Effective Shielding of NaYF4:Yb3+,Er3+ Upconverting Nanoparticles in Aqueous Environments Using Layer-by-Layer Assembly.

Aqueous solutions are the basis for most biomedical assays, but they quench the upconversion luminescence significantly. Surface modifications of upconverting nanoparticles are vital for shielding the obtained luminescence. Modifications also provide new possibilities for further use by introducing attaching sites for biomolecule conjugation. We demonstrate the use of a layer-by-layer surface modification method combining varying lengths of negatively charged polyelectrolytes with positive neodymium ions in coating the upconverting NaYF4:Yb3+,Er3+ nanoparticles. We confirmed the formation of the bilayers and investigated the surface properties with Fourier transform infrared and reflectance spectroscopy, thermal analysis, and ζ-potential measurements. The effect of the coating on the upconversion luminescence properties was characterized, and the bilayers with the highest improvement in emission intensity were identified. In addition, studies for the nanoparticle and surface stability were carried out in aqueous environments. It was observed that the bilayers were able to shield the materials' luminescence from quenching also in the presence of phosphate buffer that is currently considered the most disruptive environment for the nanoparticles.

Introduction

Materials exhibiting upconversion luminescence[1] have been studied intensively for the last decade because of their potential use in photovoltaics,[2,3] biomedical imaging and theranostics,[4−6] and in biomedical assays.[7−9] Especially, the research in the biomedical field has benefitted from the extensive development of upconversion material syntheses, resulting in nanoscale materials down to 5 nm.[10−13] Currently, hexagonal β-NaYF4:Yb3+,Er3+ is considered to be the most efficient material in upconversion.[14,15] Its superiority is thought to arise from the low phonon energy in fluoride lattice[16] as well as shorter R–R distance compared with the cubic structure of NaYF4:Yb3+,Er3+.[17] Although the unique properties of upconversion nanoparticles have advantage in the biomedical applications, such as less scattering of light, photodamage of tissue, and minimal autofluorescence, they still have drawbacks that need to be solved.[18]

One of the most critical problems of upconverting nanoparticles use in biomedical applications is the significant quenching of the upconversion luminescence in aqueous environment due to OH– vibrations in water.[19] Additional processing steps are commonly required to obtain water dispersible materials, as most of the syntheses result in hydrophobic particles.[11,18] Furthermore, undesired disintegration of the used nanoparticles has been observed in water and in various aqueous buffers the most prominent being in phosphate buffer.[20,21] So far it is unclear which contributes most into emission intensity decrease in aqueous environment, the quenching of upconversion luminescence through ytterbium on the surface or the disintegration of the whole nanoparticle. One method of improving the emission intensity in aqueous environments is to produce more efficient materials. This has been made possible, for example, with core–shell structures[22,23] or using plasmonic enhancement.[24,25] However, especially in the cases where the shell is composed from similar fluoride structure as the core, it could be debated that it is also vulnerable to the disintegration mentioned previously. Recently, an amphiphilic coating was demonstrated to be successful in hindering the disintegration and thus shielding the particles’ luminescence from quenching.[26] Preventing the disintegration of the materials is vital, not only because of the optical changes harming the reproducibility of the measurements[20,21] but also to prevent the leaking of lanthanide and fluoride ions into the tissue or living cells in biomedical imaging.[27,28] This means that development of new surface modifications and methods to effectively stop the disintegration are increasingly important topics in the research of upconverting nanoparticles.

