NEW SYNTHETIC RESINS FOR PICTURE VARNISHES

E. Rene de la Rie and Christopher W. McGlinchey

ABSTRACT surface, resulting in higher gloss and more saturated colors, than
Several industrially manufactured low molecular weight (LMW) syn- high viscosity coatings [2]. Subtle variation of surface roughness,
thetic resins have been investigated for their potential use in picture
and therefore of appearance, is achieved by manipulation of the
varnishes. The resins were examined by infrared (lR) spectrometry, gel
permeation chromatography (GPC) and differential scanning calorimetry varnish surface during brush and spray application or by the
(DSC). Refractive indices (RI) were also determined. The stability of addition of matting agents. Some handling properties of natural
films of the resins was tested during accelerated aging in a xenon arc resin varnishes are not yet completely understood but are
fadeometer and evaluated by means of ultraviolet-visible (UV -vis) and
obviously related to their complex chemical composition, e.g.
IR spectrometry and solubility measurements. The effects of the hindered
amine light stabilizer (HALS) Tinuvin 292 on the aging of these resins the presence of monomeric as well as polymeric material.
and the ketone resin Laropal K80, the reduced ketone resin MS2A, and Refractive index (RI) controls the amount of light reflected at
a chemically modified ketone resin (ketone-ester) were determined. Most the paint-varnish interface. Ideally, a varnish layer should have
of the new products investigated are considerably more stable than both a refractive index equal to that of the paint medium in order to
natural and ketone resins and would, when stabilized with HALS, show
minimize reflection at the interface. Paint media, such as aut-
considerable longevity even in an environment containing UV light.
MS2A and the ketone-ester can also be stabilized effectively by the addi- oxidized drying oil, may have a fairly high RI [4], thus the need
tion of HALS. Paintings conservators have tested the working and for a varnish with a relatively high RI. The exact range of accept-
aesthetic properties of two of the new products and their opinions are able RIs is not yet known.
summarized.
Resins for picture varnishes should also be of low polarity and
free of functional groups that render them unstable. Resins of
1 INTRODUCTION low polarity dissolve in solvents of low polarity, i.e. aliphatic
Despite extensive research on picture varnishes [1, 2 and hydrocarbon solvents or hydrocarbon solvents of low aroma-
references therein], and the recent finding that dammar varnish ticity. Such solvents are generally least damaging to paintings.
can be dramatically stabilized in the absence of ultraviolet (UV) Oil paints swell most in solvents of intermediate polarity, such
light by a hindered amine light stabilizer (HALS) [3, and the as chlorinated solvents, and are leached particularly by polar
authors' other paper in this volume], there is still a great need solvents such as alcohols [1]. The absence of photochemically
for a synthetic varnish for paintings that is stable in the presence active functional groups in the resin molecules, such as ketone
ofUV light. All synthetic varnishes previously suggested as alter- groups and carbon-carbon double bonds, insures resistance to
natives to natural resin varnishes are based on polymers which autoxidative degradation [7-9].
produce a different appearance from traditional varnishes. Both Dammar resin and gum mastic, the natural resins most com-
theoretical considerations and the experience of conservators monly used in picture varnishes, have monomeric fractions of
[2, 4-6] indicate that synthetic picture varnishes have optical molecular weights 400-500. Their RIs are 1.54. The only syn-
effects comparable to those of natural resin varnishes when they thetic materials with similar properties presently used by conser-
contain resins of low molecular weight. A relatively high refrac- vators are the ketone resin Laropal K80 (BASF) and the reduced
tive index (RI) may also be required. ketone resin MS2A (Laporte) (Table 1). Ketone resins are not
Molecular weight controls the intrinsic viscosity of resins satisfactory because they are highly unstable, primarily as a result
which in turn affects the roughness of surfaces obtained when of the presence of ketone groups [9]. In MS2A, these ketone
solutions of the resins are applied as varnishes over micro- groups have been chemically reduced to hydroxyl groups, which
scopically rough paint surfaces. Coatings of low viscosity - makes this resin considerably more stable than Laropal K80;
those containing low molecular weight (LMW) resins - produce however, hydrogen bonding causes MS2A to be very brittle [9].
smoother surfaces and therefore reduce light scattering at the The synthesis and testing of a chemically modified ketone resin

