Content uploaded by Zheng Li
Author content
All content in this area was uploaded by Zheng Li on Aug 31, 2018
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=yjac20
Journal of the American Institute for Conservation
ISSN: 0197-1360 (Print) 1945-2330 (Online) Journal homepage: http://www.tandfonline.com/loi/yjac20
Colorimetric sensor arrays: development and
application to art conservation
Maria K. LaGasse, Kristen McCormick, Zheng Li, Herant Khanjian, Michael
Schilling & Kenneth S. Suslick
To cite this article: Maria K. LaGasse, Kristen McCormick, Zheng Li, Herant Khanjian,
Michael Schilling & Kenneth S. Suslick (2018): Colorimetric sensor arrays: development
and application to art conservation, Journal of the American Institute for Conservation, DOI:
10.1080/01971360.2018.1495480
To link to this article: https://doi.org/10.1080/01971360.2018.1495480
Published online: 21 Aug 2018.
Submit your article to this journal
Article views: 15
View Crossmark data
Colorimetric sensor arrays: development and application to art conservation
Maria K. LaGasse
a
, Kristen McCormick
b
, Zheng Li
a
, Herant Khanjian
c
, Michael Schilling
c
and Kenneth S. Suslick
a
a
Department of Chemistry, University of Illinois at Urbana-Champaign, Urbana, IL, USA;
b
Walt Disney Animation Research Library, Glendale, CA,
USA;
c
Getty Conservation Institute, Los Angeles, CA, USA
ABSTRACT
Acceptable air pollutant concentration limits for sensitive artwork are generally at or below a few
ppb: this is only ∼1% of the permissible exposure limits for humans. Monitoring pollutants at
such low levels is an exceptional challenge, especially to do so in a cost-effective fashion for a
large number of locations and microenvironments (e.g., every display case in a museum). To
meet this challenge, we have extended our portable “optoelectronic nose,”by using new sensor
array chemistry to develop cumulative colorimetric sensor arrays with dosimetric sensitivities that
are dramatically better than commercial sensor tubes. The color changes of each sensor in a
disposable printed array produce a composite response to volatiles. Using cell phone camera
imaging, we have made field trials to monitor pollutant exposure of artwork from the Walt
Disney Animation Research Library during shipping to and exhibition in Beijing. This exhibition,
“Drawn from Life: the Art of Disney Animation Studios,”featured animation drawings, story
sketches, layouts, and concept art spanning the 90 years of the Disney Animation Studio’s
history. Sensor arrays monitored exterior and interior environments of passe-partout artwork
frames during exhibition and inside shipping crates during transport providing quantitative
information on oxidant, aldehyde, and sulfide pollutant exposure.
RÉSUMÉ
La concentration limite de polluant acceptable pour une œuvre d’art sensible est généralement plus
petite ou égale à quelques ppb : ceci étant environ 1% de la valeur limite d’exposition (NIOSH PEL)
fixée pour les humains. Surveiller les polluants à un niveau si faible est un défiscientifique
exceptionnel, particulièrement lorsqu’il s’agit de le faire de façon économique pour un nombre
élevé de lieux et de microenvironnements (ex. chacune des vitrines d’un musée). Afin de relever
ce défi, nous avons amélioré notre « nez optoélectronique » déjà extrêmement sensible et
portatif : en utilisant une nouvelle gamme de capteurs chimiques, nous avons développé une
matrice de capteurs colorimétriques dont la réaction est cumulative, possédant des sensibilités
dosimétriques infiniment meilleures que celles des tubes capteurs commerciaux. En suivant
numériquement le changement de couleur de chaque zone réactive sur une matrice imprimée
et jetable, une mesure quantitative de la réponse du réactif au contaminant volatil est obtenue.
Nous avons développé une plateforme utilisant le traitement d’image de l’appareil photo d’un
téléphone cellulaire et avons fait des expériences pratiques de mise à l’essai en suivant
l’exposition aux polluants des œuvres d’art provenant du Walt Disney Animation Research
Library pendant leur transport et leur présentation à Pékin. Cette exposition, « Drawn from Life:
The Art of Disney Animation Studios » mettait en valeur des dessins d’animation, des esquisses
de scénario, des schémas, des arrière-plans et des concepts artistiques couvrant les 90 ans de
l’histoire du Walt Disney Animation Studio. Les matrices de capteurs ont été utilisées pour suivre
à la fois les environnements extérieurs et intérieurs des passe-partout des cadres lors de
l’exposition, et l’intérieur des caisses de transport pendant leur déplacement, livrant de
l’information quantitative importante sur l’exposition aux oxydants, aux aldéhydes et aux
sulfures. Traduit par : Isabelle Cloutier.
RESUMO
Os limites aceitáveis de concentração de poluentes atmosféricos para obras de arte sensíveis são
geralmente iguais ou inferiores a algumas ppb (partes por bilhão): isto é apenas 1% dos limites
de exposição permitidos (NIOSH PEL) exigidos para seres humanos. Monitorar poluentes em
níveis tão baixos é um desafio científico excepcional, especialmente quando se quer fazer isso
com uma boa relação de custo-eficácia para um grande número de locais e microambientes (por
exemplo, todos os expositores de um museu). Para enfrentar esse desafio, ampliamos o nosso já
extremamente sensível e portátil “nariz optoeletrônico”: utilizando um novo sistema de sensor
químico, desenvolvemos matrizes de sensores colorimétricos cumulativos com sensibilidades
dosimétricas que são dramaticamente melhores do que os tubos de sensores comerciais.
KEYWORDS
Colorimetric sensor array;
preventive conservation; air
pollution; pollutants; sensors
© American Institute for Conservation of Historic and Artistic Works 2018
CONTACT Kenneth S. Suslick ksuslick@illinois.edu Department of Chemistry, University of Illinois at Urbana-Champaign, 600 S. Mathews Av., Urbana, IL
61801, USA
JOURNAL OF THE AMERICAN INSTITUTE FOR CONSERVATION
https://doi.org/10.1080/01971360.2018.1495480
Monitorando digitalmente as mudanças de cor de cada ponto do sensor em uma matriz descartável
impressa, é obtida uma medida quantitativa da resposta composta aos voláteis. Desenvolvemos
uma plataforma que se baseia em imagens de câmeras de telefones celulares e fizemos testes
de experimentos de campo para monitorar a exposição poluente de obras de arte da Walt
Disney Animation Research Library (Biblioteca de Pesquisa de Animação da Walt Disney) durante
o envio e a exposição em Pequim. Esta exposição, “Extraída da Vida: A Arte dos Estúdios de
Animação da Disney”, contava com desenhos de animação, esboços de histórias, layouts, planos
de fundo e arte conceitual que abrangem os 90 anos da história do Walt Disney Animation
Studio (Estúdio de Animação de Walt Disney). Conjuntos de sensores foram usados para
monitorar tanto ambientes externos quanto internos de molduras na exposição e dentro das
caixas de remessa durante o transporte, oferecendo valiosas informações quantitativas sobre
exposição a poluentes oxidantes, aldeídos e sulfuretos. Trudizo por Cristina Antunes, e revisado
por Beatriz Haspo.
