L uminophores emitting in the NIR region (800-1700 nm) window are increasingly valued across many fields, [1][2][3][4][5] such as bioimaging and NIR optics. [6][7][8] Thiolate-protected Au n (SR) m NCs (SR = thiolate) have recently emerged as a promising class of NIR-emissive materials. [9][10][11][12][13][14][15] These NCs feature a core-shell structure, [16][17][18] in which the inner Au(0) core is enclosed by Au(I)-SR "staple motifs". The tailorable size, structure, and composition of Au NCs allow them to exhibit emission peaks across the visible to NIR range. [19][20][21][22] Moreover, their atomic precision aids in a deeper understanding of photophysical mechanisms, 23,24 facilitating the design of highly luminescent materials. Currently, a few highly luminescent NCs in the visible range have been reported, [25][26][27][28][29][30] but the NIR region is still difficult due to the energy gap law induced significant loss of excitation energy via nonradiative relaxation. 31,32 The photoluminescence quantum yield (PLQY) of NIR-emissive Au NCs is often below 1%, 33,34 except a few cases 15,21,[35][36][37][38][39][40] under ambient conditions.

Enhancing the PLQY can be accomplished by increasing the radiative decay rate (k r ) and/or decreasing the nonradiative decay rate (k nr ) according to the formula, PLQY

k k k r r n r

= + . In the case of Au NCs, given their significantly higher k nr (10 5 -10 7 s -1 ) than the k r (10 4 -10 5 s -1 ), reducing the k nr offers a greater opportunity for PLQY enhancement. 19,41,42 The PL properties of Au NCs have been recognized to be intricately linked to the Au(I)-SR "staple motifs", thus, restricting motions associated with these surface motifs is generally an effective strategy for achieving higher PLQY by suppressing the k nr . 5,[43][44][45] Here, we report a noncoordinating interaction strategy for the suppression of k nr to enhance the NIR emission of rodshaped Au 42 (PET) 32 (PET = 2-phenylethanethiolate). Specifically, the nonradiative energy loss in Au 42 is suppressed by the addition of amide-containing small molecules, thus improving the PLQY to 50% in solution at room temperature. Cryogenic PL analysis reveals that the vibrations associated with the Au-S staples on Au 42 are suppressed by amide molecules. Moreover, when Au 42 is embedded in a polymer film containing amide molecules, the PLQY is further boosted to 75% at room temperature.

The Au 42 quantum rod was synthesized using a method of N-heterocyclic carbene (NHC)-mediated kinetic control reported by our group. 46 The Au 42 structure shows a rodshaped, hexagonal close-packed Au 20 kernel protected by two pairs of interlocked Au 4 (PET) 5 motifs (marked in green and light green) on the two ends and six monomeric Au(PET) 2 motifs (marked in blue) on the body (Figure 1A). 34,47 The optical absorption spectrum of Au 42 exhibits two major peaks at 375 and 806 nm (Figure 1B, green profile). Theoretical simulations identified that the 806 nm peak originates from the HOMO-to-LUMO transition and the transition dipole is strongly polarized along the longitudinal direction, while the 375 nm peak is not. 47 Upon excitation at 806 nm, Au 42 exhibits fluorescence and phosphorescence dual emission at 875 nm (denoted FL) and 1040 nm (PH) (Figure 1B, blue profile), respectively, with a total PLQY of 18% (Figure S1); note that this value is higher than the 12% reported earlier 46 due to the different excitation wavelengths (806 nm in this work versus 380 nm previously).

