View Article Online

View Journal

Nanoscale
Accepted Manuscript

This article can be cited before page numbers have been issued, to do this please use: M. Kaiser, C.
Würth, M. Kraft, I. Hyppänen, T. Soukka and U. Resch-Genger, Nanoscale, 2017, DOI:
10.1039/C7NR02449E.

Volume 8 Number 1 7 January 2016 Pages 1–660 This is an Accepted Manuscript, which has been through the
Royal Society of Chemistry peer review process and has been
accepted for publication.
Nanoscale Accepted Manuscripts are published online shortly after
www.rsc.org/nanoscale

acceptance, before technical editing, formatting and proof reading.

Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.

You can find more information about Accepted Manuscripts in the

author guidelines.

Please note that technical editing may introduce minor changes

to the text and/or graphics, which may alter content. The journal’s
ISSN 2040-3364 standard Terms & Conditions and the ethical guidelines, outlined
PAPER
in our author and reviewer resource centre, still apply. In no
Qian Wang et al.
TiC2: a new two-dimensional sheet beyond MXenes
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.

rsc.li/nanoscale
 Page 1 of 9 Please doNanoscale
not adjust margins

View Article Online

DOI: 10.1039/C7NR02449E
Nanoscale

ARTICLE

Power-dependent upconversion quantum yield of NaYF4:Yb3+,Er3+

nano- and micrometer-sized particles – Measurements and
Published on 13 June 2017. Downloaded by University of California - San Diego on 13/06/2017 16:11:38.

Received 00th January 20xx,

Accepted 00th January 20xx
simulations
a a a b b
Martin Kaiser, Christian Würth, Marco Kraft, Iko Hyppänen, Tero Soukka, and Ute Resch-

Nanoscale Accepted Manuscript

DOI: 10.1039/x0xx00000x
a,*
Genger
www.rsc.org/

Photophysical studies of nonlinear lanthanide-doped photon upconverting nanoparticles (UCNP) increasingly used in
biophotonics and photovoltaics require absolute measurements of excitation power density (P)-dependent upconversion
luminescence (UCL) and luminescence quantum yields (UC) for quantifying material performance, UCL deactivation
pathways, and possible enhancement factors. We present here the P-dependence of UCL spectra, UC , and slope factors
of the different emission bands of representative 25 nm-sized oleate-capped -NaYF4:Yb3+: 17 %, Er3+: 3% UCNP dispersed
in toluene and as powder as well as UC of 3 µm-sized upconversion particles (UCP), all measured with a newly designed
integrating sphere setup, enabling controlled variation of P over four orders of magnitude. This includes quantifying the
influence of the beam shape on measured UC and comparison of experimental UC with simulations utilizing the balancing
power density model of the Andersson-Engels group and simulated UC of UCP from the Berry group, underpinned by
closely matching decay kinetics of our UC material. We obtained a maximum UC of 10.5 % for UCµP and UC of 0.6 % and
2.1 % for dispersed and solid UCNPs, respectively. Our results suggest an overestimation of the contribution of the purple
and an underestimation of that of the red emission of -NaYF4:Yb3+, Er3+: microparticles by the simulations of the Berry
group. Moreover, our measurements can be used as guideline to absolute determinations of UCL and UC.

19-22
(Nem) to absorbed photons (Nabs), see eq. 1. For
Introduction upconversion (UC) materials, UC is defined as ratio of the
number of emitted photons Nem with a wavelength em < abs,
Lanthanide-doped photon upconversion nanoparticles (UCNPs)
obtained from integral emission intensities of all UC emission
are an emerging class of multiphotonic reporters for
bands (except for sensitizer luminescence), and Nabs, see eq.
fluorescence assays and bioimaging studies as well as
For arbitrary P, UC(P) can be simulated with a
19, 20, 23-25
2.
functional materials for sub-bandgap energy harvesting in
simplified rate equation model of a two level-emitter system
photovoltaics, and security printing which reveal a multitude 26
(eq. 4 developed by the Andersson-Engels group), using a)
of sharp luminescence bands in the ultraviolet (UV), visible
(vis), and near-infrared (NIR) upon sequential absorption of
UC,sat, obtained from converging of UC(P) at high P and b) the
1-13 balancing power density Pbalance, where UC(Pbalance) equals
two or more NIR photons. Although UCNP can be
meanwhile tuned in crystal size, phase, and emission color
UC,sat/2. This approach is strictly only valid for a purely
14 biphotonic emission like the green or the 800 nm
through refined synthetic procedures, full exploitation of 3+/ 3+ 3+ 3+
luminescence in Yb Er and Yb /Tm systems, respectively,
their potential is hampered by the need to improve their
and monodisperse and uniform particles (eq. 3). Also other UC
brightness. This requires strategies to brighter UCNP utilizing
luminescence (UCL) features like the photonic nature or slope
sophisticated core-shell architectures, plasmonic structures,
15-18 factor n (eq. 4), which provides the number of photons
and dye sensitization and tools to reliably characterize and 27
required to populate a certain emissive electronic state, are
quantify the excitation power density (P) dependent
P-dependent. This results in a decrease of n with increasing P,
luminescence of UCNPs, especially their photoluminescence
for P exceeding a certain material-, size-, and matrix-
quantum yields (), i.e., the ratio of the number of emitted 26, 27
dependent level. This must be considered for mechanistic
studies.
a.
Federal Institute for Materials Research and Testing (BAM), Division Biophotonics 𝜙 = 𝑁em /𝑁abs (eq. 1)
(1.10), Richard-Willstaetter-Str. 11, D-12489 Berlin, Germany
b.
University of Turku, Department of Biotechnology, Tykistökatu 6A, Turku, FI-
𝜙UC (𝑃) = 𝑁em /𝑁abs, for em < abs (eq. 2)
20520, Finland.
Electronic Supplementary Information (ESI) available: [details of any 𝑃balance
supplementary information available should be included here]. See 𝜙UC (𝑃) = 𝜙UC,sat /( + 1), (eq. 3)
𝑃
DOI: 10.1039/x0xx00000x

