NIR induced modulation of the red emission from erbium ions for selective lanthanide imaging

Upon direct excitation with green light (522 nm), Er3+ ion doped nanoparticles feature a number of radiative and non-radiative decay pathways, leading to distinct and sharp emission lines in the visible and near-infrared (NIR) range. Here we apply, in addition to continuous 522 nm irradiation, a modulated NIR irradiation (1143 nm) to actively control and modulate the red emission intensity (around 650 nm). The modulation of red Er3+ ion emission at a chosen frequency allows us to reconstruct fluorescence images from the Fourier transform amplitude at this particular frequency. Since only the emission from the Er3+ ion is modulated, it allows to selectively recover the lanthanide specific signal, removing any non-modulated auto-fluorescence or background emission resulting from the continuous 522 nm excitation. The modulated emission of specific lanthanides can open up new detection opportunities for selective signal recovery.

Here we demonstrate for the first time a combination of direct excitation in the visible range and additional NIR excitation to enhance and modulate the red emission of Er 3+ UCNPs. We demonstrate the potential of this approach for emission modulation-based imaging by recovering the selective lanthanide signal from environments containing bright organic fluorophores. This selective emission modulation at a constant frequency allows for efficient discrimination of the non-modulated background [39][40][41][42].

Results and discussion
Details on the Er 3+ UCNPs synthesis and size characterization can be found in the supporting information and figure S1 which are available online at stacks. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
iop.org/MAF/6/044001/mmedia. Figure 1(a) shows the electronic energy diagram of the Er 3+ ions and the decay pathways that occur upon direct excitation with 522 nm light [43,44]. The 522 nm laser excites Er 3+ from the ground state to the 2 H 11/2 state from which the Er 3+ ions undergo non-radiative transitions to the subsequent energy levels 4 S 3/2 and 4 F 9/2 . Besides nonradiative transitions, two emission lines appear on the Stokes side in the visible range at about 550 nm ( 4 S 3/2 -4 I 15/2 ) and 660 nm ( 4 F 9/2 -4 I 15/2 ) [43]. Additional emission lines, such as the NIR emission around 810 nm ( 4 I 9/2 -4 I 15/2 ), 850 nm ( 4 S 3/2 -4 I 13/2 ), 980 nm ( 4 I 11/2 -4 I 15/2 ) and 1530 nm ( 4 I 13/2 -4 I 15/2 ), should also be present [43,44]. The 4 S 3/2 -4 I 13/2 transition is of particular importance for populating the 4 I 13/2 state which we will use to modulate the red emission. In order to increase and, hence, modulate the red emission of the 4 F 9/2 -4 I 15/2 transition, we have chosen a secondary NIR wavelength of 1143 nm that can optically pump the Er 3+ ions back from the 4 I 13/2 to the 4 F 9/2 excited state (See figure 1(b)). In accordance with the electronic diagram proposed by Anderson  [43,44] might be affected if ET2 and the depopulation of 4 I 13/2 contributes significantly to the 810 nm emission. We measured the 810 nm and 850 nm emission as can be seen in figure S2. The intensity of the 850 nm and 810 nm emission remains unchanged upon NIR modulation which indicates that ET2 does not significantly contribute to the 810 nm emission or is not affected by NIR modulation. Besides emission on the Stokes side, the 522 nm excitation also generates emission on the anti-Stokes side (see figure S3), however this is not of interest for this paper. In order to check the efficiency by which the Er 3+ UCNPs can be pumped back from 4 I 13/2 to 4 F 9/2 , we recorded spectra with and without the additional 1143 nm light (See figure 1(c)) of a sample of Er 3+ UCNPs on a glass coverslip. Figure 1(c) shows that at a secondary NIR excitation intensity of 15 kW cm −2 , the emission of the 660 nm transition increases more than twofold, while the spectral shape remains unchanged. The amount of additional 660 nm emission is dependent on the decay time of the 4 I 13/2 state (4.4 ms in D 2 O) and the intensity of the 1143 nm laser [44]. The latter was proven by changing the 1143 nm laser intensity while measuring the 660 nm emission as shown in figure 1(d). The dependency follows a linear trend up to about a NIR laser intensity of 12 kW cm −2 . Afterwards, the emission intensity seems to deviate from linearity, which might indicate saturation of the optical pumping. Furthermore, the unchanged green emission intensity in figure 1(c) shows that we observe direct optical pumping and not a multiphoton absorption process. Additional factors like temperature, particle size and composition will also have an effect on the observed emission of the UCNPs, but they were not the focus of this investigation [34,35].
