Photodermatology, Photoimmunology & Photomedicine

ORIGINAL ARTICLE

New noninvasive approach assessing in vivo sun

protection factor (SPF) using diffuse reflectance
spectroscopy (DRS) and in vitro transmission
Eduardo Ruvolo Junior, Nikiforos Kollias & Curtis Cole

Johnson&Johnson Consumer and

Personal Products Wordwide,
Skillman, NJ, USA.

Key words:
diffuse reflectance spectroscopy
(DRS); in vitro; in vivo;
spectroscopy; sun protection factor
(SPF); sunscreen; ultraviolet A
protection factor (UVA-PF)

Correspondence:
Mr Eduardo Ruvolo Junior, M.S.,
Johnson&Johnson Consumer and
Personal Products WW, 199
Grandview Rd, Skillman, NJ 08558,
USA.
Tel: +1 908 874 2377
Fax: +1 908 874 1209
e-mail: eruvolojr@gmail.com

Accepted for publication:

20 December 2013

Conflicts of interest:
None declared.

SUMMARY
Background/Purpose
In the past 56 years, many different in vitro methodologies have been developed and published to assess the sun protection factor (SPF) of products, but
there is no method that has 1 : 1 correlation with in vivo measurements.
Spectroscopic techniques have been used to noninvasively assess the UVA
protection factor with good correlation to in vivo UVA-PF methodologies.
To assess the SPF of sunscreen product by diffuse reflectance spectroscopy
(DRS) technique, it is necessary to also determine the absorbance spectrum of
the test material in the UVB portion of the spectrum (290320 nm). However,
because of the high absorbance characteristics of the stratum corneum and
epidermis, the human skin does not remit enough UVB radiation to be used to
measure the absorption spectrum of the applied product on skin. In this work,
we present a new method combining the evaluation of the absolute UVA
absorption spectrum, as measured by DRS with the spectral absorbance
shape of the UVB absorbance of the test material as determined with current
in vitro thin film spectroscopy.
Methods
The measurement of the in vivo UVA absorption spectrum involves the assessment of the remitted intensity of monochromatic UVA radiation (320
400 nm) before and after a sunscreen product was applied on skin using a
spectrofluorimeter Fluorolog 3, FL3-22 (Yvon Horiba, Edison, NJ, USA). The
probe geometry assures that light scattering products as well as colored products may be correctly assessed. This methodology has been extensively tested,
validated, and reported in the literature.
The in vitro absorption spectrum of the sunscreen samples and polyvinyl
chloride (PVC) films surrogate sunscreen standards were measured using
Labsphere UV-2000S (Labsphere, North Sutton, NH, USA). Sunscreens
samples were tested using PMMA Helioplates (Helioscience, Marseille,
France) as substrates. The UVB absorbance spectrum (Labsphere) is attached
to the UVA absorbance spectrum (diffuse reflectance) with the UVB absorbance matched to the UVA absorbance at 340 nm to complete the full spectral
absorbance from which an estimate the SPF of the product can be calculated.

2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd
doi:10.1111/phpp.12105

PHPP_12105

Ruvolo Junior et al.

Results
Seventeen test materials with known in vivo SPF values were tested. Two of the
tested products were PVC sunscreen thin films with 1015 micrometers thickness and were used to investigate the absorption spectrum of these films when
applied on different reflectance surfaces. Similar to the human in vivo SPF test,
the developed methodology suggests limiting the use on Fitzpatrick skin
phototypes I to III. The correlation of this new method with in vivo clinical
SPF values was 0.98 (r2) with a slope of 1.007.
Conclusion
This new methodology provides a new approach to determine SPF values
without the extensive UV irradiation procedures (and biological responses)
currently used to establish sunscreen efficacy. Further work will be conducted
to establish methods for evaluation of products that are not photostable.
Photodermatol Photoimmunol Photomed 2014; :

The evaluation of the efficacy of sun care products have for

a long time been assessed through the in vivo sun protection factor (SPF) test, which is performed on human volunteers. The concept was introduced by Franz Greiter in
1962 (1), and has become a standard for measuring the
effectiveness of sunscreen when applied on subjects at
application dose of 2 mg/cm2, followed by irradiation of
the sunscreen treated sites and a nonprotected site with
increasing doses of simulated UV radiation and comparing
the doses required to cause a minimal erythema (sunburn)
response. The ratio of the doses required to cause the
sunburn (protected divided by unprotected) represents the
protective factor for the sunscreen, known as the SPF. This
requires extensive exposures to the test subjects skin and
induction of sunburn reactions at multiple exposure sites
in order to determine the SPF value. In the past 56 years,
many different in vitro methodologies have been developed and published to assess the SPF of products (217)
with use of many different instruments, substrates, and
product application optical density in an effort to develop
an alternative to UV exposures to human subjects. To date,
there is no method that has good correlation with in vivo
SPF measurements or that can demonstrate reliable and
repeatable SPF results.
As the SPF test was originated, the methodology has
been substantially unchanged, with exception of minor
changes to equipment used and clarification of the definitions of the minimal erythema dose (MED) end point. The
in vivo human SPF method has been codified and adapted
in most countries as the standard for sunscreen perfor-