This journal is © The Royal Society of Chemistry 20xx Nanoscale., 2017, 00, 1-3 | 1

Please do not adjust margins

 Please doNanoscale
not adjust margins Page 2 of 9

ARTICLE Nanoscale

for a biphotonic process with our setup (see Supporting Information (SI),View Fig. S1 Online
Article and
𝑛 S2), as well as an ideal Top Hat (TH,ideal)DOI:
BP. 10.1039/C7NR02449E
𝐼UC ~𝑃 , for low P (eq. 4)
∞ 𝜙UC,sat
∭−∞ 𝑃BP (𝑥,𝑦,𝑧) d𝑥d𝑦d𝑧
Nevertheless, often only single UC values for a certain, often 𝑃balance
𝑃BP (𝑥,𝑦,𝑧)+1
unknown P are reported to emphasize the performance of a ϕUC (𝑃BP ) = ∞ (eq. 5)
∭−∞ 𝑃BP (𝑥,𝑦,𝑧) d𝑥d𝑦d𝑧
new UC material and data comparability is hampered by
different excitation sources (different excitation wavelengths), Subsequently, UC(PBP) was calculated for UC,sat = 50 % and
-2
excitation geometries, beam profiles (BPs), and P. The Pbalance = 1 Wcm as detailed in the SI (see Fig. S3 and eq. S1
reliability of often performed relative UC determinations or and S2). Beam deformation due to absorption in the sample
determinations of enhancement factors with a volume was not considered as this is negligible for an
spectrofluorometer
28
or an integrating sphere, using an absorption < 10 %. The results are highlighted in Fig. 1 (panel
Published on 13 June 2017. Downloaded by University of California - San Diego on 13/06/2017 16:11:38.

29
organic dye like IR26 or another UC material as reference
30 a)). Panel b) of Fig. 1 illustrates the quotients resulting for the
has not yet been confirmed. The recently recognized need for different UC(PBP), i.e., UC(PGauss,exp)/UC(PTH,ideal),
quantitative UCL measurements has turned the determination UC(PGauss,exp)/UC(PTH,exp), and UC(PTH,exp)/UC(PTH,ideal)). The
of UC and saturated UC (UC,sat) one of the ultimate relatively close match between UC(PTH,exp) and UC(PTH,ideal)

Nanoscale Accepted Manuscript

5
cornerstones of UC nanotechnology to identify optimum
strategies to brighter UCNP. This has meanwhile triggered an
increasing number of reports on absolute
with the highest UC values of Yb ,Er -
31-41
measurements,
based systems obtained so far being 7.8 % for a µm-sized
3+ 3+ -2 38
NaYF4:Yb ,Er powder at P of 22 Wcm , and 5 % for 40 nm-
3+ 3+
sized core-shell LiLuF4:Yb ,Er particles dispersed in hexane at
-2 39
P of 127 Wcm , respectively, but detailed descriptions of
instrument design and actual performance of UC
measurements are commonly missing. Furthermore, P is
commonly at best varied over up to two orders of magnitude,
which is often not sufficient for UC saturation.
To close this gap, we assess here different strategies for the
absolute determination of P-dependent UCL with a custom
designed integrating sphere setup representatively for high
quality monodisperse 25 nm-sized oleate-capped -NaYF4:Er :
3+
3%, Yb :17 % particles dispersed in toluene and as powder as
3+
well as commercial 3 µm-sized -NaYF4:Er : 2.2 %, Yb :21.4 %
particles (UCµP). The latter, used as representative bulk
42
material, were chosen to closely match particles
characterized by van Veggel in his pioneering work on UC
32
measurements of dispersed UCNP and UCµP studied by the
23
Berry group. Our experimental results are compared with
simulations utilizing the balancing power density/saturation
model and simulated UC of UCµP from the Berry group.
26
with deviations of maximally 13 % suggests a rather
homogeneous P distribution for TH,exp and highlights the
UC suitability of our setup design for the quantitative
characterization of UC materials. A Gaussian BP leads to an
3+ 3+
overestimation of UC, yielding values that exceed the
corresponding UC(PTH,ideal) and UC(PTH,exp) data by about 50 %
and 30 %, respectively. These deviations are most pronounced
in the unsaturated P-regime.
n(𝑃BP ) = 𝑑 log(𝐼UC (𝑃BP ))/𝑑 log(𝑃), (eq. 6)
The P-dependent slope factors n(P) of the different UCL bands
were determined with eq. 6. As follows from panel c) of Fig. 1,
n(P) can be divided into an unsaturated and a saturated P
regime with constant slope values of 2 and 1, and a semi-
saturated regime reflected by a steady reduction of n from 2
3+
approaching 1 for saturation. BP-dependent effects are
observed only in the semi-saturated P-regime. (see Fig. 1,
3+
panel c)) Moreover, Fig. 1 demonstrates that n(Pbalance)
obtained for TH,ideal. equals 1.5 whereas for an
inhomogeneous BP like Gauss,exp, n(Pbalance) deviates
obviously from 1.5. These findings clearly underline the
considerable impact of the homogeneity of the P distribution
on resulting UC(P) and n(P).
23

Additionally, P-dependent intensity contributions of the green

2 4 4 4 4
(520, 540 nm, H11/2, S3/2 → I15/2), red (655 nm, F9/2 → I15/2),
2 4 3+
and purple (410 nm, H9/2 → I15/2) emission of Er to overall
UCL are discussed as well as n values and decay kinetics.