Table 1 Physical properties of selected low molecular weight resins

- - f
Resin Supplier Type8 RIb Soft.pl. C Tg(OC) Mnd Mwe Mn/Mw
Dammar resin Kremer natural 1.539 39.3 488 1361 2.79
Gum mastic Kremer natural 1.536 34.7 460 1929 4.20
Laropal K80 BASF ketone 1.529 75-85 50.8 531 734 1.38
MS2A Laporte reduced ket. 1.518 54.1 620 1144 1.85
Ketone esterg ester. red. keto 1.515 (34)
Aldehyde resinh BASF aldehyde 1.494 51.3 565 723 1.28
Arkon M90 Arakawa hydrocarbon 1.542 90
Arkon P70 Arakawa hydrog.hydr. 70 15.3 694 1073 1.55
Arkon P90 Arakawa hydrog.hydr. 1.522 90 35.6 620 934 1.51
Arkon PlOO Arakawa hydrog.hydr. 100 37.9
Arkon PI25 Arakawa hydrog.hydr. 125 67.3 778 1305 1.68
Escorez 5300 Exxon hydrog.hydr. 1.546 99-110 48.6 391 652 1.67
Escorez 5380 Exxon hydrog.hydr. 1.548 79-100 31.5 349 518 1.48
Regalrez 1078 Hercules hydrog. hydr. 1.510 78-81 29.7 598 699 1.17
Regalrez 1094 Hercules hydrog.hydr. 1.519 90-98 43.8
8 hydrog.hydr. = hydrogenated hydrocarbon; reduced ket. = reduced ketone; ester.red.ket = partially esterified reduced ketone
b determined at 20°C
C softening point (OC), from manufacturer's literature
d number-average molecular weight
e weight-average molecular weight
f polydispersity
g experimental product [9, 10]; T g varies with degree of esterification
h experimental product, not commercially available
 F\
// ..... '\, •.. ,,:/"\"'[,)'
I "'f 1 \
\/.:\. '/'\ ",.,.--~
.J ,," .•

./
( I
i./
: i! '" '\ II ::
(\
I \ "

i " ,.! \ iI It!',' if

iI, ,.,',
\ III
1\"
I'
.' t

i
!
:
f
I
I:
1/ I
I \'..

\'1 \ \1 i!
i.:!
.\lUI
\ "" ~\
! i/ \ 1'1 \

M~\~
./ '~,.. \ v

,;
.
!
1.1
1/', t ~ iii

/ /; r'\ \\
I : :: 0
z:
c:
~~
~00
/i / I' \\ \ \),\ z:
I
'I'
c: V
i /, \ ~\ r~ ~
lk:

i
~ .I
./.1
i
it
/ :
V\
\\
i\ f \
~\ ! I
! ,.! " i \1: I , ~ ~
0/ .I /! \" \\ i ~
1 ../ ./ ,/,/ \'\ \ \
I

nIN=8.88 T

I /
,I .'
....
,,'
I,'
\'
\
::
\\ , 4888 3588 3888 2588 2998 UAVENunBER 1588 <cn-l ) 1888

w..... '\~<~T~~-<~' f nAX-lee. ee 1

r
~"~._/~ ./ .. ~.~"~",/~~~:~~""". mol. cuI.

\~
--.- ~"--- Ii' ~--
10000 3000 1000 300 150
"\
Fig. 1 Gel permeation chromatograms of Regalrez 1078 (I), aldehyde !