mance (1823). The most significant differences between

the international methods are now primarily in UV exposure increments.
Diffuse reflectance spectroscopic (DRS) techniques have
been used to noninvasively assess the UVA protection
factor with good correlation to in vivo UVA-PF methodologies for UVA-PF from 2 to 12 (24). This involves the
assessment of the remitted intensity of monochromatic
UVA radiation (320400 nm) before and after a sunscreen
product was applied on skin. Incident monochromatic
UVA radiation is delivered to the skin through a bifurcated
optical fiber bundle while a second fiber optic bundle (randomly intermixed with the incident bundle) collects the
remitted intensity from the skin. The UVA radiation is
attenuated both on its way through the sunscreen layer on
the surface of the skin and again as reflected and scattered
within the skin on its way out of the skin through the
applied product on the skins surface. The absorbance is
scanned across the UVA spectrum and thus the absolute
UVA absorption spectrum of the tested product can be
determined noninvasively. The probe geometry assures
that light scattering pigment containing products may be
correctly assessed. This in vivo/DRS methodology has been
tested, validated, and reported in the literature (2426).
However, because of the high absorbance characteristics of
the stratum corneum and epidermis, the human skin does
not remit enough UVB radiation to be useful to measure
the absorption spectrum of the applied product in the
UVB region. It is therefore necessary to determine the
absorbance spectrum of the test material in the UVB

Photodermatol Photoimmunol Photomed 2014; :

2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

PHPP_12105

Noninvasive approach assessing in vivo SPF

portion of the spectrum (290320 nm) separately using

another technique. The approach utilized in this method
uses in vivo evaluation of the absolute UVA absorption
spectrum, as measured by DRS with a grafting of the spectral absorbance shape of the UVB absorbance of the test
material as determined with current in vitro thin film spectroscopy to complete the spectral scan in order to project
the in vivo SPF value. This new methodology provides a
new approach to determine SPF values without the extensive UV irradiation procedures (and biological responses)
currently used to establish sunscreen efficacy. This report
provides the development steps and resulting data validating this technique.

MATERIAL AND METHODS

Subjects
Prior to the start of the study, the investigator prescreened
subjects by selecting their Fitzpatrick skin type (I, II, and
III) (27). Selected subjects were in general good health and
had signed informed consent forms after the nature of the
study had been fully explained. Ten subjects were selected
according with their skin phototype for the DRS study.

In vivo UVA absorption spectrum

The UVA absorption spectrum was measured by the in
vivo method using DRS (20). Briefly, the DRS measurements were performed using a spectrofluorimeter
Fluorolog 3, FL3-22 (Yvon Horiba, Edison, NJ, USA). The
Fluorolog-322 is a spectrofluorometer consisting of a 450
W xenon arc lamp, a double excitation monochromator, a
sample compartment with a cuvette holder, a double emission monochromator, and a photomultiplier tube (PMT).
The use of a double-double monochromator (excitation
and emission) provides the highest stray-light rejection
and improves measurement sensitivity. A fiber-optic
probe can be coupled to the sample compartment such
that spectroscopic measurements can be made remotely.
Although the spectrofluorometer has been designed for
fluorescence spectroscopy, fluorescence (as either an excitation or emission spectrum) as well as reflectance spectra
(in synchronous scan mode) can be measured.
For the measurements performed in this paper, the synchronous scan mode was used, and the instrument was set
up with a bifurcated randomized quartz optical probe.
The optical fiber probe was manufactured by FiberGuide
industries (Stirling, NJ, USA), type Superguide, UVVisible, HOH, NA 0.22. The fiber probe is a bifurcated
random bundle with approximately 300 individual fibers
of 200 micrometers diameter 1/2 of the fibers to carry the
Photodermatol Photoimmunol Photomed 2014; :
2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