Results and Discussion

Simulation of the influence of the excitation beam profile (BP) on
UC(P) and n(P).
As a first step to quantitative UCL measurements at broadly
varied P, we assessed the influence of BP on UC(P) with the
aid of simulations. To consider BP-dependent effects, eq. 3
was modified to yield UC as function of the local distribution
of the excitation power PBP(x,y,z), see eq. 5. As representative
BPs for the simulation a Top Hat (TH,exp) and Gaussian
(Gauss,exp) BP were used, which were experimentally realized

2 | Nanoscale., 2017, 00, 1-3 This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

 Page 3 of 9 Please doNanoscale
not adjust margins
Nanoscale
Published on 13 June 2017. Downloaded by University of California - San Diego on 13/06/2017 16:11:38.
Figure 1. a) Simulation of UC(PTH,ideal) (black line), UC(PTH,exp) (blue line), and UC(PGauss,exp) (red line) with UC,sat = 50 % and Pbalance = 1 Wcm-2 for a path length of 10 mm assuming a
biphotonic optical transition. b) Quotients of UC(PTH,exp)) and UC(PTH,ideal) (cyan line) and UC(PGauss,exp) and UC(PTH,ideal) (orange line), respectively. For comparison with our actual
measurements shown in Fig. 2, we also included the quotient of UC(PGauss,exp)) and UC(PTH,exp) (violet line). c) Corresponding n(P) for a biphotonic optical transition like the green
emission obtained for the three BPs with eq. 6.
Influence of beam profile on UC, n, and UCL intensity ratios
for dispersed UCNP.
We subsequently assessed the P-dependence of UC, n, and
the corresponding intensities IUC(em,P) of the green (520, 540
2 4 4 4 4
nm, H11/2, S3/2 → I15/2), red (655 nm, F9/2 → I15/2), and purple
2 4 3+
(410 nm, H9/2→ I15/2) UCL of Er (given as spectral photon
-1 -3
fluxes (s m )) for the TH,exp and Gauss,exp BPs. As follows
from Fig. 2 (panels a) and b)), with deviations in the order of
20-30%, UC(PTH,exp) and UC(PGauss,exp) agree reasonably well
with the results of our simulations (see Fig. 1, panel b)). It has
to be taken into account here that this comparison is only
qualitative, since our determination of the UC values of the
UCNP also includes the intensity of triphotonic emission bands.
Fitting of UC(P) of the purely biphotonic green emission yields
UC,sat,green of 1 % with Pbalance of 550 Wcm-2, see SI (Fig. S4 and
Table S2; deviation of maximum 10 %). A possible explanation
for the increasing deviation of measured and simulated UC(P)
-2
data for P > 1000 Wcm (see SI, Fig. S4, left panel) can be the
more pronounced contribution of triphotonic processes for the
Gauss,exp BP. Also particle-to-particle variations in size,
dopant ion concentration, and number of lattice and surface
defects can lead to a distribution of particles with different
UC and Pbalance especially for small UCNP with a high surface-
to-volume ratio as discussed e.g., for semiconductor quantum
dots.
43, 44
3+ 3+
In the case of NaYF4: Yb , Er UCNPs, main contributions to
UC(P) in the order of at least 70% arise from the green and
red emission. A closer look at the integral emission intensities
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins
ARTICLE
View Article Online
DOI: 10.1039/C7NR02449E
Nanoscale Accepted Manuscript
of the different UCL bands IUC(em), given in percentage of the
overall integral UCL intensity UC(all); see Fig. 2, panel c),
shows almost identical trends (deviation < 10 %) for TH,exp
and Gauss,exp BPs despite the different photonic orders of the
23, 45
green and red emission. Contrary, the intensity of the
triphotonic purple emission reveals a more pronounced
increase for Gauss,exp compared to TH,exp BP. This is ascribed
to the higher P near the focal point for the former BP. For our
UCNP, the green emission is most pronounced at low P values
-2
up to 20 Wcm in the unsaturated P regime, accounting for
about 42 % of the overall UCL. Upon onset of UC saturation in
the high P regime, the contribution of the green emission
decreases to 27 % whereas the red emission becomes more
pronounced, rising from initially 30 % at low P to 55 % in the
-2
saturated P regime of about P > 2000 Wcm . This underlines
the potential of varying P for color tuning. The change in the
red-to-green intensity ratio (Ig/r) (see Fig. S6) in the P regime of
-2
10-2000 Wcm is ascribed to back energy transfer (BET) from
3+ 3+ 4 4
Er to Yb , with the red emission band ( F9/2 - I15/2 transition)
4 45
being efficiently populated from the G11/2-level. Our
observation that the intensity of the red emission converges at
-2
about 1000 Wcm while the purple emission still increases
4 4
steadily suggests prevention of BET between the G11/2→ F9/2
3+ 2 2 3+
transition of Er and the F5/2 → F7/2 transition of Yb for
saturation.
Nanoscale., 2013, 00, 1-3 | 3
 Please doNanoscale
not adjust margins Page 4 of 9

ARTICLE Nanoscale

View Article Online

DOI: 10.1039/C7NR02449E
Published on 13 June 2017. Downloaded by University of California - San Diego on 13/06/2017 16:11:38.

Nanoscale Accepted Manuscript

Figure 2. a) UC(P) of NaYF4:Er3+Yb3+ UCNP dispersed in toluene measured with TH,exp (open squares) and Gauss,exp BPs (open triangles) using an optical path length of 10 mm
and integrating over all UC emission bands between 370 and 890 nm). b) Quotient of UC(P) obtained with Gauss,exp and TH,exp BPs; c) P-dependent integral intensities of the
purple (410 nm; purple square symbols, green (sum of both green emission bands at 520 nm and 540 nm: light green circles), and red emission bands (655 nm; red triangles) of
Er3+ given in percentage of the total UCL for our UCNP measured with TH,exp (full symbols) and Gauss,exp BP (open symbols); d) and e) corresponding slope factors n(PTH,exp) (full
symbols) and n(PGauss,exp) (open symbols) for the different emission bands of Er 3+ (violet: 410 nm, dark cyan: 520 nm, dark green: 540 nm, and red: 655 nm). The P-dependent slope
factors of the 520 nm and 540 nm emission band overlap.