RESUMEN
Los límites aceptables de concentración de contaminantes atmosféricos para obras de arte
sensibles son generalmente de unos pocos ppb o menos: esto es solo ∼1% de los límites de
exposición permisibles (NIOSH PEL) requeridos para humanos. Monitorear los contaminantes a
niveles tan bajos es un desafío científico excepcional, especialmente para hacerlo de una manera
rentable para una gran cantidad de ubicaciones y microambientes (por ejemplo, cada vitrina en
un museo). Para hacer frente a este desafío, hemos ampliado nuestra "nariz optoelectrónica" ya
extremadamente sensible y portátil: mediante el uso de nueva química de matriz de sensores,
hemos desarrollado matrices de sensores colorimétricos acumulativos con sensibilidades
dosimétricas que son dramáticamente mejores que los tubos sensores comerciales. Controlando
digitalmente los cambios de color de cada mancha de sensor en una matriz desechable impresa,
se obtiene una medida cuantitativa de la respuesta compuesta a volátiles. Hemos desarrollado
una plataforma que se basa en imágenes de cámaras de teléfonos celulares e hicimos
experimentos de campo de prueba para monitorear la exposición de las obras de arte de la
Biblioteca de Investigación de Animación de Walt Disney durante el envío y duración de una
exhibición en Beijing. Esta exhibición, "Dibujado desde la vida: el arte de los estudios de
animación de Disney" presentó dibujos de animación, bocetos de historias, diseños, fondos y
arte conceptual que abarca los 90 años de la historia del estudio de animación de Walt Disney.
Las matrices de sensores se usaron para monitorear tanto el ambiente exterior como el interior
de los marcos paspartú en la exhibición y dentro de las cajas de envío durante el transporte,
ofreciendo valiosa información cuantitativa sobre la exposición a oxidantes, aldehídos y sulfuros
contaminantes. Traducido por Amparo Rueda.
1. Introduction
One of society’s most important cultural responsibilities
is the preservation of the past for the future (Whit-
more 2005). The surest way to protect cultural heritage
objects from damage is to control the environment in
which they are displayed; both physical and chemical
factors, such as light, temperature, relative humidity
(RH), and so on, can have a profound impact on these
objects’lifetimes (Brimblecombe 1994; Camuffo et al.
2001). The discussion of slowing the deterioration of
museum materials through environmental control was
pioneered by Thomson (Thomson 1965; Thomson
1986) on the issue of artwork exposure to pollutants
and by Oddy on exposure to local environmental con-
ditions, including the microenvironment of the display
case itself (Oddy 1973). In conservation, the focus over
time began to shift from restoration to prevention, and
the field of preventive conservation as it is known
today was born (Whitmore 2005).
The intention of preventive conservation is to extend an
object’s lifetime by controlling environmental conditions
(microclimate, microbiology, and chemical pollution)
around a work of art. Over the last 50 years, the impor-
tance of measuring air pollution exposure has become
clear for preservation of cultural heritage materials.
Work in preventive conservation has revealed the dete-
riorative effects of even very low levels of specific pollu-
tants (Feng 2016; Grzywacz 2006; Hatchfield 2002;
Lucchi 2018; Smielowska, Marc, and Zabiegala 2017).
Improvements in detection methods, museum materials,
and HVAC environmental control have begun to
improve the environment to which artwork is exposed
(Blades et al. 2000; Camuffo et al. 2001; Ferdyn-Gry-
gierek 2014; Hisham and Grosjean 1991;Marćet al.
2015; Schieweck and Bock 2015). In order to continue
such progress, it is necessary to develop cost-effective
techniques for continuous control and monitoring of
pollutants for a large number of locations and microen-
vironments (e.g. every display case in a museum) at the
low levels needed to minimize the effects of common
pollutants (Table 1).
For this reason, a collaborative effort between the
University of Illinois, the Getty Conservation Institute
(GCI), and the Walt Disney Animation Research Library
2M. K. LAGASSE ET AL.
(ARL) has explored and developed, as a proof of concept,
a colorimetric sensor array for the monitoring of air pol-
lutants in situ during shipping and exhibition of artwork.
This work builds on the “optoelectronic nose”that has
been developed by the Suslick group at the University
of Illinois (Askim and Suslick 2015; Askim, Mahmoudi,
and Suslick 2013).
Colorimetric sensor arrays have proven to be a
broadly applicable technology that can successfully
differentiate among volatile organic compounds (Janzen
et al. 2006; Lin, Jang, and Suslick 2011), toxic industrial
chemicals (Feng et al. 2010a,2010b; Lim et al. 2009),
explosives (Askim et al. 2016), foods and beverages (Li
and Suslick 2016; Musto, Lim, and Suslick 2009; Suslick,
Feng, and Suslick 2010; Zhang and Suslick 2007; Zhang,
Bailey, and Suslick 2006), and bacteria and fungi (Carey
et al. 2011; Zhang et al. 2014). The core of past colori-
metric sensor array technology is an array of chemically
responsive dyes. These dyes serve as sensors by changing
colors due to dye–analyte interactions, e.g. Brönsted and
Lewis acid–base, hydrogen bonding, dipolar, and π–π
interactions (Askim et al. 2013). These sensors are essen-
tially “chemical fuses.”They are generally reversible for
most analytes (especially at lower concentrations), but
do need to be replaced after exposure to very high con-
centrations or to an aggressive analyte. Thus, the inter-
actions between gaseous analytes and colorants in our
earlier arrays mostly represent equilibrium interactions,
and for this reason, previous colorimetric sensors must
be imaged during exposure in real-time. Many of the
dye-analyte interactions are relatively strong, and so sen-
sitivities are often in the ppb regime, especially for
chemically reactive analytes (Askim et al. 2013; Askim
et al. 2016; Feng et al. 2010b). Importantly, the colori-
metric sensor arrays are relatively unaffected by changes
in RH or changes in temperature (Askim, Mahmoudi,
and Suslick 2013; Janzen et al. 2006; Rakow and Suslick
2000).
For monitoring of artwork, there are two significant
disadvantages for any equilibrium-based sensor: (1)
there is no improvement in sensitivity with increased
dosage (i.e. exposure time) and (2) as mentioned, arrays
must be imaged in real-time. The first disadvantage is
critical where long-term, but very low concentration,
exposures need to be monitored, i.e. the long-term sen-
sitivity of artwork to pollutants (Table 1). The second
disadvantage, real-time imaging, complicates real-world
use of the sensor array during exhibition conditions.
To overcome these disadvantages, we have altered the
dye–analyte chemistry to make the reactions irreversible
and thus the analyte–dye reactions become cumulative
(dosimetric) sensors where the color change is mono-
tonic (and generally linear) as a function of dose until
saturation (i.e. when all dye has reacted). Pollutant con-
centration measurements are therefore representative of
a time-weighted average (TWA) in units of concen-
tration multiplied by exposure time (e.g. ppb-days).
Essentially, our approach draws on colorimetric irrevers-
ible reactions already in use with current direct-reading
passive sampling devices (Grzywacz and Stulik 1992),
such as Draeger tubes (Draeger Co. 2016), but minia-
turizes the reactive media dramatically, includes dozens
of analyte-specific sensors (i.e. each spot), and substan-
tially improves sensitivity and accuracy by use of digital
imaging. Given the relatively high cost of individual
sampling tubes, using a printed disposable array also
has the promise of a very large decrease in operating
expenses.