PHPP_12105

light from the monochromater to the skin, and 1/2 the

fibers to carry the remitted light signal to a photomultiplier
for quantification. There is no spacer media between the
individual fibers.
Diffuse reflectance spectra of untreated skin test sites
(three locations within each test site) were measured by
synchronously scanning the monochromators from
290 nm to 400 nm (in incremental steps of 2.0 nm, integration time = 0.1 s) to determine the baseline remittance
spectra. The high voltage in the instrument photomultiplier (PMT) Hamamatsu R928P (Hamamatsu, Japan) was
set at 800 volts. For each measurement, the investigator
marked sites on the volar forearm of each volunteer with
plastic template 9 cm2 (3 cm 3 cm). There was no
overlap between sites. After the baseline measurements
were completed, products were applied at an application
density of 2 mg/cm2 on the test sites and allowed to dry for
15 min before a second set of three measurements per test
site. Each formulation was tested on 10 subjects. YSI
Teflon membranes model 5793 (YSI, Yellow Spring, OH,
USA) were placed on each site between the skin and the tip
of the fiber optic probe to avoid contamination.
During the in vivo measurement, the light from the
instrument traverses through the sunscreen as it enters the
skin (photons that are not absorbed), is scattered by
dermal collagen, is backscattered, exits the skin, and traverses the sunscreen again before it is collected by the fiber
bundle. Therefore, UVA radiation is attenuated twice by
the sunscreen material before being collected. Therefore,
the transmission spectrum of the sunscreen on the test sites
must be calculated as follows:
T ( ) =

I ( )
Io ( )

(1)

where T() is the transmission of the measured product,

I() is the measured intensity with the product, and Io() is
the measured intensity without the product at wavelength
. Taking the square root in the ratio of remitted intensity
by the incident intensity is necessary due to the double pass
of light through the sunscreen.
The linearity of this method/instrument can be tested by
using UV-VIS neutral density (ND) filters placed on the
skin and measuring the skin reflectance without ND filter
(quartz blank for baseline) and with the ND filter on skin
using the same measurement conditions above described.
UV-VIS ND filters can be purchased with a standard thickness of 1.5 mm, a quartz slide with the same thickness
should be used as blank reading. As NDs are designed to
reduce transmission evenly across a portion of a specific
spectrum, the optical density of the filters can be calculated
3

Ruvolo Junior et al.

by the negative logarithm of the effective transmission

from the equation (1) above for all the wavelengths. Using
ND filters with optical density of 0.1, 0.5, 1.0, 1.5, and 2.0,
the instrument presented a linear response assessing the
correct optical density (as cited above) with an optical
tolerance of 7% over the 320400 nm wavelength interval. Data were not included in this paper.

In vitro UV absorption spectrum

The UV absorbance spectrum of each sunscreen formulation listed in Table 1 was measured in accordance with the
COLIPA guidelines for the In Vitro Determination of the
UVA Protection Factor (28), but no irradiation step was
performed. Briefly, the substrates used were PMMA
Helioplates with surface topography (mechanical microbeam technique) having a specific roughness Ra = 2
micrometers (Helioscience, Marseille, France). The baseline transmission of the PMMA plate, with a homogeneous
layer of glycerin was used as baseline and measured with
a LabsphereTM UV-2000S UV Transmission Analyzer
(Labsphere, North Sutton, NH, USA). A test sunscreen was
then applied to the substrates at a dose of 0.75 mg/cm2 and
spread uniformly over the surface of the plate. The samples
were allowed to air dry for 10 min, and the UV absorbance
of the formulae was measured. The average of 10 UV

absorption spectra (290400 nm) was then used to be

grafted onto the in vivo UVA absorption spectrum that was
measured by DRS.

Tested samples
DRS measurements and in vitro UV absorbance were performed on the 15 sunscreen formulae listed in Table 1 with
their respective in vivo SPF mean values and standard
error. All the in vivo SPF results here presented were determined in an external laboratory certified by REBLAS (Rede
Brasileira de Laboratrios Analticos em Sade) Governmental Certification according with the International
Sun Protection Factor (SPF) Test Method 2006 (22) on 10
subjects.

Grafting the in vitro UVB absorption spectrum

to the in vivo UVA absorbance spectrum
To estimate the SPF of a sunscreen product, knowledge of
the full UV absorption spectrum (290400 nm) is needed.
DRS has been shown to accurately assess the UVA absorption spectrum of the sunscreen product in skin which can
use it to set the absolute scale of the measurement for the
test product and then add the shape of the in vitro UVB
spectrum adjusted to match the absorbance of the UVA

Table 1.
SPF in vivo
Sample

BM

ET

BMT

TD

MBT

ES

OX

DHB

Mean SD

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
PVC Sunscreen Film I
PVC Sunscreen Film II

1.5
0
1.5
3
4
3
3
0
3
0.5
4
1
3
3
3
0
0

2.2
0.9
2
3
3.5
3
3
1.5
3
1
3.8
1.5
3
4
2.5
0
0

2
1
2
2.5
4
4
2.5
1.5
3
1
4.5
1.5
5
2.35
3
0
0

1
0.44
0.7
2.5
3
3
2.5
0.5
2.2
1
3.5
0.7
2.7
4
1.5
0
0

1
0
0.5
0
1
1.5
1
0
1
0
1.5
0.5
2
1
0.5
0
0

0
0
0
0
0
0
0
0
0
0
0
0
0
5
0
0
0

0
0
0
0
0
0
0
0
0
0
0
0
0
4.7
0
0.163
0.313

0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.163
0.313

0
0.4
0
0
0
0
0
0.9
0
0
0
0
0
0
0
0
0

15.4 1.3
7.1 0.6
22.2 1.1
30.8 1.9
51.9 2.4
50.2 2.5
17.4 0.9
10.3 1
31.5 1.7
8.8 0.7
65.2 3.4
16.1 0.9
60.1 3.6
75.5 5.4
30.3 1.4
21.5 3
58.7 5.8