\['
resin (2), Laropal K80 (3), Arkon P90 (4) and MS2A (5).
I
was recently described [9,10]. Using Laropal K80 as the starting
~~
I~
I
I
material, ketone groups were reduced with sodium borohydride I I
and hydroxyl groups were then partially esterified with acetic or ~~
I1
II
propionic anhydride. The modification rendered the product,
r~ I~
which in this paper will be referred to as ketone-ester, more stable I
r
and less brittle. Working and aesthetic properties of ketone-ester
varnish were judged to be good by painting conservators of the ~ v
I
Metropolitan Museum of Art, New York. From this work, it is
I MIN=8.88 Ti
i I i i i
48&8 3598 3888 2598 2998 UAVENUMBER 1588 <cn-l ) 1888
clear that a varnish with desirable optical and handling properties
can be produced with a stable LMW resin. Although it would Fig. 2 Infrared spectra of the aldehyde resin (top) and Arkon P90
not be impossible to have the ketone-ester manufactured, it was (bottom).
felt that other industrially manufactured LMW resins should be
investigated first to determine if they are suitable for picture occur in the 3400-3100cm-1 region typical for N-H stretching of
varnishes. The investigation of such resins is the subject of the secondary amides in the solid phase [14] (Fig. 3). Secondary
present paper. acyclic amides in the solid state display an amide II band, caused
A variety of LMW synthetic resins is commercially available. by NH bending in the 1570-1515cm-1 region. This band is
These resins, like Laropal K80, are used industrially as gloss absent in the IR spectrum of the aldehyde resin, which may
modifiers in paints and/or tackifiers in adhesives [II]. Only resins indicate a cyclic amide (lactam) [14].
that were considered likely to be more stable than Laropal K80 A band of moderate intensity at 1735cm-1 was a point of con-
were examined. Many aldehyde resins, obtained from BASF, cern since it may be due to residual aldehyde groups, which are
were tested. Only one, an experimental product that is not com- very reactive. However, no aldehyde groups could be detected
mercially available, combined solubility in hydrocarbon solvents by nuclear magnetic resonance (NMR) spectrometry. Carboxylic
of low aromaticity and resistal)ce to photochemical degradation. acid groups also have to be ruled out based on the NMR spectrum
Although a similar product was later manufactured on a large (for experimental details see Appendix). The absorption could
scale by BASF, only samples prepared in their laboratory were be caused by ester or ketone groups; however, the strong and
found satisfactory. The company therefore kindly agreed to broad band near 1200cm-1 caused by C-O stretching of ester
manufacture c. 10kg in the laboratory for our project. The alde- groups is not present. Although ketone groups cannot be ruled
hyde resin referred to in the remainder of this paper is this special out, a more likely explanation for the absorption at 1735cm-1 is
product. Hydrocarbon resins, obtained from several suppliers, a five-membered ring (y) lactam (i.e. a cyclic amide). Similarly,
were also tested. All are commercially available and soluble in the absorption at 1668cm-1 could be from a six-membered ring
aliphatic hydrocarbon solvents. Only those that are hydrogenated (6) lactam [14]. No further information with regard to the
proved adequately stable. chemical structure of the product is available at this time.
This paper presents chemical and physical data for the products The amide group is capable of photo-induced a-cleavage re-
and the results of accelerated aging tests of films of resins with actions [15], although probably not with the same efficiency as
and without a stabilizing additive. For comparison, aging results carbony I groups. Our tests indicate that this particular product
of Laropal K80, MS2A, and ketone-ester films, with and without is very stable (see section 3). The resin is soluble in hydrocarbon
a stabilizing additive, are also presented. solvents of low aromaticity, such as mineral spirits, as well as
more polar solvents, such as acetone and alcohols.
2 PROPERTIES OF UN AGED RESINS
2. 1 Chemistry of aldehyde resins 2.2 Chemistry of hydrocarbon resins
Aldehyde resins are condensation products of urea and low Hydrocarbon resins are obtained by oligomerization of un-
molecular weight aldehydes [12]. The infrared (IR) spectrum of saturated C9 or C 10 petroleum fractions [11]. Some products are
a film of the experimental product tested by us shows the absorp- derived from dicyclopentadiene, a CIO-molecule, while others
tion band of the C-O stretch vibration of the amide group (amide are obtained by oligomerization of C9 isomers or pure
I) at 1668cm-l, typical of aminoplasts such as urea-formalde- a-methylstyrene (Fig. 3). These resins contain residual unsatura-
hyde and melamine-formaldehyde [13] (Fig. 2). Multiple bands tion that can be removed by hydrogenation. Hydrogenated
hydrocarbon resins are the most stable. We have tested Arkon from the other Arkon products [16] and therefore does not follow
M90, a non-hydrogenated hydrocarbon resin (Arakawa), and the the trend. The relatively polar aldehyde resin has a rather high
following hydrogenated hydrocarbon resins: Arkon P70, P90, T g. The brittleness of MS2A, caused by hydrogen bonding
Pl00 and P125 (Arakawa); Escorez 5300 and 5380 (Exxon); and among its abundant hydroxyl groups, is reflected in its high T g.
Regalrez 1078 and 1094 (Hercules). The numbers refer to soften- (The T gS for Laropal K80 and MS2A were newly determined
ing points (Table 1). with the procedure used for the other resins; values differ from
The IR spectra of hydrocarbon resins are very simple due to previously published values [9], probably because of different
the absence of functional groups (Fig. 2). Because of their non- experimental conditions.) Table 1 shows that resins with a fairly
polar character, hydrocarbon resins are soluble in aliphatic and broad range of T gS are available among the new products. It
aromatic hydrocarbon solvents but are insoluble in acetone and should therefore be possible to choose a product with intermediate
lower alcohols. The latter fact can be exploited in paintings con- hardness, i.e. not so soft that it picks up dust and dirt, and not
servation by choosing an inpainting medium soluble in acetone so brittle that it cracks easily.
or alcohol but insoluble in hydrocarbon solvents. This would RIs of the new products vary considerably and range from
allow for the removal of a hydrocarbon resin varnish without 1.494 for the aldehyde resin to 1.548 for Escorez 5380. A choice
affecting underlying retouchings. based on RI can therefore be made as well. The relatively low
RI of the aldehyde resin is of interest since several conservators
have reported good results with the resin (see following section).