Each sensor includes a dye-indicator whose reactions
with the target analyte are essentially irreversible, i.e.
have enthalpies greater than ∼−150 kJ/mol (e.g. metal
sulfide precipitations, strong coordination to metal
ions, irreversible oxidations of dyes or bleaching, redox
reactions with large ΔE°). In addition, to de-convolute
the response to the total environment into responses
from individual analytes or classes of analytes, each chro-
mophore is chosen to react more specifically with a class
of analytes (i.e. oxidants, sulfides, aldehydes, acids),
which is a departure from the cross-reactive equilibrium
sensor spots of our earlier arrays. These sensor spots
were compiled into a first-generation array for detection
of multiple museum pollutants; the array incorporates
into one cohesive system a series of chemically respon-
sive dyes that respond specifically to the main museum
pollutants (Figure 1 and Supplemental Material (Table
SM1)). Digital images are taken of the sensor array
before and during exposure to pollutants, and color
difference maps (Figure 2) are generated by digital sub-
traction, pixel by pixel, of the center portion of sensor
spots in the array before and after exposure. The result-
ing changes in red, green, and blue values are used for
quantitative analysis of the data, as described below in
Section 2.3.
Table 1. Important pollutants found inside museums and
suggested concentration limits for cultural heritage materials
(Grzywacz 2006).
Major
pollutants
Suggested limits for
sensitive materials (ppb)
Suggested limits for other
collection aterials (ppb)
NO
2
0.05–2.6 2–10
O
3
<0.5 0.5–5
SO
2
0.04–0.4 0.4–2
H
2
S <0.1 <0.1
Acetic acid <5 40–280
Formic acid <5 5–20
Formaldehyde <0.1–510–20
Acetaldehyde <1–20
JOURNAL OF THE AMERICAN INSTITUTE FOR CONSERVATION 3
2. Materials and methods
2.1. Materials, formulations and array printing
Certified, premixed gas tanks were obtained from Math-
eson Tri-Gas Corp. through S. J. Smith, Co. (Urbana, IL).
Gas streams were prepared by mixing the analyte stream
with dry and wet nitrogen gas. MKS mass flow control-
lers were used to achieve the desired concentrations and
RH. Importantly, gas stream concentrations and RH
were confirmed by real-time, in-line analysis using an
MKS multigas analyzer (model 2030, a dedicated Fourier
transform infrared spectrophotometer).
All other reagents were an analytical-reagent grade, pur-
chased from Sigma-Aldrich and used without further
purification. Each sensor spot was optimized for best
chemical compatibility and sensor performance (formu-
lations given in Supplementary Material (SM), Table SM1).
Sensor arrays were robotically printed in a linear geo-
metry using custom-designed rectangular pins (Figure.
SM1) installed in an Array-It NanoPrinter (Askim
et al. 2016). Printing of arrays was usually on a white
polypropylene membrane (Sterlitech Corporation; thick-
ness: 130–170 μm, pore size: 0.22 μm) and attached to an
imaging platform (described in detail in Section 2.2)
using a solvent welding method (dichloromethane).
Further experimental details are available elsewhere
(LaGasse 2016).
2.2. Passive sampling techniques
Array response has been explored for passive sampling
environments, in which air and pollutants diffuse to
the surface of the sensor spots without active circulation
or pumping. There are two different modes of exposure
and imaging used in this study (shown schematically in
Figure 3): (1) passive sampling and real-time imaging
using a flatbed scanner and (2) passive sampling and
Figure 1. Photographic image of a colorimetric sensor array for
artwork monitoring, 5×30 mm. The array incorporates, into one
disposable strip, 24 cumulative chemically responsive dyes that
are dosimetric for specific classes of analytes relevant to the
museum environment. These include redox indicators for
oxidants (ozone, NO
2
), precipitory metal ions (for sulfides), pH
indicators (SO
2
, acetic, formic acids), and Brady and Schiffreac-
tants (aldehydes).
Figure 2. Images of a 24-dye colorimetric sensor array, (Top)
before exposure and (Middle) after exposure to 1.5 ppm ozone
for 2.0 min; images obtained with an ordinary flatbed scanner.
(Bottom) Subtraction of the two images yields a difference vector
in 3N dimensions (i.e. N changes in red, green, and blue color
values, where N is the number of sensor spots); the difference
vector is represented by a difference map that shows the absol-
ute values of the color changes of each spot. The gray boxes indi-
cate the sensor spots that are most responsive to ozone. For
purposes of visualization and display, the color range of differ-
ence maps is usually expanded; in this case, the RGB color values
were expanded from 6 bit to 8 bit color; all quantitative analysis,
however, uses only the raw measured differences unaltered.
Figure 3. (A) Passive sampling setup using a flatbed scanner for
array imaging. Injection of a calibrated concentration and volume
of the analyte was used to prepare a fixed gas concentration pas-
sive sampling environment. (B) Cellphone imaging platform and
holder for an iPhone 4S. (C) Sensor array for imaging by cell-
phone camera. To generate reproducible images of the sensor
array, the cellphone in the custom holder was positioned so
that the trapezoid filled the camera’s rectangular field of view;
the trapezoid was sized so that the perspective through the cell-
phone camera appeared as a rectangle.
4M. K. LAGASSE ET AL.
periodic imaging using an iPhone 4S camera. The former
was used primarily for calibration purposes and the latter
primarily for field-testing during shipping and exhibi-
tion. In general, we observe significantly better S/N in
the difference maps generated with the flatbed scanner
compared to the iPhone camera, due primarily to higher
noise in the iPhone images.
2.2.1 Passive sampling with flatbed scanner
For passive sampling with flatbed scanner imaging, sensor
arrays were imaged face down on the flatbed scanner
(Epson V600) through a clear zip-top polypropylene
bag (6 Gal., filled using mass flow controllers to a total
volume of 22 L) with a rubber septum and 3-way valve
attachment (Figure 3A). The array was placed in the bag
facing downward to the scanner surface, the bag was
closed and excess air removed. The bag was then filled
with a fixed volume (7.8 L) of 50% RH filtered air. After
3 min, a “before-exposure”image of the array was
taken. Analyte gas was then drawn out of another bag
using a syringe through a rubber septum attachment
and injected into the experimental bag containing the
array to create an environment with the desired concen-
tration of analyte. After analyte addition, images of the
arrays were taken at set time increments during exposure.
2.2.2. Passive sampling with cellphone imaging
For passive sampling with cellphone imaging, images
were acquired using an iPhone 4S held in a custom
mount machined to hold the iPhone at a fixed 30
o
angle and at a fixed height above the array surface
(Figure 3B). This mount, when used in conjunction
with the trapezoid outline printed on the array-imaging
platform (Figure 3C) allowed images of each array to be
taken from the same position. To obtain an image, the
phone-mount setup was moved towards the imaging
platform until the sides of the trapezoid perimeter
aligned with the edges of the camera display. Due to per-
spective, the trapezoid perimeter when viewed through
the camera becomes the rectangular outer edge of the
camera’sfield of view. Triplicate images were taken
each time and later averaged to minimize noise.
The imaging platform consists of sensor spots printed
on a white polypropylene membrane with a neutral gray
reference strip mounted 1 cm below sensor spots. The
polypropylene and gray reference strip was then attached
to an impact resistant polycarbonate film the size of a
glass microscope slide (McMaster-Carr; 1”x3”x0.040”)
and mounted on acid-free paper (HP Premium Choice).
Printed on the acid-free paper was (1) an identifying
number unique to each array, (2) the date the array
was printed, and (3) a trapezoid outline that was sized
to aid in image alignment (as discussed above).
To make field use as easy as possible for the curator,
no special effort was taken to isolate the system from
ambient lighting. To reduce the effect of changes in
ambient lighting, each image was taken with the camera
flash illumination turned on. Illumination of the array, as
expected, was not uniform even with the use of the cam-
era flash. There are inevitable differences in the illumina-
tion across the field of view, in both horizontal and
vertical directions. As shown in Figure 4, these differ-
ences are significant, and lighting differences across the
sensor array surface or from image to image can deviate
as much as ±20% from the center of the array outwards.