BM, butyl methoxydibenzoylmethane; ET, ethylhexyl triazone; O, octocrylene; BMT, bis-ethylhexyloxyphenol

methoxyphenyl triazine; TD, titanium dioxide; MBT, methylene bis-benzotriazolyl tetramethylbutylphenol; ES,
ethylhexyl salicylate; OX, oxybenzone; DHB, diethylamino hydroxybenzoyl hexyl-B.
4

PHPP_12105

Photodermatol Photoimmunol Photomed 2014; :

2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

Noninvasive approach assessing in vivo SPF

absorbance spectrum in the 340350 nm range to approximate a full spectral scan. The average ratio of DRS absorbance to in vitro absorbance in the range from 340 to
350 nm is used to multiply the in vitro absorbance from
290 to 340 nm to scale it to the absolute absorbance from
the in vivo UVA spectral scan. Fundamentally, the scaled in
vitro UVB/UVA absorbance spectrum from 290340 nm is
attached to the UVA absorbance part of the spectrum
340400 nm from the diffuse reflectance measurement,
creating a full 290400 nm absorbance scan from which we
can be used to estimate the SPF of the product. See Fig. 1
for a visual illustration of the process.
This method assumes that the absorption spectrum
from the DRS measurement has the correct shape and
amplitude; it should not be changed. There are occasions
that the in vitro absorbance is higher than DRS absorbance
depending on the in vitro method used to assess the UV
absorption spectrum of the sample. In case the in vitro
absorbance spectrum has higher values than the in vivo
DRS (for wavelengths lower than 340 nm), the multiplication ratio factor in the 340350 nm range is less than 1.

Standard sunscreen PVC thin film

In order to evaluate reproducibility measurements of the
measurement technique without variability introduced
by sunscreen application inconsistencies, two PVC film
surrogate sunscreen standards were developed having the
similar optical absorbance properties as typical sunscreen
products (both in spectral shape and absorption magnitude) as well as approximately the same physical thickness
as sunscreen formulations on the skin. Ten micrometers is
an approximate average film thickness on skin of residual
product, which varies depending on the product and
application.
A 1% stock solution of the PVC/tetrahydrofuran (THF)
(both from Sigma Aldrich, St. Louis, MO, USA) was prepared by dissolving the PVC in the THF overnight. The
4 mg of Parsol MCX (DSM Nutritional Products AG,
Kaiseraugst, Switzerland) and 2.1 mg of Oxybenzone
(Symrise Inc, Teterboro, NJ USA) were added to yield a
total concentration of 0.325% of combined sunscreens.
The solution was stirred for 10 min. An aliquot of the
prepared solution was placed at the bottom of a crystallizing dish sufficient to cover the bottom of dish wetting the
entire surface. The dried film was removed by prying
around the edges of the dish and pulled from the bottom of
the dish. The cast film (sunscreen film I) had a thickness of
approximately 1015 micrometers. A second film of equal
thickness with 0.625% of sunscreen concentration using
the same proportion of the two filters was prepared
Photodermatol Photoimmunol Photomed 2014; :
2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

PHPP_12105

(sunscreen film II). The UV absorbance spectrum of each

sunscreen film was measured using the Labsphere
UV-2000S UV Transmission Analyzer (Labsphere). The in
vivo SPF of the two PCV sunscreen-impregnated films was
determined by the method of the FDA Final Rule (18). The
skin of 10 volunteer subjects was irradiated with and
without the film present to determine the SPF value for
each film on each subject. The individual and mean SPF
( standard error) value for each film is illustrated in
Table 1.

Reflectance standards
In order to evaluate the influence of skin type (and skin
darkness) on the measured absorbance of sunscreen
products and the sunscreen film standards, DRS measurements were conducted on the two standard SPF films
described above using Spectralon Diffuse Reflectance
Standards (Labsphere) as background substrates for the
measurements.
These reflectance standards are durable, chemically inert
substrates with reflectance values ranging from 2% to 99%
and are spectrally flat over the UV-visible-near-infrared
spectrum ( 4% over the range of 2502500 nm and 1%
over the photo-optic region of the spectrum).
In this experiment standard with 80%, 60%, 40%, 20%,
10%, 5%, and 2% reflectance were tested. Using the same
procedure as described in the in vivo UVA absorption spectrum of the standard sunscreen films were measured using
each of the reflectance standards as the background substrate. Three repetitions were performed before and after
having the standard sunscreen film on top of the surface of
the reflectance standard. Figure 2 illustrates the reflectance
standards used in this experiment and their respective
L* SD color space values (CIE 1976) (29). The standards
have no color, i.e, a* and b* values are close to zero.