After we demonstrated spectrally that our proof of principle works, we continued with measuring the integrated intensity of the red emission shown in figure 1(c) with a long pass/short pass filter combination (creating a transmission window from 633 nm to 750 nm) on an avalanche photo diode (APD) as a function of time. During the measurements, the 1143 nm laser light was modulated on and off by a chopper wheel with adjustable frequency. The resulting intensity time traces can be seen in figure 2(a) for frequencies of 50, 100 and 200 Hz. The ratio between red emission intensities measured with and without 1143 nm irradiation is similar as in the experiment in figure 1(c). From the time traces, we calculated the Fourier transform which is displayed in figure 2(b). As expected, a clear peak in the Fourier transform spectrum can be seen at the frequency that coincides with the frequency settings of the chopper wheel. Although it is possible to modulate the red emission of the Er 3+ ions at even higher frequencies, problems might arise when approaching a frequency that is close to the inverse of the transit time from 2 H 11/2 -4 I 13/2 [36,37,45].
In order to evaluate the Er 3+ modulation for selective lanthanide signal recovery and imaging, we prepared a sample which consisted on one side of Er 3+ UCNPs embedded in polystyrene (PS) and on the other side of an organic fluorophore (ATTO 633) embedded in polyvinyl alcohol (PVA). A scheme of the sample is shown in figure 3(a). A piezo scanner was used to raster scan the sample while acquiring single photon macro times (time elapsed since the start of the experiment) for each detected single photon with time-correlated single photon counting hardware. In this example, the sample was excited with the 522 nm primary and the 100 Hz modulated 1143 nm secondary excitation wavelength. The intensity image in the spectral bandpass range (633 nm-750 nm) can be seen in figure 3(b) and shows that the ATTO 633 dye (on the left side) displays a homogeneous and, on average, brighter emission than the Er 3+ UCNPs (on the right side). Since the emission of the ATTO 633 dye is not modulated by the 1143 nm laser, removal of this 'unwanted' emission can be achieved by constructing an image of the amplitude of the 100 Hz contribution in the FT spectrum. An example of the intensity time trace and FT spectrum at each pixel (time the scanner moves over one pixel is 0.4 s) is given in figure 3(c). For image reconstruction the maximum FT amplitude in the range from 95 to 105 Hz was determined and plotted. The resulting FT amplitude image can be seen in figure 3(d) and shows that the amplitude of the 100 Hz modulation image recovers selectively only the red emission of the Er 3+ ions, with complete removal of the ATTO 633 emission. Scan time per pixel and modulation frequency will determine the total time and quality of the FT amplitude image.

Conclusion
We have demonstrated the use of multiple excitation wavelengths for manipulating the 4 F 9/2 and 4 I 13/2 excited states in Er 3+ UCNPs. This allowed us to selectively modulate the amplitude of the red emission from the 4 F 9/2 excited state. Plotting of the corresponding Fourier transform amplitude of the pixel-wise acquired, modulated time traces allowed for selective lanthanide signal recovery and removal of 'unwanted' background emission. Application of more powerful visible and NIR laser systems will allow for expanding the presented approach to wide-field microscopy. This will enable for fast selective lanthanide signal recovery and imaging on the seconds time-scale. Our work shows how active control of excited state populations in lanthanides can open up new imaging and sensing applications for this interesting class of elements.