When Au 42 (0.1 OD at 806 nm, absorption coefficient ε 806 = 1.08 × 10 5 M -1 cm -1 , 48 i.e., 9.26 × 10 -4 mM) was mixed with nonluminescent N,N-dimethylbenzamide (DMBA, Figure S2), the Au 42 absorption profile remains unchanged, but its NIR absorption peak blueshifts from 806 to 781 nm with increasing amide concentration from 0 to 2143.9 mM (Figure 1C and Figure S3), and the integrated PL intensity of Au 42 increases significantly by ∼3-fold (Figure 1D and Table S1), reaching a total PLQY of 50.1% (Figure 1E and Table S1). Specifically, the PLQY initially remains unchanged with the concentration up to 53.6 mM (Stage I). It then exhibits a gradual rise, reaching 50.1% at the DMBA concentration of 1286.4 mM (Stage II), and maintains this intensity as the concentration is further increased (Stage III). When Au 42 was precipitated out of the solution to remove amides and redissolved in C 2 Cl 4 , the PLQY of Au 42 recovers to the initial 18%, indicating noncoordinative interactions between Au 42 and DMBA.

The dual PL bands are deconvoluted to analyze the respective variation of FL and PH (Figures S4 and S5 and Table S1). It is evident that the FL shows a dependence on the concentration of DMBA, but the PH remains constant. Generally, the FL enhancement can be accomplished either by increasing the k r and/or reducing the k nr . Here, our results reveal a significant reduction in the k nr for the FL of Au 42 upon the addition of DMBA, plummeting from 13.51 × 10 8 s -1 to 3.33 × 10 8 s -1 , together with a moderate increase in k r from 1.42 × 10 8 s -1 to 2.51 × 10 8 s -1 (Figure 1F and Table S1).

We further conducted cryogenic PL measurements from room temperature to 80 K (Figure 2A and B). For the Au 42 / DMBA system, we selected a DMBA concentration of 857.6 mM to ensure a significant PL enhancement but preventing the precipitation of DMBA at low temperatures. Given the fact that Au NCs exhibit stronger absorption at low temperatures, we also performed temperature-dependent absorption (Figure S6) to correct PLQY at low temperatures. The cryogenic PL for Au 42 and Au 42 /DMBA in 2-methyltetrahydrofuran (2-MeTHF) are shown in Figure 2A-B. The PLQY of Au 42 (without DMBA) increases from 16.8% to 45.6% as the temperature is lowered from 298 to 80 K; note: 16.8% in 2-MeTHF ("glass" forming solvent) slightly differs from 18% in C 2 Cl 4 . For the Au 42 /DMBA, the PLQY rises from 45.7% to 89.1% in the same temperature range. The detailed results of peak deconvolution are provided in Tables S2 and S3. Both FL and PH intensities for the two systems increase as the temperature decreases, in contrast to the sole FL enhancement by amide. The FL for the Au 42 /DMBA system is consistently higher than that of Au 42 (Figure 2C). Conversely, the PH emission remains nearly identical for the two systems at each temperature, though the PH increases at lower temperatures (Figure S7). The PL excitation spectra for Au 42 and Au 42 / DMBA were also compared (Figures S8 and S9). The PL excitation at 80 K shows a blue shift compared to that at 298 K, consistent with the cryogenic absorption (Figure S6A).

We further compared the k r and k nr of the FL for both Au 42 and Au 42 /DMBA systems at low temperatures (Figure 2D). The k r values for both systems remain relatively constant, but the k nr values for both Au 42 and Au 42 /DMBA exhibit a notable decrease, attributed to the suppression of staple vibrations at low temperatures; note: the core vibrations are typically manifested at even lower temperatures than 80 K. 49 Additionally, it is important to highlight that the k nr of Au 42 /DMBA is significantly lower than that of Au 42 at the same temperatures. To elucidate the mechanism underlying the decrease in k nr of FL upon the addition of DMBA, we fitted the temperaturedependent FL intensity evolution by eq 1 50

where I 0 represents the initial intensity, a denotes the ratio of nonradiative and radiative probabilities, and E is the activation energy for the nonradiative relaxation. Here, only one dominant phonon-assisted nonradiative channel is considered in this modeling. The corresponding fitting line and parameters are shown in Figure 2E, where the activation energies of phonon modes that coupled with the FL of Au 42 and Au 42 /