Our aim was to study the shielding of the upconverting nanoparticles with layer-by-layer method[29−32] and how varying the length of the polyelectrolyte (PE) chain affects the surface coating and luminescence properties. The layer-by-layer method has not been fully researched previously with upconverting nanoparticles and can provide multiple new insights into their surface modifications. Commonly, oppositely charged polyelectrolytes are used in forming the bilayers but it is also possible to use inorganic components such as positive metal ions.[33,34] The studied NaYF4:Yb3+,Er3+ nanoparticles were coated with negatively charged polyelectrolyte and positively charged neodymium ions to produce bilayers. The negatively charged polyelectrolyte is expected to attach on the nanoparticle surface similarly via carboxyl or phosphate group when coated with oleic acid. After the first negative polyelectrolyte layer, the buildup of bilayer coating is reached by varying the charge of the deposited layer.[35] Both poly(acrylic acid) (PAA) and polyphosphate (PP) were used as polyelectrolytes, and neodymium was chosen as positive ion because of its easy optical detection from bilayers and from the lanthanide ions at the core. Neodymium could also be used as a possible sensitizer for the 808 nm excitation.[36,37] The coating with varying polyelectrolyte molecular weights (and thus lengths) and its coverage was studied with small number of bilayers (1–3) due to previous observation that the advantage of shielding the core material and improving the luminescence seems to weaken after three bilayers.[35] The formation of the bilayers was studied with appropriate methods such as Fourier transform infrared (FT-IR) and reflectance spectroscopy to observe deposited polyelectrolytes and neodymium ions, respectively. Thermal analysis was used to get further insight on the bilayer formation and ζ-potential to observe changes on the surface. Upconversion luminescence measurements were made to powder materials using a 973 nm excitation at a power range of 3.2–11.1 W/cm2 to observe the changes in obtained emission due to the coating of the materials. In addition, upconversion luminescence decay and the stability of the nanoparticles with bilayers was investigated in aqueous suspensions, observing the emission intensity at selected time intervals using a 980 nm excitation.

Materials and Methods

Reagents

Poly(acrylic acid) (PAA; Mw ≈ 2000, 15 000, or 100 000), (C3H4O2), (Sigma-Aldrich), potassium polyphosphate (PP; long chain, potassium metaphosphate (KO3P), ABCR; short chain, sodium polyphosphate (NaPO3)), sodium chloride (NaCl 99.5%, J.T. Baker), and neodymium(III) chloride hexahydrate (99.9% purity with respect to other rare earths, Sigma-Aldrich) were used. Absolute ethanol was used as received. Phosphate buffer (5 mM) pH 7.75 and potassium fluoride (KF 99%, Merck) were used for the decay measurements along with a Greiner polypropylene microtiter plate (Sigma-Aldrich).

Materials Preparation

The β-NaYF4:Yb3+,Er3+ (xYb: 0.17, xEr: 0.03) nanoparticles (size ca. 22 × 26 nm) used as a core material were prepared with the synthesis procedure reported previously,[13] and the oleic acid present at the nanoparticle surface was removed with previously described acidic treatment before coating.[35,38] The coating solutions contained 10 mM PAA/PP (referring to monomer concentration of the polyelectrolyte) solubilized in 0.1 M NaCl (aq) and 10 mM NdCl3·6H2O in quartz-distilled H2O. The pH of the solutions was unadjusted and at ca. 5.

The coating cycle of the nanoparticles was the same as that reported previously.[35] The core particles (ca. 50 mg) were dispersed into coating solution and ultrasonicated for 2 min and washed with quartz-distilled water twice to remove any excess unattached material to prepare half a bilayer.

For the aqueous measurements, an additional layer of PAA (the length of polyelectrolyte being equal to that of polyelectrolytes used in the bilayers in PAA and the Mw 100 000 PAA for the PP) was deposited as the last layer on the surface with a similar procedure to provide shielding and attaching sites for possible further use.

Characterization

The core particle structure was determined at room temperature with the X-ray powder diffraction using a Huber G670 image plate Guinier camera (Cu Kα1 radiation, 1.5406 Å) with a 2θ range of 4–100° (step 0.005°). Data collection time was 30 min, and 10 data reading scans were collected from the image plate. From this data, the crystallite size of the core material was calculated with the Scherrer formula[39] using reflections (002) and (200) for the thickness and width of the hexagonal faces, respectively.

The surface ions were probed with reflection spectra measured using an Avantes Avaspec-2048 × 14 fiber spectrometer between 350 and 1000 nm. As a light source, a 60 W incandescent light bulb was used and the measurements were done with an integration time of 400 ms and 30 scans. The presence of the polyelectrolytes was studied with Fourier transformed infrared spectra (FT-IR) using a Bruker Vertex 70 using MVP Star Diamond setup with 32 scans between 450 and 4500 cm–1. The resolution was 4 cm–1. The elemental composition of the products was studied with X-ray fluorescence spectroscopy (XRF) with PANanalytical Epsilon 1 apparatus using its internal Omnian calibration with four scans. The qualitative thermal behavior of the coated materials was studied with a TA Instruments SDT Q600 thermogravimetric analysis (TGA)–differential scanning calorimetry apparatus. The error for weight in this setup is ca. 1%. The measurements were made between 35 and 600 °C with a heating rate of 10 °C/min using flowing air sphere (100 mL/min). The ζ-potential to evaluate surface coverage was measured with a Malvern Zetasizer Nano-ZS equipment using three replicate measurements. The pH of the aqueous solutions was ca. 6, and the nanoparticle concentration 100 μg/mL. The images of the coated nanoparticles were obtained with a JEM-1400 Plus transmission electron microscope (TEM) using an OSIS Quemesa 11 Mpix bottom-mounted digital camera. The nanoparticles were suspended into ethanol and dried on a lacey carbon grid and then imaged with an acceleration of 60 kV.