Panels d) and e) of Fig. 2 summarize the n(P) values obtained
with TH,exp and Gauss,exp BPs (calculation of n(P) from
experimental data as detailed in the SI (see Fig. S7 and eq. S4).
These panels highlight the small differences between both
experimentally obtained data sets of maximally 10 % and the
decrease of n(P) of all emission bands with increasing P
indicating saturation. For the unsaturated P-regime (P < 10
-2
Wcm ), n amounts to 2 for the green and the red emission and
P-dependence of luminescence spectra, UC, and slope
factors of nm and µm-sized UC powders.
Subsequently, we studied the UCL of our UCNP in the solid
state and commercial µm-sized UC particles (UCµP) of
comparable degree of doping using the TH,exp BP to assess
size effects without the contribution of potentially quenching
solvent molecules. This enables also a comparison with results
from Boyer and van Veggel and theoretical data of UC(P)
32

to 3 for the purple emission, respectively. This points to similar reported for UCµP by the Berry group. UC(P), IUC, and n
23

population pathways of the green and the red luminescence at values resulting for the different UCL bands are summarized in
-2
low P. At P = 50 Wcm , the red emission shows a maximum n Fig. 3.
value of 2.2 suggesting contributions from bi- and triphotonic
4
processes due to increased population of the G11/2 energy
40
level, which feeds also the triphotonic purple emission .
3+ 3+
Recycling of photons via BET to Yb is unlikely, since Yb
quenching is high particularly for unshelled and relatively small
UCNPs as studied here. Recently, Berry et al. demonstrated for
90 nm-sized particles that the red emission is mostly fed by a
45
triphotonic process even at low P. Our n values suggest,
however, that in the case of smaller UCNP like our 25 nm sized
UCNPs, a biphotonic population process of the red emission is
dominant at low P.

4 | Nanoscale., 2017, 00, 1-3 This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

 Page 5 of 9 Please doNanoscale
not adjust margins

Nanoscale ARTICLE

View Article Online

DOI: 10.1039/C7NR02449E
Published on 13 June 2017. Downloaded by University of California - San Diego on 13/06/2017 16:11:38.

Nanoscale Accepted Manuscript

Figure 3. P-dependence of a) UC of NaYF4:Er3+,Yb3+ UCNP powder and a commercial UCµP powder measured with TH,exp and an optical path length of 0.1 mm; the UC data of
UCµP were corrected for the beam deformation due to heated ND filters and the Yb3+ emission at the excitation wavelength41 (see SI, Fig.S5 and S8); b) Quotient of UC(P) of UCNP
and UCµP; c) P-dependent integral intensities of the purple (410 nm; purple squares) green (sum of both green emission bands at 520 nm and 540 nm: light green circles), and red
emission bands (655 nm; red triangles) given in percentage of total UCL for solid UCNP (open symbols) and UCµP (full symbols), respectively. d) and e) Corresponding slope factors
n(P) derived for solid UCNP (open symbols) and UCµP (full symbols) for the different emission bands of Er 3+ (violet: 410 nm, dark cyan: 520 nm, dark green: 540 nm, and red: 655
nm) with TH,exp BP. n(P) of the 520 nm and 540 nm emission bands overlap for solid UCNP; for UCµP, the corresponding n reveal a different P-dependence for P > 15 Wcm-2.

As follows from Fig. 3 (panel a)), UC(P) of the UCNP powder
-2
increases steadily up to P values of 46 Wcm , where the
material started to degrade, most likely due to decomposition
of the organic surface ligands. UC(P) of the UCµP powder
increases with increasing P and levels off at P of about 30
Wcm . Fitting of UC(P) of the green emission yields
-2
UC,sat,green of 0.6 % with Pbalance of 100 Wcm-2, see SI (Fig. S3
and Table S2; deviation < 5 %). The decrease in UC(P) of the
UCµP powder at higher P exceeding the saturation power is
not to be expected from the balancing power model, yet can
be simulated with actual rate equation models currently
developed. This observation is attributed to the larger
proportion of absorbed photons used for triphotonic UC
processes at high P. Although UC(P) of the UCNP and UCµP
45
powders differ by more than two orders of magnitude at low
P, these deviations strongly decrease with increasing P (see
also Fig. 3, panel b)). This strong diminution of UC(P) with
particle size is reflected by the decay kinetics of the Yb
2 2
emission at 976 nm ( F5/2→ F7/2) and the green emission of Er
4 4
(540 nm, S3/2 → I15/2) and NIR (1520), yielding reductions in
lifetime by factors of 17,6 and 3, respectively (see Fig. 4, Fig. S9
and Table S3). This shows, that the main contribution of the
UC reduction results from Yb3+ quenching revealed by Würth
40, 46 3+
et al. and Arppe et al., while the Er decays is much less
affected. This is ascribed to quenching of UCL by (CH 2)n
stretching vibrational modes of the oleate ligands at 2800 to
-1 47
2950 cm , that overlap with the energy level differences of
several emission bands like the green and red luminescence
-1 -
which are in the order of 3000 cm to 3200 cm
. .Comparison of UC(P) of the UCNP powder and the
1 47
corresponding UCNP dispersion (see SI, Fig. S10) highlights the
reduction in UC(P) by factors of about 4 for low P (P of about
-2 -2
1 Wcm ) and 3 at higher P (P of about 30 Wcm ) for dispersed
UCNP; this reflects additional surface quenching by solvent
3+
molecules. This is supported by the lifetime data of the Yb
2 2
( F5/2→ F7/2) transition (see SI Fig. S9 panel d)), which also
underlines the difficulties of finding suitable rate constants for
the modelling with rate equations, since the beginning of the
3+
decay curve of UCNP powder is strongly multiexponential.
3+

This journal is © The Royal Society of Chemistry 20xx Nanoscale., 2013, 00, 1-3 | 5

Please do not adjust margins

 Please doNanoscale
not adjust margins Page 6 of 9

ARTICLE Nanoscale

View Article Online

DOI: 10.1039/C7NR02449E
Published on 13 June 2017. Downloaded by University of California - San Diego on 13/06/2017 16:11:38.