For this reason, the RGB values obtained for each spot
have been normalized using the white surface of the poly-
propylene membrane on which the sensor spots were
printed as a reference. The areas on both sides of each sen-
sor spot were used to obtain an average white reference
RGB value for each spot; this reference value was used
to normalize the lighting illuminating each spot
(Figure 4C). Each spot across the array for each image
at each time point has a measured R, G and B value (i.e.
X
meas
where X= R,G,B) and a corresponding value for a
white standard (i.e. Xwhite) averaged from both sides of
the spot for the white reference (Figure 4C). A corrected
RGB value, Xcorr can be defined as (Xmeas/Xwhite )∗255.
This simple correction produces roughly a twofold
Figure 4. Sensor array imaging with cell phone camera. (A) Digi-
tal image of the array sensor spots taken with an iPhone 4S cam-
era. (B) Digital image of an unprinted polypropylene membrane
(i.e. no sensor spots) to show the non-uniformity of lighting. (C)
False-color gradient map of relative reflectivity of the unprinted
membrane showing non-uniformity of lighting; deviation is ±
20% over the array; false color reference scale also shown. (D)
Enlarged view of several sensor spots showing the areas aver-
aged for each of the dye spots (red outline, X
meas
) and for the illu-
mination normalization (black outline) on both sides of each dye
spot, averaged to give X
white
.
JOURNAL OF THE AMERICAN INSTITUTE FOR CONSERVATION 5
improvement in the reproducibility of RGB values
obtained for the sensor spots over a wide range of external
illumination (e.g. fluorescent vs. halogen lamp at 5000 K
vs. halogen at 3000 K). Attempts to make a more sophis-
ticated correction to the RGB values using an adjacent
gray standard were unsuccessful due to the non-uniform-
ity of illumination in the vertical direction.
2.3. Data processing and analysis techniques
From digital images of the sensor array before and after
exposure to pollutants, a color difference map (Figure 2)
is easily generated by digital subtraction, pixel by pixel, of
the center portion of each of the sensor spots in the
array: red value after exposure minus red value before,
green minus green, blue minus blue. The resulting data is
inherently digital (simply a vector of 3N dimensions
where N= total number of spots), and all quantitative
and statistical analysis is done directly from the digital
difference vectors, i.e. the ΔRGB values. The color differ-
ence maps are useful primarily for convenient visualization
of color changes of the dyes in the array; note that the color
values shown in the difference maps are the absolute values
of the differences and that expansion of the color space is
useful for visualization of small color changes.
The changes in red, green, and blue values (ΔRGB)
from each sensor element at a given time point can be
combined into a Euclidean distance at any given time,
(ED
t
), defined by Equation (1):
EDt=(DR2
1+DG2
1+DB2
1+DR2
2+...+DB2
n)1/2
t(1)
where nis the number of spots under consideration and t
is the time of exposure. To generate a response profile for
a given analyte, the average Euclidean distance for mul-
tiple trials (ED
t
) is plotted with respect to time. From this
response profile, reaction kinetics can be determined. A
dosimetric sensor gives a nearly linear response until sat-
uration, and the slope of this response profile as a func-
tion of analyte concentration determines the dosimetric
sensitivity (DS) and limit of detection (LOD) (discussed
in greater detail in Section 3.2).
3. Results and discussion
3.1. Pollutant discrimination by the colorimetric
sensor array
The sensor array was evaluated for its ability to discriminate
among pollutant classes (i.e. oxidants, sulfides, aldehydes,
and acids). As shown qualitatively by the difference maps
(Figure 5), the response patterns to oxidants (ozone, nitro-
gen dioxide), sulfides (hydrogen sulfide), and aldehydes
(formaldehyde) are substantially different.
Finer discrimination among analytes of the same class
using colorimetric sensor arrays is generally possible
using standard pattern recognition techniques, most
notably hierarchical cluster analysis (HCA) (Askim,
Mahmoudi, and Suslick 2013; Lim et al. 2009). This
finer discrimination was not a major goal of the work
described here. Nonetheless, in order to obtain a rough
sense of the discrimination ability of the cumulative sen-
sor array developed in this work, we examined discrimi-
nation among each of the six analytes and controls
(difference maps shown in Figure 5) and also among
different concentrations of a single analyte (e.g. NO
2
)
using standard pattern recognition clusterification
(Figure SM2). The resulting HCA dendrograms showed
excellent discrimination among all analytes tested with
the exception of acetic acid vs. SO2.
In the version of the cumulative sensor array used for
the Disney Exhibition, detection and identification of
acid analytes was imperfect in another sense as well.
Over the prolonged exposure time from artwork framing
to shipping to exhibition to reshipping, that the acid sen-
sitive sensor elements were exhausted by continual
exposure to atmospheric CO
2
(∼400 ppm), which is,
after all, an acidic gas. As mentioned, the array was
also difficult to differentiate the two acidic analytes of
primary interest (SO
2
and acetic acid) from one another
Figure 5. Color difference maps of arrays exposed to the major
museum pollutants imaged using a flatbed scanner. All pollu-
tants are shown at 7 ppm except ozone, which is shown at
1.5 ppm due to its higher responsivity. The sensor spot numbers
(as in Table SM1) are provided at the bottom of the figure.
Exposure utilized a two-minute exposure with a flow rate of
500 sccm. For visualization purposes, the RGB color values
were expanded from 6 bit to 8 bit color. For quantitation, alde-
hyde response was measured from the color change of spot 7;
H
2
S, spot 9; acids, spot 23; and oxidants, spot 19.
6M. K. LAGASSE ET AL.
(Figure SM2). Further work with metal nanoparticle sen-
sor inks is underway to address this limitation.
3.2. Dosimetric sensitivity and TWA exposure
With cumulative sensors, the definition of detection sen-
sitivity is a function of both analyte concentration and
exposure time: dosimetric sensitivities are expressed in
units of concentration-time (e.g. ppb*days), which
expresses the lowest TWA exposure of some specific
analyte that the sensor array is capable of measuring
over some length of time.
This DS has been determined experimentally for
hydrogen sulfide, formaldehyde, nitrogen dioxide, and
ozone. Using a passive sampling environment with ima-
ging on a flatbed scanner (Figure 3A), arrays were
exposed to the target analyte at 0, 62.5, 125, 250, 500,
and 1000 ppb in 50% RH filtered air for up to 2 h. As
an example, the color difference maps and their analysis
representing the response of the array to hydrogen
sulfide are shown in Figure 6.
From the data collected in these experiments, one may
define an apparent LOD (LOD
app
) for exposure over a
specificfixed time to a specific analyte concentration
by Equation (2):
LODapp =(3∗N∗[A])/EDt(2)
where [A] is analyte concentration in ppm, Nis the noise
determined from multiple images of the same array, and
ED
t
is the Euclidean distance of the color changes
(ΔRGB) of the single most responsive spot at a given
time (e.g. after 2 min exposure). The extrapolated LOD
was determined by plotting LOD
app
vs. analyte concen-
tration (ppm). A second order polynomial fit was used
to extrapolate the LOD as the analyte concentration,
[A],approaches the LOD
app
.
The DS (in units of ppb*days), was then determined
using the Equation (3) where t
exp
is the exposure time
in minutes for which the LOD was determined:
Dosimetric Sensitivity (ppb∗days)=DS
=LOD (ppb)∗texp(min)
1440 min/1day (3)
Processing of the data from these experiments can be
taken a step further to produce a calibration curve to give
a TWA exposure (in ppb) over some total exposure time.