The upconversion luminescence spectra were measured at room temperature with an Avantes Avaspec HS-TEC spectrometer. A fiber-coupled continuous near-infrared laser diode IFC-975-008-F (Optical Fiber systems) with an excitation wavelength of 973 nm (10 270 cm–1) was used as an excitation source with a power density range of 3.2–11.1 W/cm2. Dry nanomaterials were held inside a rotating capillary tube. After the sample, a short-pass filter with a cutoff of 750 nm (Newport) was used to exclude excitation radiation. The emission was collected at a 90° angle to the excitation and directed to the spectrometer with an optical fiber. The downshifted emission and excitation spectra were measured for a dry nanomaterial with a Varian Cary Eclipse spectrofluorometer. Phosphorescence mode with 0.1 ms delay and 0.5 ms gate time was used to record the spectra.

The time-dependent upconversion luminescence decay measurements in aqueous media (particle concentration 5 μg/mL) were carried out using 980 nm excitation and a lifetime measurement mode with a modified Plate Chameleon fluorometer.[40] The luminescence decays were measured after incubating the nanoparticles for 0, 4, or 24 h in pure water or 5 mM phosphate buffer (pH 7.75) with or without 1 mM KF in slow shaking. The measurements were done in Greiner polypropylene microtiter plates. The samples were exposed repeatedly to 2 ms pulsed excitation, and the emissions were collected using a 535/40 nm band-pass filter. Obtained data were analyzed using second-order exponential decay fitting of Origin 8 (OriginLab, Northampton, MA).

Results and Discussion

Success of the Coating Process and Coverage

The β-NaYF4:Yb3+,Er3+ upconversion materials used for the layer-by-layer coating were prepared with the method described earlier[13] and were of hexagonal form with a size ca. 22 × 26 nm (Table S1). As the synthesis produces oleic acid-capped nanoparticles, the oleic acid had to be removed before the coating. The removal was confirmed with the FT-IR spectra where this could be easily seen as a reduction of the asymmetric and symmetric vibrations of COO– (Figure S1). After removal of the oleic acid surface, the bare particles were used as a core for the coating process.

The FT-IR spectra showed that the desired polyelectrolytes were present in all of the coated materials (Figure S2). The strongest vibrations from PAA and PP could be distinguished at different sections in the recorded spectra, the main vibrations of PAA being around 1000–1750 cm–1 and of PP being around 900–1400 cm–1.[41] No quantitative conclusions could be made from FT-IR spectra, but with the materials prepared with PAA Mw 100 000 and both lengths of PP having three bilayers, the polyelectrolyte-related vibrations seemed to be the strongest (Figure S2). This is a good indicator that there should be more polyelectrolyte present on the material. In addition, the reflectance spectroscopy revealed the presence of neodymium ions on the surface (Figure S3). The absorption of neodymium ions (575, 740, 795, and 870 nm) was easily distinguishable from the ones arising from the core lanthanide ions (490, 520, 540, 640, and 800 for Er3+ and 975 nm for Yb3+) (Figure ). The absorbance of neodymium ions was increased with the increasing number of bilayers and the increase of the PAA polyelectrolyte length, which implies successful layer formation. This was expected as the longer PAA chain has more attaching sites to capture neodymium ions on to the structure.

Figure 1

Reflectance spectra of the core and coated NaYF4:Yb3+,Er3++@(PE/Nd3+)3 materials prepared with selected polyelectrolyte lengths PAA (Mw: 2000, 15 000, and 100 000) and short- and long-chain PP.