3+ 3+ 3+ 2 2
Figure 4 a) Energy scheme of the NIR emission of Yb at 976 nm and Er at 1520 nm, respectively;. b) Decay curves of the directly excited Yb emission ( F5/2→ F7/2) at 976 nm
(detected at exc = 940 nm) obtained for UCµP (blue) and UCNP (black) powders and UCNP dispersed in toluene (red); c) Rise and decay of the Er3+ emission at 1520 nm

Nanoscale Accepted Manuscript

( I13/2→ I15/2) for exc = 976 nm measured with a short laser pulse (pulse width < 2 µs), same colour code as used for panel b).
4 4

As follows from Fig. 3 (panel a)) and Table S1, we obtained

UC,max of 10.5 % (integration over all emission bands) for the
-2
UCµP powder at P = 30 Wcm , equaling a value of 9.3 % for
integration of only the visible (vis) luminescence, i.e., the
purple, green, and red emission. For solely the green emission
-2
we measured 2.4 %, (P = 20 Wcm , see Table 1), which is
3+
identical to the value from Page reported for bulk NaYF4: Er ,
3+
Yb (it should be noted, that for direct comparison with our
measurements, the value provided by Page had to be
converted from power conversion efficiency to UC see SI,
Figure S11; the power conversion efficiency takes the photon
energy into account and can thus reach maximum 100 %
whereas UC defined in eq. 2 can reach maximally 50 % for a
purely biphotonic process and even lower values for higher
Figure 5 Comparison of experimentally determined UC of the visible emission (black
order photonic processes). Our UC value agrees also well with
symbols) derived from the sum of the purple (purple symbols), green (green symbols),
the UC value of 3.0 +/- 0.3% obtained by van Veggel for the and red UC (red symbols) with the results from the simulation of these quantities of
3+ 3+-
green emission of pure -phase 20 % Yb , 2% Er doped the Berry group (corresponding lines).
34
NaYF4 UCµP with a size >> 100 nm at identical P. Our overall
UC values agree also well with simulations of UC(P) of Fig. 3 (panel c)) shows a comparison of the relative integral
3+ 3+
commercial -phase NaYF4: 18 % Yb , 2 % Er UCµP of the intensities of the different UC emission bands of our UCµP and
23
Berry group obtained from coupled rate equations (Fig. 5) UCNP powders. This underlines the influence of particle size on
utilizing lifetime data and spectral data measured with pulsed the intensity ratios of the green and red emission (Ig/r), which
excitation (varied from 2 to 200 mJ/cm²). This yields UC,max of present the main contributions to UC(P) with >80 % and
7.6 % for the visible UCL derived from the sum of the purple, >75 %, respectively. For UCµP, Ig/r increases with decreasing P,
green, and red emission. Interestingly, contrary to this whereas for UCNP, which are more prone to nonradiative
generally good agreement, the UC values obtained for the depopulation processes, Ig/r converges to 1.25 for decreasing P
individual UC bands differ strongly from simulations of the (see SI, Fig. S6). This underlines the lower nonradiative rate
Berry group, although the rise and decay kinetics of the purple, from the green to the red band for UCµP. In the saturated
green and red emission, (see SI, Fig. S12) of our UCµP closely regime for both materials the Ig/r converges to 0.33 and 0.5 for
match with the lifetime data reported by the Berry group, UCµP and UCNP, respectively. Ig/r values < 1 for high P have
underpinning the comparability of both materials. been previously reported. In general, for high P, a Ig/r value of
0.3 to 0.4 value is often obtained for NaYF4–based systems
3+
with doping concentrations of about 20 % of Yb and 2 % of
3+
Er , which suggests that this Ig/r value may be specific for UCL
31, 48, 49
saturation. This was also recently found by us for 23 nm-
40
sized NaYF4 UCNPs with these doping concentrations.
Additionally it was demonstrated that for the same UCNP
material in different environments the population dynamics
40
are comparable for high P despite different quenching rates.
In case of the triphotonic purple emission also the population
for the UCµP is more effective as reflected by a higher

6 | Nanoscale., 2017, 00, 1-3 This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

 Page 7 of 9 Please doNanoscale
not adjust margins

Nanoscale ARTICLE

percentage of its contribution to the overall UCL (see Fig. 3, red emission in the unsaturated P regime and n of 2.5
View forOnline
Article the
3+ DOI: 10.1039/C7NR02449E
panel c)); this is ascribed to the longer lifetimes of the Er purple emission, respectively. The latter falls below the n value
intermediate energy levels of the UCµP powder (see also SI, of 3 obtained in UCNP dispersion. In the case of the UCµP
section on luminescence dynamics). powder, n(P) of all emission bands decrease with increasing P
n(P) of the green, red, and purple emission bands of the UCNP as to be expected for saturation. Comparison of the n(P) values
and UCµP powders are highlighted in panels d) and e) of Fig. 3.
For the UCNP powder, we obtained n of two for the green and

Table 1 Overall UC (integration UCL- spectra from 370 nm to 890 nm) and UC,emission band of the purple (410 nm), green (520 nm,540 nm), red (655 nm), IR1 (810 nm) and IR2 (850
nm) emission bands, respectively. TheseUC values are listed for P of 20 Wcm-2 (at this P, the green UC of UCµP has its maximum (see Figure 5)) and for the highest P value
achieved for each sample, respectively. All UC values given were measured with the experimentally realized Top Hat beam profile (see SI, Figure S1).
Published on 13 June 2017. Downloaded by University of California - San Diego on 13/06/2017 16:11:38.