RESULTS
Prediction of SPF in vivo
The hybrid in vivo DRS, in vitro UVB absorbance methodology was applied to all the samples illustrated in Table 1,
and comparisons were made between the SPF results
obtained by this new method and the standard (International Harmonized SPF method) in vivo SPF results. The
results and correlation between the two methods are illustrated in Fig. 3. A linear regression through zero using the
error bars as weight was calculated obtaining a slope of
1.007 and a correlation coefficient r2 = 0.98. The linear fit
5

Ruvolo Junior et al.

A)

1.0

0.9

0.9

0.8

0.8

0.7

0.7

Absorbance (a.u.)

Absorbance (a.u.)

1.0

0.6
0.5
0.4
0.3

0.6
0.5
0.4
0.3

0.2

0.2

0.1

0.1

0.0
290

300

310

320

330

340

350

360

370

380

390

B)

0.0
290

400

300

310

320

Wavelenght (nm)
1.0

C)

1.0

350

360

370

380

390

400

D)

0.9

average abs=0.72

0.8

0.8
0.7

Absorbance (a.u.)

0.7

Absorbance (a.u.)

340

Wavelenght (nm)

0.9

0.6
0.5
0.4

average abs=0.38

0.3

0.6
0.5
0.4
0.3

0.2

0.2

0.1

0.1

0.0
290

330

300

310

320

330

340

350

360

370

380

390

400

Wavelenght (nm)

0.0
290

300

310

320

330

340

350

360

370

380

390

400

Wavelenght (nm)

Fig. 1. Illustrates the hybrid method to obtain the full UV (290400 nm) absorption spectrum of the sunscreen product on skin.
Graph at Fig. 1a shows the absorption spectrum of a sunscreen, sample 2 from Table 1, obtained by the in vitro method on
PMMA plate at 0.75 mg/cm2 using the LabsphereTM UV-2000 UV spectrometer. Fig. 1b illustrates the averaged absorption spectrum
from the same sample (from one subject) measured in skin by the DRS method. The average absorbance from 340 to 350 nm is
calculate for both absorption spectra in Fig. 1a and Fig. 1b. The average absorbance for the in vitro spectrum is about 0.38 a.u.
and for the DRS spectrum the value is 0.72 a.u., as illustrated in Fig. 1c). Assuming that the absorption spectrum from the DRS
measurement has the correct shape and amplitude, the absorbance at 340350 nm should not be changed. So we correct the in
vitro spectrum by calculating an average ratio in the absorbance range 340350 nm (absorbance DRS/absorbance in vitro = 1.89).
This scalar ratio is used to multiply the in vitro absorbance for each wavelength from 290 to 340 nm to scale it to the absolute
absorbance from the in vivo UVA spectral scan. Finally, the new corrected in vitro absorption spectrum from 290 to 340 is
attached to the in vivo DRS spectrum, maintaining the absorbance values from the DRS spectrum from 340 to 400 nm, Fig. 1d.
This final spectrum is used to calculate the SPF of the sample. In this example, the estimated SPF by the new method is 7.0 1.3
and the in vivo SPF is equal to 7.1 1.9.

slope and coefficient of determination R2 were obtained

using Origin 6.0 (Northampton, MA, USA).

Influence of skin phototypes on SPF prediction

The cast PVC sunscreen thin films (I and II) in combination of the Spectralon reflectance standards were used
6

PHPP_12105

as a model to understand the influence of the skin substrate darkness on the accuracy of the new DRS/in vitro
UVB method. The same procedure for the in vivo DRS/in
vitro UVB measurement was performed replacing the
human skin by the reflectance standards and the sunscreen product by the thin PVC film sunscreen standard
Photodermatol Photoimmunol Photomed 2014; :
2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

Noninvasive approach assessing in vivo SPF

light intensity in the excitation as well as detecting

the reflected signal. The advantage of using double
monochromators is the substantial reduction of the
stray-light.
At values above 40% reflectance, representing a typical
phototype III individual (or lighter), the predicted SPF
values stabilize at the known SPF value of the test films
indicating that valid results can be obtained within this
range of substrate reflectivity.