The shorter chain PP seemed to be better in attaching the neodymium ions than the longer chain. The polymer chain is expected to be similarly open in all of the deposition solutions as the ionic concentration was unchanged.[42] In addition, the pH of the solutions was constant so the percentage of open attaching sites is expected to be similar in all polyelectrolytes.

The increase in the number of bilayers with PAA/Nd3+ should be seen as an increase in the weight loss during the thermal analysis (Figure S4). This was true with both Mw 2000 and 100 000 coated materials. In addition, the shape of the TGA curves of the second and third bilayer are different from that of the first bilayer, suggesting that the bilayer formation is more defined due to the increasing hydration layer (water) on the material surface. However, with the Mw 15 000, the weight loss is decreasing with increasing number of bilayers and the weight loss curve of the first bilayer behaves differently from the others. It is possible that with this length of PAA, the bilayer formation is not sufficient at the first bilayers but only after further coating cycles. With short-chain PP layers, the thermal analysis showed a decrease in the weight loss with increasing layers, suggesting that more polyphosphate is present in the materials as the polyphosphate should remain unaffected during the heating (Figures and S4). The behavior in TGA curves with increasing number of bilayers is similar to that of TGA curves obtained with fire-retardant ammonium polyphosphate,[43] suggesting that the bilayers shield the inner particle and possible remaining impurities in the structure, making the weight loss smaller. There is also a possibility that the nearby neodymium ions interact with the polyphosphate, forming Nd3O3PO4 that could contribute to the weight gain but its contribution is expected to be minimal.

Figure 2

Thermal analysis of the NaYF4:Yb3+,Er3++@PAA/Nd3+ materials coated with Mw (PAA) 2000 (top) and short-chain polyphosphate (bottom) with deposition concentrations of 10 mM PP/PAA in 0.1 NaCl (aq) and 10 mM NdCl3 (aq).

The increase of neodymium at the materials with the increase in the number of bilayers could be observed from the XRF measurements (Table S2 and Figure S5). With varying PAA lengths, the increase was linear but small. The only distinguishable difference observed was that, as with the reflectance spectroscopy, the longest polyelectrolyte was the most efficient in attaching the neodymium ions. With PP, the increase in neodymium amount was also linear. However, some differences between the amounts of phosphorus were observed. With the short chain, the increase was linear, but with the longer chain, it could be debated that the increase after two bilayers indicates exponential growth in the amount of phosphorus. This could be attributed to the fact that a longer polyphosphate chain could develop branched structures around the particle because the chain ends reaching out from particle surface are able to also host the next oppositely charged layer formation.

The ζ-potential was used to observe differences in the surface charge with increasing number of bilayer formation. All of the core materials were observed to have a similar ζ-potential (ca. 20 mV), as reported previously from the bare upconversion nanoparticle core.[26] With all materials, the ζ-potential decreased with increasing number of bilayers, as expected. The positive charge arising from the Nd3+ ions is negligible in comparison to the negative charge of the polyelectrolytes. However, the ζ-potential decrease between the first and second bilayer was small, which suggests that even though the coating was observed, it is possible that it might not occur linearly but in a steplike manner (Figure ). With different lengths of PAA, it was observed that the coating was most efficient during the first bilayer with the shortest polyelectrolyte (Mw (PAA): 2000), as expected, and the longer two (Mw (PAA): 15 000 and 100 000) behaved similarly during the first layer. When three bilayers were formed the smallest two seemed to form better coverage, resulting in a lower ζ-potential (ca. −7 mV) than that of the longest PAA (ca. 3 mV). With polyphosphate used as a coating component, the bilayer formation was also observed being similar to that with varying PAA lengths. The shorter polyphosphate seemed to cover more of the particle surface on the first bilayer, but eventually both lengths of the polyphosphate reached a similar ζ-potential of −5 mV at three bilayers (Figure S6). The ζ-potential of all coated materials except the longest PAA reached between −5 and −10 mV at three bilayers, suggesting that the coverage at that certain stage could be similar regardless of the length of the polyelectrolyte or the pathway of each bilayer formed.

Figure 3

ζ-Potential of the core and NaYF4:Yb3+,Er3+@PAA/Nd3+ materials coated with deposition concentrations of 10 mM PAA in 0.1 NaCl (aq) and 10 mM NdCl3 (aq). The lines are only a guide for the eye.