2.4 Working properties

SECONDARY AMIDE
Several paintings conservators at the Metropolitan Museum of
Art, New York, the National Gallery of Art, Washington DC,
and others have experimented with solutions of the aldehyde resin
and Arkon P90 in mineral spirits. Mark Leonard at the Getty
Museum, Malibu, California, and Herbert Lank in London
reported results that were judged by a number of conservators
to be comparable to mastic and MS2A, respectively. Both conser-
vators indicated that it took time to learn how to work with the
resins and determine the appropriate concentration of the solu-
tions. Mark Leonard has worked with fairly dilute solutions
(15-20%), and applied these by both brushing and spraying. * To
improve the working properties of the varnishes, several more
experiments should be conducted. Other components may need
to be added to the varnishes to achieve certain desired rheological
(flow) properties.
None of the LMW synthetic resins has the polymeric fraction
a-METHYLSTYRENE DICYCLOPENT ADIENE found in natural resins such as dammar and mastic. On the other
hand, the molecular weight distributions are all broader than
Fig. 3 those of the monomeric fractions of the latter. The polydispersity
values for dammar and mastic are relatively high because of the
2.3 Physical properties of the aldehyde resin and hydrocarbon presence of this polymeric fraction (Table 1). Mastic contains
resins polymeric material of molecular weights up to c. 100,000 and
Table 1 lists physical properties of selected LMW resins. dammar up to c. 30,000. A GPC trace of dammar has been
Molecular weight distributions were determined using gel perme-
published previously [2]. These polymeric fractions lower the
ation chromatography (GPC); glass transition temperatures (Tg)
T g' introducing some flexibility to the varnish film, and play a
were obtained using differential scanning calorimetry (DSC).
role in the handling properties of the varnishes. In order to
Experimental details have been described elsewhere [2, 8, and
simulate this, a synthetic varnish may have to be formulated from
the authors' other paper in this volume]. RIs were determined a LMW resin, polymeric material and other additives.
at 20°C by applying films of the resins to an Abbe refractometer.
The films were dried under a heat lamp until all residual solvent 3 RESULTS OF ACCELERATED AGING AND
had disappeared. Softening points were obtained from the sup- DISCUSSION
pliers' literature. Testing under light is indicated because photochemically initiated
Figure 1 shows GPC traces of some of the resins. Number- autoxidation is the chief path of degradation of picture varnishes
average and weight-average molecular weights, Mn and Mw' [8,17]. Films of the resins were applied to quartz or glass plates
and polydispersities, Mw/Mn, were calculated from such from solutions in good quality mineral spirits (see section 4) or
chromatograms and are included in Table 1. Although all resins reagent grade toluene. The films were dried for at least a week
are of low molecular weight, it can be seen that sizable variations and subsequently aged in a xenon arc fadeometer (Atlas) in which
occur. Mn and Mw affect other properties, including softening daylight through window glass (including the UV component) is
point and T g' and therefore brittleness of the resins. Note the simulated at high intensity. U nstabilized films and, films contain-
very broad molecular weight distribution of MS2A and the ing the hindered amine light stabilizer (HALS) Tinuvin 292
narrow distribution of Regalrez 1078 in Figure 1, resulting in (Ciba-Geigy) [3, 7, and the authors' other paper in this volume]
a high and a low polydispersity value, respectively. These varia- were aged. Changes in the resins were measured using UV -vis
tions will undoubtedly affect working properties of varnishes and IR spectrometry and solubility tests, following methods
formulated with these resins. described previously [8]. Changes in the UV -vis spectra are
Within one class of LMW resins, the Tg, as well as the soften-
presented in terms of AA(300,t) or, for the more stable products,
ing point, increases when Mn and Mw go up: compare Arkon
P90 and P125, Escorez 5300 and 5380, and Regalrez 1078 and
1094. The very soft Arkon P70 differs in chemical composition * See the paper by Mark Leonard in this volume.
 SOT/SO" SOLUS