If one plots the optimal sensor response (i.e. the Eucli-
dean distance of the color change of the most responsive
spot) per exposure time vs. analyte concentration, then a
linear fit of this data (Figure 6C) can be used to deter-
mine the TWA analyte concentration from the color
changes recorded in real-world field observations.
Table 2 gives the dosimetric sensitivities achieved with
our array in passive sampling environments. For all ana-
lytes, sensor array sensitivities are vastly better than stan-
dard Draeger tubes for each specific analyte. The sensor
arrays are far less expensive to produce compared to
Draeger tubes and a single sensor array works for multiple
analytes. Due to the open exposure of our printed sensor
arrays to the atmosphere, passive sampling sensitivities
Figure 6. Data used to determine DS and time-weighted averages
(TWA) for hydrogen sulfide (H
2
S) exposure in a passive sampling
environment over 10 min. (A) Difference maps of the sensor
array color changes showing array response vs. H
2
S concentration;
for visualization, the RGB color values were expanded from 3 bit to
8 bit color (i.e. RGB values 3–10 expanded to 0–255). The red box
indicates the most responsive sensor spot. (B) Response profiles
vs. time, depicting the linear, dosimetric response of the most
responsive spot. (C) Calibration curve plotting optimal sensor
response (i.e. ED of the color change of the most responsive
spot) per exposure time vs. analyte concentration. In field tests,
the measured ED of that sensor spot divided by total time of
exposure determines the TWA analyte concentration.
JOURNAL OF THE AMERICAN INSTITUTE FOR CONSERVATION 7
are improved substantially compared to conventional
diffusion tubes. Quantitative comparisons to tests invol-
ving tarnishing of metal surfaces, i.e. the Oddy test devel-
oped originally at the British Museum in 1973 (Oddy
1973) and further expanded since (Bamberger 2012; Bam-
berger, Howe, and Wheeler 1999; Wang et al. 2011; Wang
et al. 2012), are difficult due to the long evaluation times
(days to weeks) required by the Oddy test and the subjec-
tive and qualitative nature of the evaluation (Pretzel and
Shibayama 2003; Wang et al. 2011).
3.3. Field-testing of sensor arrays
The first exhibition of artwork from the Walt Disney
Animation studios took place at the National Museum
of China in Beijing from June 30 through August 18,
2015. Featuring over three hundred art objects, many
of which have not been shown before to the public,
Drawn from Life: The Art of Disney Animation demon-
strated the 90-year legacy of the Walt Disney Animation
Studios (Figure 7).
To test the sensor arrays for application to monitoring
the exposure of artwork during shipping and exhibition,
a collaboration among the University of Illinois (UIUC),
the Walt Disney Animation Research Library (ARL), and
the Getty Conservation Institute (GCI) were established.
Sensor arrays were placed at select locations (i.e. inside
and outside of sealed and framed artworks, in sampling
boxes placed in galleries, and in shipping crates storing
artwork during travel). The sensor arrays were imaged
at key times during the exhibition process: when initially
mounted to the artwork, upon arrival at the Beijing exhi-
bition site, and upon departure from the exhibition. All
imaging of the sensor arrays used the digital camera of
an iPhone 4S, as described earlier in Section 2.2.2.
3.3.1. Sensor arrays for passe-partout frames
The method of passe-partout mounting, generally used
for works on paper, produces a microclimate for the art-
work inside a mat package. This packaging is used to pro-
tect artwork from ambient pollutants potentially present
in the atmosphere surrounding works of art. Two sensor
arrays were mounted to the back of eight separate pieces
of artwork: one inside the transparent polyester back
wrapping and one outside of the backing (Figure 8).
The list of the passe-partouts that were monitored with
sensor arrays, a description of the composition of the
art, crate number used for shipping to the Beijing exhibi-
tion, and floor plan of the gallery space are given in Table
SM2 and Figure SM3.
During control experiments conducted at UIUC, we
discovered that the interior of the passe-partout was off-
gassing a sulfide contaminant. To determine which
passe-partout material was responsible for the sulfide
emission, the cellphone imaging platform was used to
image arrays exposed to each of the passe-partout
materials individually in a passive sampling environment
(Figure 3A). Array exposure to the materials was done by
first imaging the array, then placing the array in a small
polyethylene bag with the material of interest, heat-sealing
the bag under nitrogen and then imaging again 8 days
later. In addition to the passe-partout materials, we also
tested the material used to make the card used as a gray
calibration reference used in the imaging platform itself.
Table 2. Dosimetric sensitivities (DS) achieved during passive
sampling with the colorimetric sensor array compared to
Draeger tubes, as taken from online product literature (Draeger
Co. 2016).
Analyte
Sensor array
(ppb*days)
Draeger
tube
(ppb*days)
Improvement
factor
Hydrogen
sulfide
0.05 430 8600×
Formaldehyde 1.12 70 60×
Nitrogen
dioxide
0.03 430 14000×
Ozone 0.04 20 500×
Note: Improvements in DS are given.
Figure. 7. Drawn from Life: The Art of Disney Animation, the first
exhibition of artwork from the Walt Disney Animation studios
shown at the National Museum of China in Beijing from June
30 through August 18, 2015. Artwork
©
Walt Disney Co.
8M. K. LAGASSE ET AL.
The TWA concentration of sulfides off-gassing from each
material was determined, as shown in Figure 9. The source
of the sulfide off-gassing was unambiguously determined
as the acrylic sheet used as the front glassof the passe-part-
out (Acrylite OP3-plex), with an effective TWA emission
of roughly 200±80 ppb*days.
3.3.2. Sensor arrays in shipping crates
The shipping crates used to transport artwork for the
exhibit were made from AC grade Plywood and heat-
treated, kiln dried #2-grade pine with Tyvek liner. Sensor
arrays were mounted on the interior (i.e. inside of the
Tyvek liner) of the three shipping crates that were used
to transport the art pieces that had sensor arrays
mounted on the back of their passe-partout frames
(both inside and outside).
To differentiate any changes in the sensor arrays due
to aging from changes that occur due to pollutant
exposure, a comparison was made to non-traveling con-
trol arrays. Images were taken at the same length of time
after printing of both the traveling sensor arrays
mounted on crates or artwork and of a separate set of
three arrays kept in a controlled 50% RH filtered air
environment at the University of Illinois: i.e. analysis
was made on the difference of differences, ΔRGB
Traveling
minus ΔRGB
Control
. The traveling and control sensor
arrays all came from the same printing batch. To moni-
tor any aging effects on the arrays, a set of control arrays
were imaged about once a week on-site at UIUC using an
imaging platform identical to the one used to image the
traveling arrays, as described in Section 2.2.2; the air
inside the storage bag containing the control arrays
was flushed with fresh filtered air weekly.
A semi-quantitative analysis of pollutant exposure
was generated as described in Section 3.2. The response
of the oxidant, aldehyde, and sulfide sensitive spots of
the sensor arrays was converted to a dosage expressed
as a TWA Daily exposure (ppb*days) for each class of
pollutant, as shown in Figure 10. The comparison was
made between the sensor arrays mounted inside the
crates and the arrays mounted on the outside of the
passe-partout frames (labeled as outside-PP) of artwork
carried inside the crates.
Oxidant exposures measured by the arrays mounted
to the crates and by the arrays outside of the passe-
partout frames within the crates are quite similar
between the crate arrays and the outside-PP arrays
(Figure 10 left). Given that the oxidant sources during
shipping are external (e.g. NO
x
and ozone in the
atmosphere), these results are as expected. Figure 10
does, however, emphasize the importance of pollutant
exposure during shipping, even during the relatively
short period of flying time.