Transmission electron microscopy (TEM) imaging was used to probe if the bilayers were visible on the nanoparticle surface. To observe the delicate organic structure on the surface, the suspended materials were dried on the lacey carbon grid to minimize the background and imaged with a low 60 kV voltage. From the images, some indication of the bilayers could be drawn, as the surface of the materials is fuzzy and similar to those obtained previously (Figure ).[35] The imaging grid and the used solvent can have an effect in aggregating the nanoparticles as they seem to attach to the carbon fiber edges. However, with this small number of bilayers, it is unclear if the ca. 2 nm rings around the nanoparticles are actually the bilayers or combination of bilayers and solution boundaries.

Figure 4

TEM image of the NaYF4:Yb3+,Er3+@(PAA/Nd3+)3 materials coated with PAA Mw 2000. The scale bar is 50 nm and the inset zoom is 150%.

Luminescence Properties

The upconversion luminescence of the coated nanoparticles was studied by a 973 nm (10 280 cm–1) excitation, which is suitable for the well-known energy-transfer upconversion from ytterbium ion to erbium.[44,45] The characteristic emission of Er3+ in green (2H11/2, 4S3/2 → 4I15/2) and red (4F9/2 → 4I15/2) wavelength regions was observed from all of the prepared materials.[45] With the materials coated with varying PAA lengths, it was notable that with the shortest PAA (Mw: 2000), the most intense upconversion luminescence was observed from the material with three bilayers (Figure S7). With others PAA lengths, the first bilayer was found to be the most effective in shielding the luminescence. The strongest enhancement in comparison to that of the core material, ca. 2.5 times stronger, was observed with one bilayer of Mw 15 000 PAA (Figure ). With the varying lengths of PP, similar enhancement was observed with lower bilayer amounts but the overall effect was weaker than that with PAA.

Figure 5

Upconversion luminescence spectra excited with 973 nm of the core and NaYF4:Yb3+,Er3+@PAA/Nd3+ materials coated with PAA Mw 15 000.

When the intensity of emission from green and red energy levels is plotted against the excitation power density in a log–log plot, its linearity gives information about the photons needed for the upconversion process. The obtained slopes for the green emission were ca. 2, indicating a two-photon process, as expected (Table S3). With the red emission, the slopes were generally higher than those previously achieved with oleic acid-capped materials[13,46] and in some of the materials raised as high as ca. 2.5. With larger (ca. 90 nm) oleic acid-capped particles, this higher slope for red emission has been observed and is expected to be due to the back energy-transfer process to the Yb3+ sensitizer.[44] In addition, the presence of H2O molecules seems to increase the slope and lead to the population of the red emitting state via green emitting levels without the back transfer process.[45] To study this further, downshifted emission and excitation spectra were measured with varying excitation wavelengths to probe the possible pathway to the red upconversion luminescence (Figure S8). We observed that the red emission is obtained regardless of the excitation wavelength. With the spectra obtained from the materials prepared by Berry et al. and Hyppänen et al., in both the excitation spectra of the materials dispersed in toluene and D2O, respectively, the red emission lacked the component from the green emitting states of Er3+ (2H11/2, 4S3/2).[44,45] However, with our dry materials, these energy levels are clearly visible (Figure S9). In materials prepared with shorter PAA lengths (Mw: 2000 and 15 000), the red emission has less input from the green emitting levels than from the bare core materials. In the case of longer PAA (Mw: 100 000), this difference is not so clear, suggesting more involvement from H2O molecules that are more easily interlocked into the longer polyelectrolyte structures. So we believe that in our materials, the H2O molecules play a significant role as the surface quencher[19] both in the oleic acid-free core materials as well as in the coated materials because of the probability of interlocked H2O molecules within the layer structure. This leads to the red emission using a combination of both two- and three-photon processes depending on the environment of the ytterbium and erbium ions confirming the obtained slope values for green and red emissions.

