However, the Au–Au bond length changes are distinctly observed when the temperature rises from 100 to 243 K. All Au–Au bond lengths in the Au114 core elongate more or less as the temperature increases. More specifically, the Au–Au bonds in the inner shell elongate from 2.748 to 2.758 Å, and the Au–Au bonds in the second shell elongate from 2.917 to 2.922 Å. The two kinds of Au–Au bonds in the third shell stretch from 3.220 to 3.224 Å and 2.883 to 2.895 Å, respectively. The small-angle XRD analysis also qualitatively reveals that the distance between the two neighboring nanoparticles in the amorphous powder enlarges as the temperature increases, indicating that the nanoparticles expand with the temperature rise (fig. S13). An interesting finding is that the Au–Au bonds with a length close to that of bulk gold (2.88 Å) have the largest elongation (+0.41%), indicating that this kind of bond has better thermal extensibility than the shorter or longer ones, which may provide the atomic-level perspective for the fact that metals generally show better thermal extensibility than nonmetal solid materials.

It came to our attention during the revision of this manuscript that Lei et al. (47) recently reported a Au144 structure protected by alkynyl ligands. The two Au144 nanoparticles have a similar kernel but different staple motifs due to the ligand difference. The averaged Au–Au bond lengths in the shells of the kernel are also slightly different (first shell, 2.755/2.748 Å; second shell, 2.912/2.914 Å; third shell, 3.162/3.220 and 2.946/2.883 Å; alkynyl-protected/thiolate-protected Au144). Besides, the total structure of Au144(C≡CAr)60 (Ar = 2-F-C6H4-) does not exhibit the I-symmetry of Au144(SCH2Ph)60 because of the asymmetry of the C≡CAr ligand. The Au144(SCH2Ph)60 differs from Au144(C≡CAr)60 not only in ligand composition and surface structure but also in the material properties (for example, optical absorption). Furthermore, Au144(C≡CAr)60 is a new species and does not involve the historical long pursuit of the thiolated Au144 nanoparticle. Finally, the intra- and internanoparticle interactions, temperature-dependent Au–Au bond length, etc., are not involved in the work of Lei et al.


In summary, we have reproducibly grown ideal single crystals of the Au144(SR)60 nanoparticle and resolved its structure using XRC on the basis of designed intra- and interparticle weak interactions, as these play critical roles in growing high-quality single crystals of metal nanoparticles. XRC also reveals a multiplicity of Au–Au bond lengths in the kernel of Au144(SR)60, indicating that Au144(SR)60 is a transition size in terms of Au–Au bonding character. It is found that the Au–Au bond with the length close to that of bulk gold shows the best thermal extensibility among various Au–Au bonds in the Au144(SR)60 nanoparticle, which provides an atomic-level perspective for the fact that metals generally show better thermal extensibility than nonmetal solid materials. The significance and novelty of our work include the following: (i) We resolved a structure that has perplexed the scientific community for nearly a decade since its composition determination by ESI-MS. (ii) The available structure will pave the way for future research of structure-property correlation, transition-sized nanoparticles, etc. (iii) We provided important implications and guidance for the challenging single-crystal growth of metal nanoparticles. (iv) This work also provided an atomic-level perspective for the good thermal extensibility of metals compared with other common nonmetal solid materials.



Tetrachloroauric(III) acid (HAuCl4·4H2O, >99.9% metal basis), TOABr (≥98.0%), phenylethanethiol (99%, Sigma-Aldrich), benzyl mercaptane (99%, Sigma-Aldrich), 4-tert-butylbenzylmercaptan (99%, Sigma-Aldrich), p-methylphenylthiophenol (99%, Sigma-Aldrich), 2,4-dimethylbenzenethiol (99%, Sigma-Aldrich), and sodium borohydride (NaBH4, 99.8%, Shanghai Chemical Reagent Co. Ltd.) were used. Ethanol, toluene, acetonitrile, n-pentane, CH2Cl2, and methanol were purchased from Sinopharm Chemical Reagent Co. Ltd. The water used in all experiments was ultrapure (resistivity, 18.2 megohms·cm), produced by a Milli-Q NANO pure water system. All chemicals were used as received.

Synthesis of Au144(SR)60

The synthesis of Au144(SR)60 nanoclusters is documented in a previous work and modified here (31). All the reactions were conducted at room temperature under ambient air. Typically, HAuCl4·4H2O (0.607 mmol, 250 mg) and TOABr (0.696 mmol, 380 mg) were added to a 100-ml single-neck round-bottom flask and mixed with 30 ml of methanol. After being vigorously stirred for 30 min, the solution color changed from yellow to dark red. Then, various thiols (3.19 mmol for each synthesis) were added to the solution at room temperature. A milky, white floc was generated immediately. After ~15 min, a fresh NaBH4 solution (6 mmol, 227 mg, dissolved in 10 ml of nanopure water) was rapidly added to the solution under vigorous stirring. The color of the solution immediately turned black and produced Au particles, which were precipitated out of the methanol solution. The reaction was stopped after ~4.5 hours, and then the solution was removed. Black precipitates were washed with excess methanol and collected by centrifugation. This wash/precipitation procedure was repeated several times. The crude products that were dissolved in 1 ml of CH2Cl2 were pipetted onto eight pieces of a PTLC plate (10 cm by 20 cm), and the separation was conducted in a developing tank (solvent: DCM/petroleum ether = 20:20, v/v) for ~5 min. Then, the band of Au144(SCH2Ph)60 in the PTLC plate was cut, and the nanoparticles were extracted with pure DCM and then dried by rotary evaporation. The yield of Au144(SCH2Ph)60 was 16% (on an Au atom basis). After trying various solutions and precipitants, it was determined that single crystals of Au144(SCH2Ph)60 could be acquired by vapor diffusion of acetonitrile into a toluene solution of Au144(SCH2Ph)60 over 1 week at room temperature. The synthesis and crystal growth of other thiolated Au144 compounds were conducted in the same manner.