Aldehyde exposure was generally higher with the sen-
sor arrays mounted inside the crates than with the arrays
mounted outside the passe-partout of artwork within the
crates, by roughly 40% (Figure 10 middle). This is likely
due to the well-known emission of formaldehyde from
wood (Schäfer and Roffael 2000). Even with the use of
Tyvek liners, the close proximity of the crate arrays to
the wood crate material appears to be the likely source
of higher aldehyde exposure. The containment of the
passe-partout frames within wrappings with in the crates
appears to partially ameliorate aldehyde exposure.
Figure 8. Passe-partout framing used by the Walt Disney Anima-
tion Research Library. Artsorb sheets are included to serve as a
humidistat. Two colorimetric sensor arrays have been mounted
to the back of the artwork: one inside the transparent polyester
back sheet and one outside of the backing. The cel depicting the
Boat Builders scene (bottom left) is an example of one piece in
the exhibition. Artwork
©
Walt Disney Co.
Figure 9. TWA concentration of sulfide emission calculated from
array response to the individual materials in the passe-partout
packaging. The acrylic sheet is the only material that shows sig-
nificant off-gassing of sulfides (200±80 ppb*days).
JOURNAL OF THE AMERICAN INSTITUTE FOR CONSERVATION 9
In contrast, sulfide exposure of arrays mounted to the
artwork is consistently and substantially higher than the
sensor arrays mounted to the crates (Figure 10 right).
This is probably due to the discovered source of sulfide
out-gassing (the acrylic glass, cf. Section 3.3.1) from
the pass-partout.
3.3.3. Results from Beijing exhibition
By having two arrays mounted on the back of artwork
with one outside the passe-partout and one inside, we
have an immediately useful comparison of the external
pollutant exposure to that actually seen by the artwork
inside the sealed frame. Qualitatively, this can be easily
illustrated by the color difference maps generated from
subtracting the RGB values of each spot in the image
taken at time of mounting from the image taken at the
end of the exhibit, in this case, 77 days later. Such a com-
parison of difference maps is given in Figure SM4.
In a qualitative comparison of the magnitude of color
change from arrays mounted on the outside of the passe-
partout vs. the inside of the passe-partout, it is clear that
the passe-partout mounting is very effective in keeping
atmospheric pollutants from interacting with the art
pieces (Figure SM4 and SM5). We see multiple spot
responses that are much greater for the sensor array out-
side of the passe-partout than for the sensor inside. We
suggest that a significant part of this successful dimin-
ution of pollutants inside the passe-partout may be
due, at least in part, to the sorption of pollutants by
the Artsorb, which is a high surface area silica in sheet
form meant to control humidity. One might expect the
poly film that covers the back of the passe-partout to
be a minor barrier to pollutants over a period of a few
days, but the construction of the passe-partout is cer-
tainly not hermetically sealed over the period of a
month. Silica gel (which makes up Artsorb) is certainly
well-known as a general sorbent for VOCs. Further
quantitative investigation of the sorption of VOCs and
pollutants by Artsorb is warranted.
In contrast, the sulfide sensitive spot (#9, boxed in red
in Figure SM4) showed nearly no exposure in the sensor
arrays mounted outside of the passe-partout, whereas the
same sensor spot on arrays inside the passe-partout
responded strongly to sulfides produced inside the sealed
frame. As previously discussed, this is due to the internal
source of sulfides inside the passe-partout: the acrylic
glass within the closed microenvironment inside the
passe-partout frame.
Figure SM5 shows quantitative response profiles for
spots sensitive to each class of pollutants from (1) the
date the arrays were mounted at the ARL in Glendale,
(2) to their arrival at the Beijing Exhibition (19–21
days), and (3) up to the conclusion of the exhibition
(77 days after the original mounting of arrays). Average
total array responses from all artwork are given for sen-
sor arrays mounted outside of the passe-partout frames
vs. those inside the passe-partout at all three time-points.
For the arrays mounted in the shipping crates, the aver-
age response is given at the end of shipping to Beijing
(days 19–21); as discussed earlier, during shipping, the
aldehyde and oxidant responses were similar among
arrays mounted to the outside and inside of the passe-
partout packaging.
The total responses of the sensor arrays during the full
77 days of the shipping plus exhibition are shown in
Figure SM5. Oxidant exposure outside of the passe-
Figure 10. Oxidant (left), aldehyde (middle), and sulfide (right) exposure of individual arrays mounted to pine plywood crates with
Tyvek liners compared to average response of sensor arrays mounted to the outside of artwork passe-partout (PP) frames within
the same crates. TWA exposures (ppb*days) were measured on arrival in Beijing at days 19–21 after initial mounting of arrays in ARL.
10 M. K. LAGASSE ET AL.
partout is dramatically higher than that measured inside,
by a factor of more than 7.0 at day 77. Aldehyde response
inside the passe-partout during exhibit was nearly
unchanged from its exposure during shipping, but the
external sensor continued to respond to atmospheric
aldehydes; total exposure to aldehydes diminished
slightly (but within the error bars of our analysis) by
the end of the exhibit. These results support the effective-
ness of the sealed frame at preventing oxidants and alde-
hydes present in the gallery space from coming into
contact with the artwork itself. Consistent with our pre-
vious discussion, the environment inside the passe-part-
out has a significantly higher sulfide level than the
outside environment due to the internal sulfide source
within the passe-partout.
Exposure during the exhibition itself is shown in
Figure 11. For this analysis, the before and after images
of the sensor arrays were those taken of the arrays
upon arrival in Beijing (days 19–21) and at the end of
the Exhibition in Beijing (day 77). Figure 11 gives the
time-weighted daily exposure (ppb*days) to sulfides,
aldehydes, and oxidants over the course of the Beijing
Exhibition (≈57 days). The daily exposures plotted in
Figure 10 are the same scale as Figure 11. Notably, the
exposure outside of the passe-partout during shipping
is high compared to the relatively controlled exhibition
environment (especially for aldehyde exposure).
The difference between oxidant exposure of the arrays
mounted to the outside of the passe-partout over the
course of the Beijing exhibition is substantial, greater
than a threefold difference. This reaffirms that the
passe-partout is effective in preventing oxidants from
the environment outside the passe-partout from building
up within the passe-partout. Daily exposure to aldehydes
was higher outside the passe-partout than inside, but still
within the error bars of the analysis.
Daily exposure to sulfides for arrays mounted inside
the passe-partout during the Beijing Exhibition is sub-
stantially higher (by roughly sevenfold) than the
exposure outside. The larger gallery environment sub-
stantially decreased the concentration of sulfides outside
of the passe-partout microenvironment: sulfide concen-
trations within passe-partouts (either in control exper-
iments at UIUC or in field tests in Beijing) were ∼200
ppb*days vs. 25 ppb*days outside of the passe-partout.
In taking the analysis of sulfide exposure a step
further, systematic differences were observed depending
on the nature of the material making up the artwork.
Sensor arrays mounted on the inside of a passe-partout
encasing a painted cellulose acetate cel (whether it was
a reproduction or an original) showed significantly
lower average daily dosage of sulfide than arrays inside
a passe-partout encasing an animation drawing on
paper (i.e. graphite, pen, or Conte crayon) (Table 3).
Location within the galleries did not affect the exposure
measured. One possible explanation is that the cellulose
acetate substrate or the paint upon it absorbs sulfide
volatiles more effectively than the paper drawings.