The time-dependent behavior of the coated nanomaterials in water suspension was measured with a modified Plate Chameleon fluorometer using 980 nm excitation and detecting the green upconversion luminescence at 535 nm.[40] The emissions were collected directly after suspension to aqueous solution. After a 2 ms wide pulsed excitation, the data was collected for 8 ms. The obtained data was fitted (after 50 μs to exclude the delayed energy transfer from Yb3+ to Er3+) with a second-order exponential decay function. It must be noted that the fitted decay times of the green emission arise from complex excitation and emission systems between the active ions as well as long-lived intermediate energy states and not only from the emitting energy levels of erbium.[47] When the lifetime components of the coated materials are compared with those of core materials it could be observed that there is a decrease especially in the first component with shorter polyelectrolyte chains. However, the changes in the second lifetime component for the PAA-coated materials is reasonably small and for the longest PAA (Mw: 100 000) the decrease is negligible (Tables , S4, and Figure S10). The difference between the lifetimes of the different core materials (and their coated products) is expected to arise from the size differences (ca. 20 × 25 and 23 × 29 nm), inducing differences in energy migration through ytterbium energy levels within the particles,[47] as the larger particles’ lifetimes are similar to those obtained previously.[20,45] It could be observed that the percentage share of the lifetime components shifts slightly to the second component after coating. This seems reasonable as the longer lifetime is thought to be connected to the inner Er3+ ions at the core that are more shielded from the environment.

Table 1

Upconversion Luminescence Lifetimes of the Coated Nanoparticlesa

sample	polyelectrolyte	τ1 (μs)	A (%)	τ2 (μs)	A (%)
core		71 ± 1	94	668 ± 21	6
1 bilayer	PAA Mw: 100 000	69 ± 1	93	683 ± 22	7
2 bilayers		69 ± 1	92	636 ± 21	8
3 bilayers		66 ± 1	92	654 ± 15	8

The amplitude indicates the % effect in the total lifetime.

When the upconversion emission intensity during the time-dependent measurements was studied in aqueous suspension with the shorter chains of PAA (Mw 2000 and 15 000), the intensity did not change significantly when the core was coated. However, with the longest PAA (Mw 100 000), the upconversion luminescence intensity increased with the number of bilayers being strongest with two bilayers. With both lengths of polyphosphate, the emission intensity decreased after the first bilayer, suggesting that multiple bilayers did not help preserving the upconversion luminescence (Table S4 and Figure S10).

Disintegration of the Coated Particles

The stability of the coating and possible disintegration of the coated particles was studied with upconversion luminescence decay measurements in three time points after suspension to liquid media. The disintegration of the nanoparticles can be observed through the changes in their emission both in intensity and in lifetime components already during the first hours after introducing them into aqueous buffers.[20,21] Adding 1 mM KF in aqueous suspension has been reported to hinder or even prevent the disintegration of the particles in most used buffers, excluding the phosphate-based buffers.[20] For this reason, our measurements were conducted both in pure water and in phosphate buffer with or without the addition of 1 mM KF to observe differences in the behavior of the material’s upconversion luminescence intensity and decay at time points 0, 4, and 24 h.

Without the addition of KF in pure water, the detected emission at 535 nm decreased from 0 to 4 h in all of the materials with PAA/Nd3+ bilayers but no significant decrease followed in the remaining 24 h. However, with the longest PAA (Mw: 100 000) coated with three bilayers of PAA/Nd3+, the overall emission decrease was smallest from the coated PAA materials during the measurement period (Figures and S11). This suggests that the bilayers shield the particle for the disintegration in water. A similar conclusion could be drawn when the lifetime components were considered (Figure S11 and Table S5); the second lifetime decreased less from 4 to 24 h than what it does between 0 and 4 h. Only with the longest PAA, the second lifetime decreased less than 100 μs during the aging period, confirming that the particle is shielded from the disintegration. As the second lifetime is thought to be connected to the inner core of the particles, its significant decrease is considered as an indicator of disintegration.[20] With all of the materials coated with PAA/Nd3+ bilayers, the addition of 1 mM KF in pure water prevented the emission decrease for 24 h, as expected. The small emission intensity increase especially in the core material measurements is thought to arise from possible aggregation of the particles in the measurement well during the incubation periods.

Figure 6

Upconversion luminescence decay spectra excited with 980 nm and detected at 535 nm at time steps 0, 4, and 24 h for NaYF4:Yb3+,Er3+@(PAA/Nd3+)3PAA material (Mw: 100 000) measured in pure water.