• • 'Y
1.4

•
ESCOREZ 5380 AND REGALREZ 1078

•
'Y "'S2" 385T/61.5C

•
'Y
lAROPAl KilO
1.2
~
75T/25A
•• 'Y
• KETONE EST[R
EstOREl 5JaO/UNSTA8IlIlED

•
15T/85C
1.0
ESCOR5J80/0 5¥T1N292

OC/l00T ~ 0.8
~
lOOC
~ 0.6
RECALREl 1078/UNSTA8lUlED
311/69C
5OC/SOT
0.4
ESCOR5J80/llIT ••••292

0.2
ESCOR5JIlO/2'U1N292 100C
100c/OT
REGA11078/0 5l1TIN292 100c
0 500 1000 1500 2000 2500
500 1000 1500 2000 2500
AGING TIME (HOURS)
AGING TIME (DAYS)

SOLUS.
ARl<ON "'90 1.4 - 50T/50C
(0.94 AT 102 HRS) ARKON P90 AND ALDEHYDE RESIN 42T/5BC

0.3 1.2

•
ARI<ON IoI9O/J¥T1N292

1.0 ARI<ON PIlO

0"
• •
'Y
0
~ 0.2 • "'S2" ('Y ) 0"
lJ")
0.8
~
'Y ARK P90/0 5¥TIN292
c(

•
<I KETONE ESTER ( • ) c(

<I
0.6

0.1 • 0.4
'Y
'Y
• •
• WS2A/0.5~N292 g 0.2
AlDEHYDE RESIN 13T/87C

500 1000
. T R .

1500
'V

2000 2500
•
ARl<.P90/111TIN292

• lOOC
100C

AGING TIME (HOURS) 500 1000 1500 2000 2500

Fig. 4 Solubility and AA(300,t) of unstabilized and HALS-stabilized

films of MS2A and the ketone-ester during aging in a xenon Fig. 5 AA(250,t) of unstabilized and HALS-stabilized films of Escorez
arc fadeometer. Also shown are data for unstabilized and 5380, Regalrez 1078, Arkon P90, and the aldehyde resin during
HALS-stabilized films of Laropal K80 (solubility) and Arkon aging in a xenon arc fadeometer. Solubilities in the right-hand
M90 (AA(300,t» (5OC/50T = 50% cyclohexane/50% toluene, column correspond to the last data-point shown for each resin
50T/50A = 50% toluene/50% acetone, etc.). (5OC/50T = 50% cyclohexane/50% toluene, etc.)