90
80

in vivo/vitro DRS SPF

films. The absorption spectra of the sunscreen films

measured using the Labsphere UV-2000S are shown in
Fig. 4.
The absorption spectra of the sunscreen standard I
(lower SPF) measured on different reflectance standards
as skin replacement substrate are illustrated in Fig. 5.
Skin Fitzpatrick phototypes IIII have a UVA reflectance
equivalent to the 40% reflectance standard and higher.
Graph on Fig. 5 illustrates the thin film I absorption
spectrum on a subject skin type III as well the absorption
spectrum of the same film with Labsphere.
Using the hybrid SPF method, the predicted SPF of the
each thin film was calculated for each reflectance standard as shown in Fig. 6, as well as the predicted value for
each film using skin as substrate. These data show that
low reflectance surfaces predict SPF values significantly
below the known SPF value of the test films, indicating
an instrumental limitation in these ranges. This is due to
the fact the instrument has a double monochromator on
the excitation and emission side limiting the amount

70
60
50
40

Y = B * xscale(X)
B=1.007
R= 0.99
R2=0.98

30
20
10
0
0

10

20

30

40

50

60

70

80

90

in vivo SPF
Fig. 3. This figure shows the in vivo SPF values correlation with
the in vivo-DRS/in vitro-UVB SPF values obtained from the
samples illustrated on Table 1. The straight line represents the
best fit of the data. The minimum value for SPF = 1 that
represents the minimum protection factor that is afforded by
the product. Data are given as mean + SE (n = 10).

3.0

Sunscreen Film I
Sunscreen Film II

2.8
2.6
2.4

Absorbance (a.u.)

2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6

Fig. 2. Spectralon Diffuse Reflectance Standards and their

respective reflectance with the following L* (CIE 1976)
color space values: R2% = 11.9 0.01, R5% = 29.8 0.006,
R10% = 37.07 0.006, R20% = 53.75 0.015,
R40% = 66.54 0.025, R60% = 82.59 0.023,
R80% = 91.83 0.023.
Photodermatol Photoimmunol Photomed 2014; :
2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

PHPP_12105

0.4
0.2
0.0
290

300

310

320

330

340

350

360

370

380

390

400

Wavelength (nm)
Fig. 4. Absorption spectra of PVC thin films.
7

Ruvolo Junior et al.

2.0
80% Reflectance
60% Reflectance
40% Reflectance
20% Reflectance
5% Reflectance
Labsphere
in vivo Skin Phototype III

1.8
1.6

Absorbance (a.u.)

1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
290

300

310

320

330

340

350

360

370

380

390

400

Wavelength (nm)

Predicted SPF via DRS

Fig. 5. Illustrates the absorption spectra of the PVC Sunscreen

Film I measured over different Spectralon reflectance
standards as well the absorption spectrum of the same film
measured using the Labsphere UV Transmitter 2000 and on a
subjects volar forearm skin Phototype III.

90
85
80
75
70
65
60
55
50
45
40
35
30
25
20
15
10
5
0

Lasphere SPF=25 (SE=0.4)

Lasphere SPF=99 (SE=6.3)
in vivo PVC Film SPF= 21.5 + 1
in vivo PVC Film SPF= 58.7 + 1.8
Skin DRS SPF=60.9 + 2

Skin DRS SPF=20.1 + 0.4

20

40

60

80

Standard Reflectance
Fig. 6. Shows the predicted SPF using the hybrid model on
difference reflectance standards. Graph also illustrates the
in vitro SPF (film I = 25.3 0.4 and film II = 98.6 6.3;
average SE) for each of the tested films as well the average
predicted SPF on 10 subjects for the hybrid method (film
I = 20.14 0.4 and film II = 60.9 2; average SE)

DISCUSSION
This work presents a new methodology to assess the sun
protection factor of sun care products on human skin
allowing realistic interaction characteristics of the test
product with the skin surface chemistry and topography
8

PHPP_12105

but without having to irradiate the human subject with

damaging doses of ultraviolet radiation. It is based on the
hypothesis that the sunscreen UVA (320400 nm) absorption spectrum, as measured by DRS, represents the absolute absorption characteristics of the product on the skin
and has the same spectral shape as the in vitro measurements and that the product is photostable. The SPF of the
product is obtained by estimating of the absorption spectrum in the UVB (290320 nm) region by using an in vitro
measurement and adjusting it to the known absolute UVA
absorbance range in a valid overlapping range of both
spectra.
The offset in the in vitro absorbance (around 400 nm) as
seen in Fig. 1a and 1c for the Labsphere absorbance measurement appears as a baseline offset that is wavelength
independent. As such, it does not enter into the assessment
of the product SPF estimate as the DRS measurement
determines the absolute height function of the full curve,
including the baseline where no offset is observed.
One important question for this methodology whether
or not the skin pigmentation of the test subject affects the
assessment of the sample SPF. The cast PVC thin film
was a useful tool for this investigation as no product
application was involved (removing application variability) and the prepared films had approximately the same
thickness (1015 m) of a sunscreen product when dried
down on human skin. Based on the measurements of
the standard PCV thin films sunscreen on different
Spectralon reflectance standards, Fig. 4, it is seen that the
absorption of the sunscreen film when assessed in dark
substrates presents a decrease in the shorter wavelengths.
This decrease in absorption values for different reflectance standards is due to the hardware limitations (optics
and electronics) due to the high absorbance of the darker
substrate and insufficiency of adequate return signal. For
reflectance standards higher than 3040% the absorption
spectrum may be accurately assessed. This reflectance
range corresponds to an individual typology angle of
approximately 28 or an L* Chroma Meter (Minolta,
Osaka, Japan) value of approximately 56. Similar to the
human in vivo SPF test subject requirements, this suggests limiting the use of this methodology on Fitzpatrick
skin phototypes IIII. As the Fitzpatrick skin type is
based on sun burning and tanning history, perhaps the
inclusion criteria by skin reflectance is more useful than
skin type. This can be observed in Fig. 6 for the estimated
SPF on the thin films. For the lower SPF sunscreen PVC
film, the estimated valid range for SPF 23 is with a reflectance higher than 40% and for the high SPF standard the
valid range is reached at higher reflectance values
R > 60%.
Photodermatol Photoimmunol Photomed 2014; :
2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