Sample/(P) UC,purple(%) UC,,green(%) UC,,red(%) UC,,IR1(%) UC,,IR2(%) UC,(%)

520,540 nm
410 nm 2 4 655 nm 810 nm 850 nm
2 4 H11/2, S3/2 4 4 4 4 4 4
H9/2→ I15/2 F9/2→ I15/2 I9/2→ I15/2 S3/2→ I15/2

Nanoscale Accepted Manuscript

4
→ I15/2
UCµP powder/
-2 0.17 2.4 6.4 0.13 0.92 10.3
(20 Wcm )
UCµP powder/
-2 a 0.19 1.8 5.5 0.11 0.52 8.3
(130 Wcm )
UCNP powder/
-2 0.0028 0.11 0.14 0.015 0.039 0.32
(20 Wcm )
UCNP powder/
-2 a 0.0079 0.20 0.32 0.023 0.066 0.63
(46 Wcm )
UCNP dispersion/
-2 3.7E-4 0.040 0.033 0.0068 0.015 0.098
(20 Wcm )
UCNP dispersion/
-2 0.037 0.43 0.75 0.040 0.14 1.44
(410 Wcm )
a
powder sample started to degradate for higher P.

derived for all UC emission bands of the UCµP powder in the
saturation regime reveals that solely the 520 nm emission
2 4
( H11/2→ I15/2) converges to 1, whereas n(P) of the other
-2
emission bands reach a value of about 0.5 for P of 100 Wcm .
2
Although the H11/2 state can be thermally populated from the
4 50
S3/2 state according to the Boltzmann population, this
cannot account for the trend of n(P) of the 520 nm emission,
since additionally performed longterm illumination studies
show a minimal influence on UC intensity < 5 % for high P (see
SI, Fig. S13).
Conclusions
In summary, we presented strategies for the quantitative
determination of excitation power density (P)-dependent
upconversion (UC) luminescence (UCL) spectra and intensity
ratios and UC quantum yields UC(P) of nanoparticle
dispersions and solid samples and other spectroscopic
parameters like slope factors (n) relevant for mechanistic
insights in UC processes. Moreover, ligand and solvent
quenching were quantified and the potential of P for tuning UC
efficiency and color underlined. Although the UC value found
by us for UC microparticles differs from the value reported by
32
van Veggel for a similar material of yet unknown dopant
concentration, the good agreement of our UC(P) with UC(P)
45
simulated by the Berry group validates our measurements.
Despite the generally good match for the combined visible
emission, our results suggest an overestimation of the purple
3+
and an underestimation of the red emission of -NaYF4:Yb ,
3+
Er : microparticles by the simulations of the Berry group.
Our P-dependent UCL measurements with different beam
profiles underline that the main factors rendering the
comparability of reported UC values presently critical are the
strong influence of the size, shape, and homogeneity of the
spatial P distribution on UCL and UC together with the
challenge to reliably measure P. This makes also relative
measurements of UC suggested in some publications very
5
error prone. The considerably different UC values of
dispersed and solid UCNP in the low P regime underpin the
importance of measurements of UC(P) for comparison of UC
efficiencies of UC materials. This renders maximal UC (UC,max)
values better suited for straightforward material performance
assessment.
Materials and methods
3+ 3+
Materials. 25 nm-sized -NaYF4:Yb : 17 %, Er : 3% UCNPs
(crystallite size 23x26-30 nm) were prepared through co-
precipitation in a binary solvent mixture of oleic acid and 1-
14
octadecene and characterized by TEM and XRD (see Fig. S14).
Luminescence measurements were performed with these
oleate-stabilized UCNP dispersed in toluene and as dried
3+ 3+
powder. 3 µm-sized -NaYF4: Yb , Er particles doped with ca.

This journal is © The Royal Society of Chemistry 20xx Nanoscale., 2013, 00, 1-3 | 7