in terms of AA(250,t), since they show only small changes in 100% cyclohexane for that period. Thereafter a relatively sud-
at 3OOnm. AA(250,t) is defined as AA(250,t) = A(250,t) den change occurs. Even after they have degraded completely,
- A(250,0), in which A(250,t) is the absorbance of the film at the solubility change is much less for these resins than for natural
250nm at time 1. Solubility tests were performed using mixtures and ketone resins. Some insoluble matter has been detected in
of cyclohexane and toluene or acetone and toluene. unstabilized hydrocarbon films after c. 2,000 hours of aging in
UV and solubility data were recorded for several of the new the fadeometer; however, the films could still be removed easily
products, Laropal K80, MS2A, and the experimental ketone- with mixtures not stronger than 50% toluene/50% cyclohexane.
ester, both unstabilized and containing Tinuvin 292 at concen- Most of the unstabilized hydrogenated hydrocarbon resins and
tration levels of 0.5 %, 1%, 2 % and 3 % (% of weight of resin). those containing 0.5 % HALS fail after c. 2,000 hours in the
Only data for the lowest effective concentrations of HALS are fadeometer. The aldehyde resin and the Regalrez products are
shown (Figs. 4 and 5). Figure 4 shows that MS2A and the stabilized by 0.5 % HALS for a longer period of time. Failure
experimental ketone-ester can be stabilized effectively with consists primarily of embrittlement resulting in cracks. These
Tinuvin 292. As for some of the new products, 0.5% of the cause high scattering factors, S [8], that are determined by UV -vis
additive provides for stabilization up to at least 1,800 hours; for spectrometry (not shown here).
long-term stability, the concentration may have to be increased Increase in absorbance in the 250nm region may be due to the
to 2 %. Data for unstabilized MS2A and the unstabilized ketone- n->n* electronic transition of carbonyl groups or, more likely,
ester vary considerably, probably due to batch variations. Plots the n->n* transition of a,J3-unsaturated carbonyls (enones), that
were obtained using regression analysis of the data. Although absorb strongly in this region [14] (see also the results for IR
Laropal K80 can be stabilized to some extent with Tinuvin 292 spectrometry below). Measurement of AA(250,t) reveals con-
at 3 % (Fig. 4), it is not a recommended product because of its siderable differences in stability of the new products, although
inherent instability. Figure 4 also shows the extreme instability all are relatively stable (Fig. 5). The aldehyde resin and the
of the non-hydrogenated hydrocarbon resin Arkon M90 - even Regalrez products show greater stability than Arkon P90 and the
at 3 % Tinuvin 292, no satisfactory degree of stabilization is Escorez products. An induction period of at least 170 hours
obtained. This hydrocarbon resin was monitored at 300nm rather appears in the plot of AA(250,t) for unstabilized Escorez 5380,
than 250nm. probably caused by the presence of the antioxidant butylated
UV and solubility data for some of the hydrogenated hydro- hydroxy toluene (BHT) that is added for stabilization of the
carbon resins and the aldehyde resin appear in Figure 5. Solubility product during thermal processing [18]. Measurement of both
changes are not as straightforward as for less stable products, solubility and AA(250,t) clearly demonstrates the effect of the
such as natural and ketone resins. Natural resins, ketone resin, addition of the HALS Tinuvin 292. The Regalrez products and
MS2A and the experimental ketone-ester show measurable the aldehyde resin are stabilized for at least 3, 100 hours by 0.5 %
changes from the moment aging starts, i.e. there is no induction of the HALS. Other products require 1%-2 % of the stabilizer.
period. Polarity of the resins then changes gradually until c. 50% IR spectra were recorded during the aging for some hydro-
acetone/50% toluene is needed to dissolve the aged resin. Induc- carbon resins and the aldehyde resin, both unstabilized and con-
tion periods of at least 800 hours with respect to change in taining 2 % Tinuvin 292. Evaluation of IR spectra confirms the
solubility occur for all unstabilized hydrogenated hydrocarbon rapid degradation of the non-hydrogenated hydrocarbon resin
resins and the unstabilized aldehyde resin, which remain soluble Arkon M90. Changes in the IR spectra of the aldehyde resin and
 the hydrogenated hydrocarbon resins occur at a considerably taining Tinuvin 292, which discolor upon standing. (No such
lower rate than changes in IR spectra of natural and ketone resins discoloration is caused by the additive in solid films.)
under the same conditions. The IR spectrum of HALS-stabilized Good quality mineral spirits (m.s.) should be used because they
aldehyde resin after 2,200 hours of fadeometer aging shows only do not affect the stability of the resins. We have had good
minor changes compared to the spectrum of the unaged resin. experience with some of the Shell products, such as m.s. 135
Autoxidation initially causes increased absorption in the and m.s. 150 EC. These contain respectively 15% and 3.1 %
1700cm-1 region of IR spectra due to the formation of carbonyl aromatics according to Shell's latest data (Table 2). M.s. 150 EC
groups and in a later stage probably also due to carboxylic acid will dissolve the hydrocarbon resins easily, although they should
groups. This change can be followed particularly well for also dissovle in m.s. 200 HT. The experimental aldehyde resin
hydrocarbon resins since no absorption occurs in this region in dissolves easily in m.s. 135 but will also dissolve in a hydro-
the unaged products. A 'carbonyl value', AA(1700,t), obtained carbon solvent that is lower in aromatics. (M.s. 145 EC is pro-
by subtracting the absorption at 2,004cm1 from that at bably strong enough; the aldehyde resin is presently not freely
1,712cm-1 of spectra that have been normalized at the C-H available.) Hydrocarbon solvents with lower and higher
stretching frequency of 2918cm-1, was plotted against aging evaporating rates than the products mentioned are available and
time for some of the hydrocarbon resins (Fig. 6). The 'carbonyl may be tried for specific needs.
value' indicates an induction period of at least 70 hours for the It should be kept in mind that all commercial products men-
Arkon products and at least 158 hours for the Regalrez and tioned in this paper are subject to changes in formulation, as may
Escorez products. Figure 6 shows plots of 'carbonyl values' for be apparent for the solvents from comparing Table 2 with
films containing 2 % Tinuvin 292 up to 1,027 hours of fadeometer previously published data [19]. The latest information from the
aging. Also included are values obtained from IR spectra of supplier should therefore always be consulted.
Arkon P90, Escorez 5300, and Regalrez 1078 containing 2%
HALS after 4,494 hours of aging. These long-term aging results Table 2 Properties of some Shell mineral spirits
show the exceptional stability of the Regalrez products. Distillation Seconds to Total
range (OC) 90% evapor. aromatics (%)