When the PAA/Nd3+-coated nanomaterials were introduced into phosphate buffer, their behavior changed from that when suspended in pure water. In the phosphate buffer, the signal remained similar until the 4 h measurement and then decreased when 24 h was achieved. The behavior of the second lifetime component is similar, suggesting that the phosphate cannot interfere with the particle structure within the time scale. In addition, when introduced in phosphate buffer with 1 mM KF, the emission signal and the lifetime components could be maintained throughout the whole 24 h measurement window. This is an improvement to previous results where the various upconversion nanomaterials were seen to lose their luminescence intensity when introduced in phosphate buffer already in the first 4 h of time course, and this decrease happened even with the addition of KF in solution.[20,21]

With both lengths of polyphosphate, a behavior similar to those with PAA/Nd3+ bilayers was observed with and without the 1 mM KF addition into solution (Figure S11 and Table S5). Only the first bilayer of PP/Nd3+ with short PP was able to shield the emission and disintegration for the first 4 h of measurement in pure water. In all other cases, the emission decreased significantly already during that time period. The addition of 1 mM KF into the measurement solutions was able to prevent the emission decrease in the pure water-based measurements except with the additional number of bilayers of short polyphosphate. With measurements in phosphate buffer, the bilayers prepared with longer polyphosphate were able to shield the emission intensity for the first 4 h, as with the PAA/Nd3+ bilayers. Only the first bilayer of PP/Nd3+ with short polyphosphate was similar to that of long polyphosphate; with increasing number of bilayers of the short polyphosphate, the emission decreased significantly. In the phosphate buffer-based measurements, the 1 mM KF addition was effective only with the materials coated with longer polyphosphate, suggesting that the short polyphosphate in the bilayers interacts and detaches faster when in contact with the buffer solution.

Conclusions

We have demonstrated that using various lengths of polyelectrolytes, the layer-by-layer method can be utilized for surface passivation and can offer further functionalization of upconverting nanoparticles. In addition, we observed that selected bilayer formations were able to shield the upconversion luminescence and thus prevent the disintegration of the nanoparticles even in the phosphate buffer that has been proven to enhance the disintegration.[20,21,48]

Overall, the modifications made with poly(acrylic acid) showed better bilayer formation than with those made with polyphosphate. The length of the used polyelectrolytes had most influence on the first formed bilayer, but with increasing number of bilayers, the effects leveled. The most efficient enhancement to overall upconversion luminescence intensity was obtained with the first bilayer of PAA/Nd3+ prepared with PAA Mw 15 000. The coating with short PP had a negative effect on the upconversion luminescence lifetimes, whereas others showed no significant decrease in the lifetime components.

The disintegration of the nanoparticles was investigated with both pure water and the phosphate buffer for 24 h. It was observed that with selected number of bilayers of PAA/Nd3+, the disintegration in pure water could be prevented without additional fluoride during the whole 24 h investigation. Also, all of the materials excluding short polyphosphate showed resistance to disintegration for the first 4 h when introduced in the phosphate buffer. Addition of 1 mM KF into the phosphate buffer solution also enabled to prevent the disintegration and emission intensity loss during the measurement window.

The results presented in this study provide new pathways for the design of upconverting nanoparticles and their surface modifications. Although we acknowledge that more studies are needed in both the layer-by-layer process and the layer stability in these kind of hybrid materials in the future, we also believe that this research brings additional possibilities to find a solution to hinder and prevent the disintegration of these delicate nanoparticles, especially in the phosphate buffer-based aqueous environments.

Table 2

Upconversion Luminescence Lifetimes of the (PAA/Nd3+)3PAA-Coated Nanoparticles (Mw (PAA): 100 000) at 5 μg/mL Measured at 0, 4, and 24 h in Water, Water + 1 mM KF, Phosphate Buffer, and Phosphate Buffer + 1 mM KFa

h	μs	H2O	A (%)	+KF	A (%)	phosphate	A (%)	+KF	A (%)
0	τ1	66 ± 1	92	67 ± 1	92	67 ± 1	92	68 ± 1	92
	τ2	654 ± 11	8	657 ± 14	8	655 ± 14	8	664 ± 15	8
4	τ1	63 ± 1	92	67 ± 1	92	65 ± 1	92	66 ± 1	92
	τ2	590 ± 12	8	653 ± 12	8	611 ± 11	8	660 ± 14	8
24	τ1	62 ± 1	92	65 ± 1	92	91 ± 1	92	67 ± 1	92
	τ2	578 ± 13	8	638 ± 12	8	b		620 ± 12	8

The amplitude indicates the % effect in the total lifetime.

The phosphate buffer 24 h measurement could only be fitted reasonably with first-order exponential decay.