Figure 11. TWA daily exposure (ppb*days) to oxidants (left), aldehydes (middle), and sulfides (right) over the course of the Beijing
exhibition for each of the environments listed on the x-axis. Dosage (ppb*days) compared the array images upon arrival in Beijing
(days 19–21) to images at the end of exhibition (day 77). Note that the Daily Exposures are plotted on the same scale as Figure 10.
Labels: Inside-PP, arrays mounted to the inside passe-partout; Outside-PP, arrays mounted to the outside of the passe-partout; Con-
trol-PP, non-traveling arrays mounted to passe-partout packaging imaged at UIUC; Control no PP, arrays aged without the presence
of passe-partout, exposed to 50% RH filtered air environment and imaged at UIUC.
JOURNAL OF THE AMERICAN INSTITUTE FOR CONSERVATION 11
4. Conclusions and outlook
A colorimetric array of cumulative (dosimetric) sensors
has been developed for the portable and inexpensive
monitoring of air pollutant exposure of cultural heri-
tage objects. A system for exposing these arrays in pas-
sive sampling environments has been used to
determine dosimetric sensitivities to important classes
of pollutants and to provide TWA concentrations of
pollutants in the museum and exhibition environment.
To quantitatively measure the color changes of these
sensor arrays, a facile imaging method using an ordin-
ary cellphone camera has been created and proven
functional even during exhibition and shipping of art-
work. With a colorimetric sensor array and cellphone
imaging, we were able to calculate TWA daily exposure
(in ppb*days) to sulfides, oxidants, and aldehydes in a
variety of environments. Further work is needed (and
underway) for identification of acidic pollutants during
prolonged exposure, in large part due to the cumulative
effects of long-term (months) exposure to atmospheric
carbon dioxide.
The sensor arrays were used to monitor the environ-
mental exposure to pollutants of artwork from the Dis-
ney Studios during shipping to and on display in
Beijing as part of the Drawn from Life: The Art of Disney
Animation exhibition. We have demonstrated quantitat-
ively the effectiveness of the passe-partout packaging in
protecting artwork from sulfide, oxidant, and aldehyde
pollutants as well as determine a source of emission of
sulfides from the passe-partout materials themselves.
The importance of exposures to pollutants during ship-
ping is particularly worth noting.
Acknowledgements
This work was carried out in part at the University of Illi-
nois at Urbana-Champaign, the Walt Disney Animation
Research Library, and the “Drawn from Life: the Art of
Disney Animation Studios”exhibition site at the
National Museum of China in Beijing. K.S.S. would
like to thank Joy Mazurek for her early efforts to establish
this collaborative effort.
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
This work was supported by the National Science Foundation
(CHE-1152232). MKL gratefully acknowledges fellowship sup-
port from John and Margaret Witt Fellowship Fund.
Notes on contributors
Maria K. LaGasse is currently a research scientist at Novation
iQ in Kansas City developing new high-performance materials
and polymers; she received her B.S. in chemistry from St. Johns
University and her Ph.D. under the direction of K. S. Suslick in
2016. Her work involved a number of applications of colori-
metric sensor arrays, including the applications to artwork
monitoring described in this paper.
Kristen McCormick is the Art Collections & Exhibitions Man-
ager of the Walt Disney Animation Research Library, Glen-
dale. Kristen has been at the Walt Disney Company for over
a decade and a half where she has been responsible for the
safe keeping, care, and transport of a broad range of artworks
from African Art to Animation. In her current role, she over-
sees the conservation care of the Walt Disney Animation Col-
lection which comprises of over 64 million pieces of artwork,
from all facets of the production process including storyboard
drawing, visual development, layouts, backgrounds, animation
drawings, and animation cels. Kristen is also responsible for
the traveling exhibitions and loans related to this Collection,
with extensive tours around the world, including Taiwan,
Korea, Australia, Japan, China, and Europe.
Zheng Li is currently a postdoctoral research fellow at North
Carolina State University working with Professor Qingshan
Wei in the Department of Chemical and Biomolecular Engin-
eering. He received his Ph.D. at the University of Illinois under
the direction of K. S. Suslick in 2017, exploring several diverse
applications of the optoelectronic nose, including detection
and identification of explosives and fuels and discrimination
among single malt scotches.
Herant Khanjian is an assistant scientist at the Getty Conser-
vation Institute in Los Angeles. He received his B.A. in chem-
istry at California State University, Northridge, and has been at
the GCI since 1988. He has made heavy use of infrared spec-
troscopy to support the Getty Museum conservation labora-
tories for analysis of artwork samples during conservations.
He is also a member of GCI project teams studying new tech-
niques for the surface cleaning of art objects and architecture,
including gels methods and laser cleaning.
Michael Schilling is a senior scientist at the Getty Conservation
Institute in Los Angeles. A native Californian, he received both
his B.A. and M.S. in chemistry at the California State
Table 3. TWA daily exposure to sulfides inside passe-partout for
artwork containing paints on cellulose acetate (cels) vs. drawings
on paper.
Artwork
material Artwork
Sulfides, TWA exposure
(ppb*days)
Cel Boat Builders 80
Cel Snow White 100
Cel Lady and the Tramp,
2015
80
Cel Lady and the Tramp,
1955
80
Paper Jungle Book, pen 150
Paper Snow White, graphite 120
Paper Lady and the Tramp,
Conte
150
Note: Data taken at the end of the Beijing exhibition (77-day exposure).
12 M. K. LAGASSE ET AL.
Polytechnic University in Pomona. He has been with the Getty
since 1983, where his work has included performing color
measurements in the tomb of Nefertari prior to the tomb’s
conservation and similar work on the Mogao and Yungang
Grottoes in China. He also has conducted research on volatile
organic compounds emitted from building materials used in
museum display and storage and utilized gas chromatography
and mass spectrometry of organic binding media as part of the
Institute’s research on binding media.
Kenneth S. Suslick is the Marvin T. Schmidt Research pro-
fessor at the University of Illinois at Urbana-Champaign. He
received his B.S. from the California Institute of Technology
and his Ph.D. from Stanford University in 1978, both in chem-
istry, and has been at the University of Illinois for 40 years. In
2018–2019, he will be the George Eastman Professor at the
University of Oxford and Balliol College. He has supervised
75 Ph.D. students, edited four books and published more
than 400 scientific papers. In addition, he studied sculpture
(bronze casting especially) under Professor Roger F. Blakely.
ORCID
Zheng Li http://orcid.org/0000-0001-9066-5791
Kenneth S. Suslick http://orcid.org/0000-0001-5422-0701
References
Askim, J. R., Z. Li, M. K. LaGasse, J. M. Rankin, and K. S.
Suslick. 2016.“An Optoelectronic Nose for Identification
of Explosives.”Chemical Science 7: 199–206.
Askim, J. R., M. Mahmoudi, and K. S. Suslick. 2013.“Optical
Sensor Arrays for Chemical Sensing: The Optoelectronic
Nose.”Chemical Society Reviews 42: 8649–8682.
Askim, J. R., and K. S. Suslick. 2015.“Hand-held Reader for
Colorimetric Sensor Arrays.”Analytical Chemistry 87:
7810–7816.
Bamberger, J. A. 2012.“Comments on an Improved Oddy Test
Using Metal Films.”Studies in Conservation 57: 187–188.
Bamberger, J. A., E. G. Howe, and G. Wheeler. 1999.“A
Variant Oddy Test Procedure for Evaluating Materials
Used in Storage and Display Cases.”Studies in
Conservation 44: 86–90.
Blades, N., T. Oreszczyn, M. Cassar, and W. Bordass. 2000.