Please do not adjust margins

 Please doNanoscale
not adjust margins
ARTICLE
3+ 3+ 42
21.4 % Yb and 2.2 % Er (PTIR 550) were obtained from
Phosphor Technology Ltd. and measured as powder.
Absorption measurements. Absorption spectra of the UCNP
dispersions were recorded on a calibrated Cary 5000
spectrometer. As follows from an absorption spectrum of the
UCNP dispersed in toluene shown in the SI (Fig. S2), these
dispersions are nonscattering at the excitation wavelength.
Absolute measurement of P-dependent UC. P-dependent
UC values were obtained absolutely with a custom designed
51
integrating sphere setup previously reported, modified to
meet the requirements of UC luminescence measurements like
Published on 13 June 2017. Downloaded by University of California - San Diego on 13/06/2017 16:11:38.
the use of a high stability 8 W 976 nm laser diode as excitation
light source. More details on the calibration of the setup,
beam profile characterization, measurement procedure and
evaluation of the data are given in the SI (see Section on
Integrating sphere setup for absolute measurements of UC)
Time-resolved luminescence measurements. Lifetime
measurements of the UC and downconverted (DC)
luminescence (downshifted or Stokes shifted emission upon
direct excitation of the respective emitter) were done with a
Edinburgh Instruments spectrofluorometer FSP-920 equipped
with an electrically pulsed 1 W 978 nm laser diode (950 ms
long square pulse), a pulsed xenon lamp (pulse width ca. 1 µs),
a red extended PMT (R2658P; UC luminescence), and a NIR
PMT (R5509P; DC luminescence) from Hamamatsu (see SI, Fig.
S9 and Table S3). For the comparison of the lifetime data of
23
the berry group and for the control of the reliability of the
resulting lifetimes of the long pulse measurements the
luminescence decays obtained with short pulse (~10 ns)
excitation performed at TU Berlin using a dye (IR140) laser
pumped with an excimer laser. The observed deviations were
< 10 %. More details and the resulting lifetimes are provided in
the SI (Figure S9 and Table S3 and S4). The decay analysis is
detailed in the SI (see Section on Method for the determination
of the lifetime components).
Acknowledgements
We gratefully acknowledge Emilia Palo for the synthesis of the
upconversion nanoparticles, fruitful discussions with Dr.
Markus Grabolle, Prof. Dr. Axel Hoffmann for enabling time-
resolved measurements with short pulse excitation in his lab.
Financial support from the Federal Ministry of Economics and
Technology (MNPQ program, projects BMWI-13/09 and
BMWI-11/12), the German Research Council (DFG; grant
RE1203/18-1), and the European Upconversion Network (COST
Action CM1403) is gratefully acknowledged.
Notes and references
1. F. Auzel, Chemical Reviews, 2004, 104, 139-173.
2. A. Gnach and A. Bednarkiewicz, Nano Today, 2012, 7, 532-
563.
3. X. Wu, G. Y. Chen, J. Shen, Z. J. Li, Y. W. Zhang and G. Han,
Bioconjugate Chemistry, 2015, 26, 166-175.
8 | Nanoscale., 2017, 00, 1-3
Please do not adjust margins
Page 8 of 9
Nanoscale
4. C. T. Xu, Q. Q. Zhan, H. C. Liu, G. Somesfalean, J. Qian, S. L.
View Article Online
He and S. Andersson-Engels, LaserDOI: & Photonics Reviews,
10.1039/C7NR02449E
2013, 7, 663-697.
5. S. L. Gai, C. X. Li, P. P. Yang and J. Lin, Chemical Reviews,
2014, 114, 2343-2389.
6. W. Zheng, P. Huang, D. T. Tu, E. Ma, H. M. Zhu and X. Y.
Chen, Chemical Society Reviews, 2015, 44, 1379-1415.
7. J. C. Goldschmidt and S. Fischer, Advanced Optical
Materials, 2015, 3, 510-535.
8. J. M. Meruga, A. Baride, W. Cross, J. J. Kellar and P. S.
May, Journal of Materials Chemistry C, 2014, 2, 2221-
2227.
9. H. H. Gorris and O. S. Wolfbeis, Angewandte Chemie-
International Edition, 2013, 52, 3584-3600.
10. Y. Q. Lu, J. B. Zhao, R. Zhang, Y. J. Liu, D. M. Liu, E. M.
Goldys, X. S. Yang, P. Xi, A. Sunna, J. Lu, Y. Shi, R. C. Leif, Y.
Nanoscale Accepted Manuscript
J. Huo, J. Shen, J. A. Piper, J. P. Robinson and D. Y. Jin,
Nature Photonics, 2014, 8, 33-37.
11. A. Nadort, V. K. A. Sreenivasan, Z. Song, E. A. Grebenik, A.
V. Nechaev, V. A. Semchishen, V. Y. Panchenko and A. V.
Zvyagin, Plos One, 2013, 8.
12. H. Pakkila, M. Yliharsila, S. Lahtinen, L. Hattara, N.
Salminen, R. Arppe, M. Lastusaari, P. Saviranta and T.
Soukka, Analytical Chemistry, 2012, 84, 8628-8634.
13. T. Soukka, T. Rantanen and K. Kuningas, in Fluorescence
Methods and Applications: Spectroscopy, Imaging, and
Probes, ed. O. S. Wolfbeis, 2008, vol. 1130, pp. 188-200.
14. F. Wang, Y. Han, C. S. Lim, Y. H. Lu, J. Wang, J. Xu, H. Y.
Chen, C. Zhang, M. H. Hong and X. G. Liu, Nature, 2010,
463, 1061-1065.
15. X. Wu, Y. Zhang, K. Takle, O. Bilsel, Z. Li, H. Lee, Z. Zhang,
D. Li, W. Fan, C. Duan, E. M. Chan, C. Lois, Y. Xiang and G.
Han, ACS Nano, 2016, 10, 1060-1066.
16. M. Saboktakin, X. Ye, S. J. Oh, S.-H. Hong, A. T. Fafarman,
U. K. Chettiar, N. Engheta, C. B. Murray and C. R. Kagan,
ACS Nano, 2012, 6, 8758-8766.
17. N. J. Greybush, M. Saboktakin, X. Ye, C. Della
Giovampaola, S. J. Oh, N. E. Berry, N. Engheta, C. B.
Murray and C. R. Kagan, ACS Nano, 2014, 8, 9482-9491.
18. X. Liu, C.-H. Yan and J. A. Capobianco, Chemical Society
Reviews, 2015, 44, 1299-1301.
19. C. Würth, D. Geißler, T. Behnke, M. Kaiser and U. Resch-
Genger, Anal Bioanal Chem, 2014, DOI: 10.1007/s00216-
014-8130-z, 1-20.
20. C. Würth, M. Grabolle, J. Pauli, M. Spieles and U. Resch-
Genger, Nat Protoc, 2013, 8, 1535-1550.
21. M. Haase and H. Schafer, Angewandte Chemie-
International Edition, 2011, 50, 5808-5829.
22. C. Würth, M. Grabolle, J. Pauli, M. Spieles and U. Resch-
Genger, Analytical Chemistry, 2011, 83, 3431–3439.
23. R. B. Anderson, S. J. Smith, P. S. May and M. T. Berry,
Journal of Physical Chemistry Letters, 2014, 5, 36-42.
24. S. Fischer, R. Martin-Rodriguez, B. Frohlich, K. W. Kramer,
A. Meijerink and J. C. Goldschmidt, Journal of
Luminescence, 2014, 153, 281-287.
25. V. K. Tikhomirov, G. Adamo, A. E. Nikolaenko, V. D.
Rodriguez, P. Gredin, M. Mortier, N. I. Zheludev and V. V.
Moshchalkov, Optics Express, 2010, 18, 8836-8846.
26. H. C. Liu, C. T. Xu, D. Lindgren, H. Y. Xie, D. Thomas, C.
Gundlach and S. Andersson-Engels, Nanoscale, 2013, 5,
4770-4775.
This journal is © The Royal Society of Chemistry 20xx
 Page 9 of 9 Please doNanoscale
not adjust margins