DMBA are determined to be 38.9 and 22.3 meV, respectively; note: 1 meV = 8 cm -1 . This suggests that the addition of DMBA suppresses the vibrations associated with the Au-S staples on the Au 42 . Meanwhile, the a value falls drastically from 4.5 to 1.6, also indicating a significant suppression of the staple vibration-induced nonradiative decay. Moreover, we extracted and compared the temperature-dependent full-width at half-maximum (fwhm) values for Au 42 and Au 42 /DMBA (Figure 2F). Generally, both acoustic phonon modes (low energy) and optical phonon modes (high energy) contribute to the broadening of PL line width, but our experiments are conducted down to 80 K only  where the contributions from acoustic phonons are trivial and can be omitted, thus we only consider the optical phonon factor to model the line width broadening by eq 2 24 T e ( ) 1

1

where Γ 0 is the temperature-independent intrinsic line width, γ LO refers to the coupling coefficient of electrons with longitudinal optical (LO) phonons, and E LO denotes the average energy for coupled LO phonon modes. The modeling results (Figure 2F) reveal that the average LO phonon energies for Au 42 and Au 42 /DMBA are 30 and 15 meV, respectively. The reduced phonon energy in Au 42 /DMBA aligns with the eq 1 fitting analysis, indicating a suppression of surface vibrations. Meanwhile, the coupling strength for Au 42 /DMBA (γ LO = 43 meV) is much lower than that for Au 42 (γ LO = 132 meV), suggesting a diminished electron-phonon interaction in the Au 42 /DMBA system. The high PLQY (50%) of Au 42 /DMBA in the NIR region is rare among the reported Au NCs (Figure S10). In addition to DMBA, we found that other amide molecules (Figure 3A), such as N,N-dimethylformamide (DMF), N,N-dimethylacetamide (DMAc), and N-methylformanilide (NMFA), have similar effects on Au 42 , including (i) the longitudinal absorption peak of Au 42 at 806 nm undergoes a blueshift when mixed with these molecules (Figure S11), and (ii) a significant enhancement of the PLQY of Au 42 is observed (Figure S12 and Table S4), e.g., 29.3% for DMF, 47.0% for DMAc, and 55.8% for NMFA. The peak deconvolution analysis (Figure S13) further indicates that these amides predominantly boost the FL (Figure 3B) but not the PH. Additionally, the observed increase in FL intensity is primarily attributed to the suppression of nonradiative relaxation (Figure 3B).

To pinpoint the specific atoms in the amide group accountable for the PL enhancement, we tested two small molecules composed of only nitrogen or oxygen atom, e.g., N,N-dimethylaniline (DMA) and acetylacetone (AA), but neither molecule nor their mixture induced any blueshift in the longitudinal absorption peak of Au 42 (Figure S14), nor did they enhance the PL intensity of Au 42 (Figure S15 and Table S4). This comparison underscores a cooperative effect of nitrogen and oxygen atoms of amides on the PL enhancement of Au 42 while retaining its structure (Figure S16).

The amide molecules can further enhance the emission of Au 42 embedded in a polymer film. As illustrated in Figure S17, the PLQY of sole Au 42 increases from 18% to 52% when embedded in polystyrene (PS) films, and it is further elevated to 75% with the addition of DMBA into the Au 42 /PS film at room temperature. This highly emissive film holds promise in applications such as NIR optoelectronic devices and security as well as quantum telecom.

In summary, we report an effective strategy involving noncoordinative interactions between amides and Au 42 to achieve high PLQY (50% in solutions and 75% in films) in the NIR range by significantly reducing the nonradiative decay rate. This method is also effective for other Au n quantum rods. 48 Our findings offer inspirations for strategically designing highly efficient NIR emitters, opening new avenues for the use of engineered nanoclusters in diverse applications.