""7"
03 )1--) --
(t I m.s. 135 164 - 202 4660 15.0
m.s. 145 EC 162 - 201 3250 7.1
0'
••,,"" .00
I m.s. 150 EC 162 - 200 3415 3.1
o
,......
I
m.s. 200 HT 162 - 206 342'. <0.1
:;
~
0.2 •
I I

i..J ARKON t.l90/2"T1N292 0

::>
«
.oJ
,/ I

>
• REGALREZ 1078 ,
5 CONCLUSIONS
.oJ I
t:. ~
o
~ 0.1 ESCOREZ
AR_ P90/2"T1N292
The hydrogenated hydrocarbon resins and the experimental
(I) ~JOO
a::
« aldehyde resin described in this paper are considerably more
~ o
stable than the LMW resins currently in use by paintings conser-
o U .~
~~ _(1"

...'-~! - I
A=================t:.
c::J I I I~ t ~. ~~ REG 1078/2"TIN292
vators, such as natural and ketone resins. When formulated with
o 200 400 600 800 1000 4500 2 % of the HALS Tinuvin 292, the products are all class 'A'
AGING TIME (HOURS)
materials according to the classification proposed by Feller, even
in the presence of UV light. The non-hydrogenated hydrocarbon
Fig. 6 'Carbonyl value', ~A(l700,t) of unstabilized and HALS-
resin Arkon M90 was found to be very unstable. MS2A and the
stabilized films of Arkon M90, Arkon P90, Regalrez 1078, and
Escorez 5300 during aging in a xenon arc fadeometer. In addition ketone-ester (unavailable) can also be stabilized effectively by
to plots up to 1,027 hours, data-points at 4,494 hours for some addition of 2 % Tinuvin 292.
stabilized films are shown. Working properties of solutions of the new products in mineral
spirits are such that optical effects equivalent to those of natural
UV, IR and solubility data indicate that the aldehyde resin and resin varnishes can be obtained. In order for the working proper-
the Regalrez products are the most stable of the investigated ties of a synthetic varnish to replicate those of natural resin
products. However, all hydrogenated hydrocarbon resins'and the varnishes more closely, polymeric resins and rheological modi-
aldehyde resin form very stable films when formulated with 2 % fiers may have to be added to solutions of LMW resins. The
of the HALS Tinuvin 292. UV light was not filtered out in any beneficial influence of the addition of polymeric material on the
of these experiments except to the extent obtained by normal flexibility of varnish films should be further investigated. If, after
window glass. These new products, therefore, are a major further experimentation, the experimental aldehyde resin appears
improvement over natural and ketone resins, both of which to have major advantages over commercially available hydro-
deteriorate substantially under such conditions in only a few genated hydrocarbon resins, attempts will be made to have it
hundred hours. HALS-stabilized films of hydrogenated hydro- synthesized for the conservation field. The new products dis-
carbon resins and the aldehyde resin are class 'A' materials cussed in this paper may have additional applications in the con-
following the classification proposed by Feller. * HALS- servation field, such as in adhesives, consolidants, inpainting
stabilized MS2A and the HALS-stabilized ketone-ester are also media, furniture finishes and other coatings.
in this category.
ACKNOWLEDGEMENTS
The authors are indebted to John M. Brealey, Sherman F~irchild Founda-
4 SOME PRACTICAL SUGGESTIONS tion Chairman of Paintings Conservation at the Metropolitan Museum
Solutions of resins should always be prepared shortly before of Art, New York, for making this work possible. E.R. de la Rie is
application as degradation occurs much faster in solution than grateful to the National Gallery of Art, Washington DC, for supporting
in the solid state. This is particularly important for solutions con- the final phase of this work. The ketone-ester was synthesized by Alex-
ander M. Shedrinsky; Marycolette Hruskocy prepared samples for aging
and carried out many of the measurements. The help and generosity of
* See the paper on "The effect of a hindered amine light stabilizer on the aging BASF Aktiengesellschaft in synthesizing the aldehyde resin is greatly
of dammar and mastic varnish in an environment free of ultraviolet light' by appreciated. Arakawa Chemical Co. are thanked for providing technical
E. Rene de la Rie and Christopher W. McGlinchey in this volume, and information about their products. Barbara Berrie, Janice Gruver and
references therein. Sarah Bertalan made many valuable comments on the manuscript.
APPENDIX: EXPERIMENTAL METHODS resins and the synthesis of a derivative with increased stability and
An NMR spectrum for the aldehyde resin was obtained by Suzanne flexibility', Studies in Conservation 34 (1989) 9-19.