Nanoscale ARTICLE

27. M. Pollnau, D. R. Gamelin, S. R. Luthi, H. U. Gudel and M. View Article Online

P. Hehlen, Physical Review B, 2000, 61, 3337-3346. DOI: 10.1039/C7NR02449E
28. J. Shen, G. Chen, T. Y. Ohulchanskyy, S. J. Kesseli, S.
Buchholz, Z. Li, P. N. Prasad and G. Han, Small, 2013, 9,
3213-3217.
29. G. Chen, J. Shen, T. Y. Ohulchanskyy, N. J. Patel, A.
Kutikov, Z. Li, J. Song, R. K. Pandey, H. Ågren, P. N. Prasad
and G. Han, ACS Nano, 2012, 6, 8280-8287.
30. G. Chen, T. Y. Ohulchanskyy, A. Kachynski, H. Ågren and P.
N. Prasad, ACS Nano, 2011, 5, 4981-4986.
31. R. H. Page, K. I. Schaffers, P. A. Waide, J. B. Tassano, S. A.
Payne, W. F. Krupke and W. K. Bischel, Journal of the
Published on 13 June 2017. Downloaded by University of California - San Diego on 13/06/2017 16:11:38.

Optical Society of America B-Optical Physics, 1998, 15,

996-1008.
32. J. C. Boyer and F. van Veggel, Nanoscale, 2010, 2, 1417-
1419.

Nanoscale Accepted Manuscript

33. S. V. K. D.S. Yasyrkina, A.V. Ryabova, D.V. Pominova,V.V.
Voronov, R.P. Ermakov, P.P. Fedorov, NANOSYSTEMS:
PHYSICS, CHEMISTRY, MATHEMATICS, 2013, 4, 648–656.
34. Q. Liu, Y. Sun, T. Yang, W. Feng, C. Li and F. Li, Journal of
the American Chemical Society, 2011, 133, 17122-17125.
35. C. T. Xu, P. Svenmarker, H. Liu, X. Wu, M. E. Messing, L. R.
Wallenberg and S. Andersson-Engels, ACS Nano, 2012, 6,
4788-4795.
36. X. Li, R. Wang, F. Zhang and D. Zhao, Nano Letters, 2014,
14, 3634-3639.
37. M. Pokhrel, A. k. Gangadharan and D. K. Sardar, Materials
Letters, 2013, 99, 86-89.
38. M. Pokhrel, G. A. Kumar and D. K. Sardar, Journal of
Materials Chemistry A, 2013, 1, 11595-11606.
39. W. Wei, Y. Zhang, R. Chen, J. Goggi, N. Ren, L. Huang, K. K.
Bhakoo, H. D. Sun and T. T. Y. Tan, Chemistry of Materials,
2014, 26, 5183-5186.
40. C. Würth, M. Kaiser, S. Wilhelm, B. Grauel, T. Hirsch and
U. Resch-Genger, Nanoscale, 2017, DOI:
10.1039/C7NR00092H.
41. S. Fischer, N. D. Bronstein, J. K. Swabeck, E. M. Chan and
A. P. Alivisatos, Nano Letters, 2016, 16, 7241-7247.
42. T. Soukka, K. Kuningas, T. Rantanen, V. Haaslahti and T.
Lovgren, Journal of Fluorescence, 2005, 15, 513-528.
43. Y. Ebenstein, T. Mokari and U. Banin, Applied Physics
Letters, 2002, 80, 4033-4035.
44. N. Durisic, A. G. Godin, D. Walters, P. Grütter, P. W.
Wiseman and C. D. Heyes, ACS Nano, 2011, 5, 9062-9073.
45. M. T. Berry and P. S. May, The Journal of Physical
Chemistry A, 2015, 119, 9805-9811.
46. R. Arppe, I. Hyppanen, N. Perala, R. Peltomaa, M. Kaiser,
C. Wurth, S. Christ, U. Resch-Genger, M. Schaferling and T.
Soukka, Nanoscale, 2015, 7, 11746-11757.
47. S. Wu, Y. Ning, J. Chang and S. Zhang, Journal of
Luminescence, 2013, 143, 492-497.
48. C. F. Gainer, G. S. Joshua, C. R. De Silva and M.
Romanowski, Journal of Materials Chemistry, 2011, 21,
18530-18533.
49. T. Jung, H. L. Jo, S. H. Nam, B. Yoo, Y. Cho, J. Kim, H. M.
Kim, T. Hyeon, Y. D. Suh, H. Lee and K. T. Lee, Physical
Chemistry Chemical Physics, 2015, 17, 13201-13205.
50. L. Marciniak, K. Waszniewska, A. Bednarkiewicz, D.
Hreniak and W. Strek, The Journal of Physical Chemistry C,
2016, 120, 8877-8882.
51. C. Würth, J. Pauli, C. Lochmann, M. Spieles and U. Resch-
Genger, Analytical Chemistry, 2012, 84, 1345-1352.

This journal is © The Royal Society of Chemistry 20xx Nanoscale., 2013, 00, 1-3 | 9

Please do not adjust margins