Guidelines on Pollution Control in Museum Buildings.
London: Museums Association.
Brimblecombe, P. 1994.“The Balance of Environmental
Factors Attacking Artifacts.”In Durability and Change:
The Science, Responsibility and Cost of Sustaining Cultural
Heritage, edited by W. Krumbein, P. Brimblecombe, D. E.
Cosgrove, and S. Staniforth, 67–79. Chichester: Wiley.
Camuffo, D., R. Van Grieken, H.-J. Busse, G. Sturaro, A.
Valentino, A. Bernardi, N. Blades, et al. 2001.
“Environmental Monitoring in Four European Museums.”
Atmospheric Environment 35 (Supp 1): 127–140.
Carey, J. R., K. S. Suslick, K. I. Hulkower, J. A. Imlay, K. R.
Imlay, C. K. Ingison, J. B. Ponder, A. Sen, and A. E.
Wittrig. 2011.“Rapid Identification of Bacteria with a
Disposable Colorimetric Sensing Array.”Journal of the
American Chemical Society 133: 7571–7576.
Draeger Co. 2016.“Draeger-tubes and accuro pump.”https://
www.draeger.com/Products/Content/tubes-accuro-pump-
pi-us.pdf [Online Early Access]. Published Online: 2016.
Feng, N. 2016.“Overview of Preventive Conservation and the
Museum Environment in China.”Studies in Conservation
61: 18–22.
Feng, L., C. J. Musto, J. W. Kemling, S. H. Lim, and K. S.
Suslick. 2010a.“A Colorimetric Sensor Array for
Identification of Toxic Gases Below Permissible Exposure
Limits.”Chemical Communications 46: 2037–2039.
Feng, L., C. J. Musto, J. W. Kemling, S. H. Lim, W. Zhong, and
K. S. Suslick. 2010b.“Colorimetric Sensor Array for
Determination and Identification of Toxic Industrial
Chemicals.”Analytical Chemistry 82: 9433–9440.
Ferdyn-Grygierek, J. 2014.“Indoor Environment Quality in
the Museum Building and its Effect on Heating and
Cooling Demand.”Energy and Buildings 85: 32–44.
Grzywacz, C. M. 2006.Monitoring for Gaseous Pollutants in
Museum Environments. Los Angeles: Getty Publications.
Grzywacz, C. M., and D. C. Stulik. 1992.“In Objects Specialty
Group Postprints 1991, Albuquerque, New Mexico.”
American Institute for Conservation of Historic and
Artistic Works 1: 33–41.
Hatchfield, P. B. 2002.Pollutants in the Museum Environment:
Practical Strategies for Problem Solving, Exhibition and
Storage. London: Archetype Publications.
Hisham, M. W. M., and D. Grosjean. 1991.“Air Pollution in
Southern California Museums: Indoor and Outdoor Levels
of Nitrogen Dioxide, Peroxyacetyl Nitrate, Nitric Acid,
and Chlorinated Hydrocarbons.”Environmental Science &
Technology 25: 857–862.
Janzen, M. C., J. B. Ponder, D. P. Bailey, C. K. Ingison, and K. S.
Suslick. 2006.“Colorimetric Sensor Arrays for Volatile
Organic Compounds.”Analytical Chemistry 78: 3591–3600.
LaGasse, M. K. 2016.Ph.D. Thesis, colorimetric sensor arrays:
Development and application to art conservation https://
www.Ideals.Illinois.Edu/handle/2142/93062; University of
Illinois at Urbana-Champaign: Urbana, IL, 2016.
Li, Z., and K. S. Suslick. 2016.“Portable Optoelectronic Nose
for Monitoring Meat Freshness.”ACS Sensors 1: 1330–1335.
Lim, S. H., L. Feng, J. W. Kemling, C. J. Musto, and K. S.
Suslick. 2009.“An Optoelectronic Nose for the Detection
of Toxic Gases.”Nature Chemistry 1: 562–567.
Lin, H. W., M. Jang, and K. S. Suslick. 2011.“Preoxidation for
Colorimetric Sensor Array Detection of Vocs.”Journal of
the American Chemical Society 133: 16786–16789.
Lucchi, E. 2018.“Review of Preventive Conservation in Museum
Buildings.”Journal of Cultural Heritage 29: 180–193.
Marć, M., M. Tobiszewski, B. Zabiegała, M. de la Guardia, and
J. Namieśnik. 2015.“Current air Quality Analytics and
Monitoring: A Review.”Analytica Chimica Acta 853:
116–126.
Musto, C. J., S. H. Lim, and K. S. Suslick. 2009.“Colorimetric
Detection and Identification of Natural and Artificial
Sweeteners.”Analytical Chemistry 81: 6526–6533.
Oddy, W. A. 1973.“An Unsuspected Danger in Display.”
Museum Journal 73: 27–28.
Pretzel, B., and N. Shibayama. 2003.“Standard Materials
for Corrosiveness Testing.”V&A Conservation Journal 43:
7–9.
Rakow, N. A., and K. S. Suslick. 2000.“A Colorimetric Sensor
Array for Odour Visualization.”Nature 406: 710–713.
JOURNAL OF THE AMERICAN INSTITUTE FOR CONSERVATION 13
Schäfer, M., and E. Roffael. 2000.“On the Formaldehyde
Release of Wood.”European Journal of Wood and Wood
Products 58: 259–264.
Schieweck, A., and M.-C. Bock. 2015.“Emissions From low-
voc and Zero-voc Paints–Valuable Alternatives to
Conventional Formulations Also for use in Sensitive
Environments?”Building and Environment 85: 243–252.
Smielowska, M., M. Marc, and B. Zabiegala. 2017.“Indoor air
Quality in Public Utility Environments-a Review.”
Environmental Science and Pollution Research 24: 11166–
11176.
Suslick, B. A., L. Feng, and K. S. Suslick. 2010.“Discrimination
of Complex Mixtures by a Colorimetric Sensor Array:
Coffee Aromas.”Analytical Chemistry 82: 2067–2073.
Thomson, G. 1965.“Air Pollution –A Review for Conservation
Chemists.”Studies in Conservation 10: 147–167.
Thomson, G. 1986.The Museum Environment. 2nd ed.
Oxford: Butterworth-Heinemann.
Wang, S., L. D. Kong, Z. S. An, J. M. Chen, L. M. Wu, and X. G.
Zhou. 2011.“An Improved Oddy Test Using Metal Films.”
Studies in Conservation 56: 138–153.
Wang, S., L. D. Kong, Z. S. An, J. M. Chen, L. M. Wu, and X. G.
Zhou. 2012.“Comments on an Improved Oddy Test Using
Metal Films Response.”Studies in Conservation 57: 188–188.
Whitmore, P. M. 2005. Conservation science research:
Activities, needs, and funding opportunities; a report to
the national science foundation, 2005.
Zhang, Y., J. R. Askim, W. Zhong, P. Orlean, and K. S. Suslick.
2014.“Identification of Pathogenic Fungi with an
Optoelectronic Nose.”The Analyst 139: 1922–1928.
Zhang, C., D. P. Bailey, and K. S. Suslick. 2006.“Colorimetric
Sensor Arrays for the Analysis of Beers: A Feasibility Study.”
Journal of Agricultural and Food Chemistry 54: 4925–4931.
Zhang, C., and K. S. Suslick. 2007.“Colorimetric Sensor Array
for Soft Drink Analysis.”Journal of Agricultural and Food
Chemistry 55: 237–242.
14 M. K. LAGASSE ET AL.