Quillen Lomax of the National Gallery of Art at the US Naval Academy 10 Shedrinsky, A.M., PhD dissertation, Chapter 3, New York Univer-
in Annapolis, Maryland. The spectrum was measured on a Varian sity (1986).
EM360L instrument using CDCl3 as solvent. Chemical shifts were II McGlinchey, C. W., 'The industrial use and development of low
measured relative to tetramethylsilane, which was added for internal molecular weight resins: an examination of new products of interest
calibration. No peaks were observed in the regions characteristic for to the conservation field' in ICOM Committee for Conservation, 9th
aldehyde groups (9-1Oppm) and carboxylic acid groups (11-12ppm). Triennial Meeting, Dresden (1990).
Although the spectrum was consistent with a secondary amide, no other
12 BASF Aktiengesellschaft, private communication.
pertinent information could be obtained from it.
13 The Infrared Atlas of Monomers and Polymers, Sadtler Research
SUPPLIERS' Laboratories, Philadelphia (1980) 457.
Arakawa Chemical (USA) Inc., Chicago, IL 60611, USA. 14 Silverstein, R.M., Bassler, G.C., and Morrill, T.C., Spectrometric
Atlas Electric Devices Co., Chicago, IL 60613, USA. Identification of Organic Compounds, 4th edn, John Wiley & Sons,
BASF Aktiengesellschaft, D-6700 Ludwigshafen, West Germany. New York (1981).
Ciba-Geigy Corp., Ardsley, NY 10502, USA. 15 Turro, N.J., Modern Molecular Photochemistry, The Benja-
Exxon Chemical Co., Florham Park, NJ 07932, USA. min/Cummings Publ. Co., Inc., Menlo Park (1978) 540.
Hercules Inc., Wilmington, DE 19894, USA.
16 Arakawa Chemical Co., private communication.
Dr G. Kremer, Pigmente, Malmittel, D-7971 Aichstetten im Algau, West
Germany. 17 Thomson, G., 'New picture varnishes' in Recent Advances in Conser-
Laporte Industries Ltd, Widnes, Cheshire WAS OJU, England. vation, ed. G. Thomson, Butterworths, London (1963) 176-184.
Shell Chemical Co., Houston, TX 77251, USA. 18 Exxon Chemical Co., private communication.
19 de la Rie, E.R., 'Research on picture varnishes: status of the project
REFERENCES at the Metropolitan Museum of Art' in ICOM Committee for Conser-
1 Feller, R.L., Stolow, N., and Jones, E.H., On Picture Varnishes and vation, 8th Triennial Meeting, Sydney (1987) 791-796.
their Solvents, National Gallery of Art, Washington DC (1985).
2 de la Rie, E. R., 'The influence of varnishes on the appearance of AUTHORS
paintings', Studies in Conservation 32 (1987) 1-13. E. Rene de la Rie received a doctorandus degree in physical chemistry
3 de la Rie, E.R., and McGlinchey, C.W., 'Stabilized dam mar picture (1975) and a doctorate (1988) from the University of Amsterdam. He
varnish', Studies in Conservation 34 (1989) 137-146. carried out subsidized research at the Central Research Laboratory for
Objects of Art and Science, Amsterdam (1975-78) and served as a docent
4 Feller, R.L., 'Factors affecting the appearance of picture varnish',
and administrator for the Training Program for Conservators, Ministry
Science 125 (1957) 1143-1144.
of Culture, Amsterdam (1978-81). At the Metropolitan Museum of Art
5 Thomson, G., 'Some picture varnishes', Studies in Conservation 3 in New York (1981-89), he created and supervised a scientific laboratory
(1957) 64-78. for the Paintings Conservation Department. He became head of the
6 De Witte, E., Goessens-Landrie, M., Goethals, EJ., van Lerberghe, Scientific Research Department at the National Gallery of Art,
K., and van Springel, C., 'Synthesis of an acrylic varnish with high Washington DC, in May 1989. Address: National Gallery of Art, Scien-
refractive index' in ICOM Committee for Conservation, 6th Triennial tific Research Department, Washington, DC 20565, USA.
Meeting, Ottawa (1981) 81/16/4.
Christopher W. McGlinchey received his degree in Polymer Science and
7 de la Rie, E.R., 'Polymer stabilizers. A survey with reference to
Engineering from Polytechnic University, Brooklyn, NY. He joined the
possible applications in the conservation field', Studies in Conserva-
Paintings Conservation Department of the Metropolitan Museum of Art
tion 33 (1988) 9-22.
in 1984 as a conservation science intern and has been a staff member
8 de la Rie, E.R., 'Photochemical and thermal degradation of films of since 1987. He also teaches at the NYU Conservation Center. Address:
dam mar resin', Studies in Conservation 33 (1988) 53-70. Metropolitan Museum of Art, Fifth Avenue at 82nd Street, New York,
9 de la Rie, E.R., and Shedrinsky, A.M., 'The chemistry of ketone NY 10028, USA.