Noninvasive approach assessing in vivo SPF

100
90

Hybrid Model

in vitro SPF

80
70

Predicted SPF

When limiting the tested skin phototypes to I and

III, (see Fig. 3), we can obtain a 1 : 1 correlation between
the in vivo SPF and the hybrid DRS method for sunscreenformulae from SPF = 7 up to 80 with a correlation
coefficient r2 = 0.98. The test samples used in this correlation are all photostable. Based on observations from a
previous paper (20), we know that products that are
non-photostable will overestimate the protection provided when measured via DRS. For such products to be
tested accurately, substantial photoexposure of the test
substrate will be required before obtaining the DRS
measurement and in vitro UVB absorbance spectra and
an algorithm for estimating average/cumulative radiation
damage is required to be developed. Products can be tested
in vitro prior to utilizing this new method to establish
photostability and then become eligible for accurate SPF
prediction with this new test method.
As this technique has a linear response from 0.1 up to
2.0 optical density in the UVA range as described in the
material and methods, the methodology can be used to
evaluate sunscreens with PFA higher than 12. Would be
interesting for future validation test to correlate the UVA
protection factor assessed by DRS with the validated ISO
24443:2012 in vitro method (30).
Another important question is what is the correlation
with the in vivo SPF, for these samples, when we use the
full in vitro UV absorption spectrum to calculate the SPF
according with Sayre et al.?(10) Using the in vitro absorption spectrum and calculating the in vitro SPF, we obtain a
correlation coefficient r2 = 0.54 and a slope of 0.73 with
the in vivo SPF values, Fig. 7. Notice that the used in vitro
method tends to underestimate the SPF values (in the
data presented in this study, except for the PVC film II,
SPF = 99) when 0.75 mg/cm2 of product is used to determine the absorption spectrum. The 1 : 1 correlation
between the predicted and measured SPF values is lost,
and the variability is significantly increased.

slope=1.007
r2=0.98

60
50
40
30
20

slope=0.73
r2=0.54

10
0
0

10

20

30

40

50

60

70

80

90

100

in vivo SPF
Fig. 7. Shows the in vivo SPF correlation with the hybrid
model and the in vitro SPF calculation using the full in vitro
absorbance spectrum obtained in PMMA plates.

CONCLUSION
In conclusion, this paper presents and new noninvasive,
nonirradiative method to predict the sun protection factor
of sunscreen products. For photostable products, the
hybrid DRS/in vitro UVB absorbance method provides an
excellent opportunity to eliminate unnecessary UV exposure to human subjects while providing accurate estimates
of SPF protection values.

ACKNOWLEDGEMENTS
This study was carried out at and founded by
Johnson&Johnson Consumer Products Worldwide. The
authors would like to thank Sergio Oliveira and Leonardo
de Paulo, both from Johnson&Johnson Brazil, for supporting with the in vivo SPF data and sunscreen samples
used in this study.

REFERENCES
1. Greiter F. Sonnenschutzmittel-Typen und
Andwendung. Sonderdruck aus Parfumerie und Kosmetik 1974; 55: 199
202.
2. Kreps SI. Spectrophotometric evaluation
of Sunscreens. Proc Sci Sect Toilet Goods
Assoc 1955; 23: 1326.
3. Master KJ, Sayre RM, Everett MA. New
evaluation for sunscreens. J Soc Cosmet
Chem 1966; 17: 581594.
4. Kahn G, Wilcox GJ. Comparison of in
vitro and in vivo sunscreen testing

method. Soc Cosmet Chem 1969; 20: 807

823.
5. Kahn G, Wilcox G, Wheeler BW, Atchison
W. Testing Sunscreens using sunlight and
photography. Br J Dermatol 1969; 81:
731736.
6. Cumpelik BM. Analytical procedures and
evaluation of sunscreens. J Soc Cosmet
Chem 1972; 23: 333345.
7. Groves GA. Evaluation of solar protective
preparations. Aust Soc Cosmet Chem 1977;
90: 3650.