ESI mass spectra were recorded on a Waters QTof mass spectrometer using a Z-spray source. The sample was first dissolved in toluene (~0.5 mg/ml) and then diluted (2:1, v/v) with an ethanol solution that contained 50 mmol CsOAc. The sample was directly infused into the chamber at 5 μl/min. The source temperature was kept at 70°C, the spray voltage was 2.20 kV, and the cone voltage was adjusted to 60 V. XRD data were collected with an X’Pert PRO MPD (PANalytical) diffractometer using Cu Kα as the radiation source at 40 kV and 40 mA. All UV/vis/near-IR absorption spectra were acquired in the range of 190 to 900 nm using a UV2600 spectrophotometer. The single-crystal XRD data were collected on a Bruker D8 VENTURE PHOTON 100 CMOS diffractometer with HELIOS MX multilayer monochromator Cu Kα radiation (λ = 1.54178 Å). The structure and phase purity were checked by a Rigaku-TTR3 x-ray diffractometer using high-intensity graphite monochromatized Cu Kα radiation. Fourier transform IR spectra were acquired on a Nicolet 8700 spectrometer. The small-angle XRD patterns at selected temperatures were registered to investigate the structural phase transition, where the variation in temperature was realized by using a closed He-gas cycle refrigerator installed on the x-ray diffractometer.


Supplementary material for this article is available at

Fig. S1. ESI-MS analysis of Au144(SCH2Ph)60.

Fig. S2. Photograph of triangular prism-shaped Au144(SCH2Ph)60 crystals.

Fig. S3. Photograph (partly marked with red circle) and single-crystal diffraction pattern of Au144(PET)60 amorphous solid.

Fig. S4. XRD patterns of Au144(SCH2Ph)60 and Au144(SC2H4Ph)60 powders.

Fig. S5. UV-vis absorption spectra of Au144(SCH2Ph)60 nanoparticles redissolved in solutions of amorphous and crystalline Au144(SCH2Ph)60 solids and crystalline Au144(SC2H4Ph)60 solid.

Fig. S6. The atomic structure of Au144(SCH2Ph)60 viewed along different symmetry axes.

Fig. S7. The statistics of different types of Au–Au bonds and Au–S bonds of Au144(SCH2Ph)60.

Fig. S8. A detailed comparison between the present solved structure of Au144(SCH2Ph)60 and the structure proposed by Lopez-Acevedo et al.

Fig. S9. IR spectrum of Au144(SCH2Ph)60.

Fig. S10. The weak interactions between surface ligands and the chirally assembled pattern.

Fig. S11. The weak interactions between surface ligands in neighboring nanoparticles.

Fig. S12. The statistics and increasing proportion of the Au–Au bond in different shells of the Au114 core resolved at 100 and 243 K, respectively.

Fig. S13. Small-angle XRD patterns of amorphous Au144(SCH2Ph)60 powder at selected temperatures.

Table S1. Crystal data and structure refinement for Au144(SCH2Ph)60 at 153 K (sample 1).

Table S2. Atomic coordinates and equivalent isotropic displacement parameters for Au144(SCH2Ph)60 at 153 K (sample 1).

Table S3. Crystal data and structure refinement for Au144(SCH2Ph)60 at 100 K (sample 2).

Table S4. Atomic coordinates and equivalent isotropic displacement parameters for Au144(SCH2Ph)60 at 100 K (sample 2).

Table S5. Crystal data and structure refinement for Au144(SCH2Ph)60 at 243 K (sample 2).

Table S6. Atomic coordinates and equivalent isotropic displacement parameters for Au144(SCH2Ph)60 at 243 K (sample 2).

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: A portion of this work was performed on the Steady High Magnetic Field Facilities, High Magnetic Field Laboratory, CAS. We thank L. Zhang for the discussion of XRD pattern. Funding: This work was supported by the National Natural Science Foundation of China (nos. 21501181, 21829501, 21222301, 21771186, 21603234, 21701179, 21528303, and 21171170), Key Program of 13th Five-Year Plan, CASHIPS (KP-2017-16), Director Foundation of ISSP (no. 2016DFY05), Innovative Program of Development Foundation of Hefei Center for Physical Science and Technology (2017FXCX002), and CAS/SAFEA International Partnership Program for Creative Research Teams. Author contributions: N.Y. synthesized the nanoparticles and grew the crystals. N.X. and L.L. carried out the structural analysis. M.Z. and F.J. helped synthesize the nanoparticles. Z.W. and R.J. designed the study, supervised the project, and analyzed the data. All authors contributed to the preparation of the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

Stay Connected to Science Advances

Navigate This Article