Photodermatol Photoimmunol Photomed 2014; :

2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

PHPP_12105

8. Sayre RM, Agin PP, LeVee J, Marlowe E. A

comparison of in vivo and in vitro testing
of sunscreening formulas. Photochem
Photobiol 1979; 29: 559566.
9. Girard J, Lafille C, Unkovic J, Barbier A.
Predictive nature of laboratory methods
in sunscreen studies. Int J Cosmet Sci 1969;
20: 807824.
10. Sayre RM, Agin PP, Desrochers
DL, Marlowerm EL. Sunscreen testing
methods: in vitro predictions of effectiveness. J Soc Cosmet Chem 1980; 31: 133143.
9

Ruvolo Junior et al.

11. Diffey BL, Robson J. A new substrate to

measure sunscreen protection factor
throughout the ultraviolet spectrum. J Soc
Cosmet Chem 1989; 40: 127133.
12. Cole CA, Van Fossen RL. In vitro models
for UVB and UVA photoprotection. In:
Lowe NJ, Shaath NA, eds. Sunscreens
development, evaluation, and regulatory
aspects. New York: Marcel Dekker, 1990,
395404.
13. Herzog B. Prediction of sun protection
factors by calculation of transmissions
with a calibrated step film model. J Cosmet
Sci 2002; 53: 1126.
14. Pissavini M, Ferrero L, Alard V et al.
Determination of the in vitro SPF.
Cosmetics&Toiletries 2003; 118: 1063.
15. Stanfield JW. In vitro techniques in sunscreen development. In: Shaath N, ed.
Sunscreens: regulations and commercial
development, 3rd edn. Vol. 205. Boca
Raton: Taylor & Francis, 2005, 853
879.
16. Stanfield JW, Osterwalderb U, Herzog B.
In vitro measurements of sunscreen protection. Photochem Photobiol Sci 2010; 9:
489494.
17. Ferrero L, Pissavini M, Douceta O. How a
calculated model of sunscreen film geometry can explain in vitro and in vivo SPF
variation. Photochem Photobiol Sci 2010; 9:
540551.

10

18. Food and Drug Administration. Sunscreen drug products for over-the-counter
human use. Final Monogr Fed Register
2011; 76: 3562035665.
19. COLIPA. Colipa sun protection factor
test method. Bruxelles: The European
Cosmetic, Toiletry and Perfumery AssociationColipa, 1994.
20. Japan Cosmetic Association. Standard
SPF test method, revised eds. Tokyo: Japan
Cosmetic Industry Association, 1999.
21. Australian/New Zealand Standard. Sunscreen products evaluation and classification. Homebush, NSW and Wellington:
Standards of Australia and Standards of
New Zealand, 1998.
22. International Sun Protection Factor (SPF)
test method. Colipa: CTFA SA, JCIA,
CTFA, 2006.
23. ISO 24444: cosmetics sun protection test
methods in vivo determination of the
sun protection factor (SPF). 2010.
24. Ruvolo E Jr, Chu M, Grossman F, Cole C,
Kollias N. Diffuse reflectance spectroscopy for ultraviolet A protection factor
measurement: correlation studies between
in vitro and in vivo measurements.
Photodermatol Photoimmunol Photomed
2009; 25: 298304.
25. Gillies R, Moyal D, Forestier S, Kollias N.
Non-invasive in vivo determination of
UVA efficacy of sunscreens using diffuse

26.

27.

28.

29.

30.

reflectance spectroscopy. Photodermatol

Photoimmunol Photomed 2003; 19: 190
194.
Gillies R, Kollias N. Noninvasive in vivo
determination of sunscreen ultraviolet A
protection factors using diffuse reflectance spectroscopy. In: Lowe N, Shaath N,
Pathak M, eds. Sunscreens: development,
evaluation and regulatory aspects. New
York: Marcel Dekker Inc, 1997, 601608.
Fitzpatrick TB, Pathak M, Parish JA. Protection of human skin against the effects
of the sunburn ultraviolet (290320 nm).
In: Pathak MA, Harber LC, Seiji M and
Kukita A ed. Fitzpatrick TB, consulting
ed. Sunlight and man, normal and abnormal photobiological responses. Tokyo,
Japan: University of Tokyo Press, 1974,
751.
Colipa In Vitro Photoprotection Methods
Task Force. Method for the in vitro determination of UVA protection provided by
sunscreen products. Guideline 2007, 2nd
edn.
CIE. CIE 1976 uniform colour spaces.
In: Colorimetry, ed. Commission
Internationale de lEclairage, ed. Colorimetry. Vol. 15. Vienna: CIE Publication,
1986, 2932.
ISO 24443:2012, Determination of sunscreen UVA photoprotection in vitro.
2012.

Photodermatol Photoimmunol Photomed 2014; :

2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

PHPP_12105
All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.