See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/297411436

Relationship of vibro-mechanical properties

and microstructure of wood and varnish
interface in string instruments

Article in Applied Physics A · March 2016

DOI: 10.1007/s00339-016-9670-1

CITATIONS READS

2 309

6 authors, including:

Marjan Sedighi-Gilani Rolf Kaufmann

Nobel Biocare Management AG Empa - Swiss Federal Laboratories for Materi…
32 PUBLICATIONS 259 CITATIONS 106 PUBLICATIONS 1,251 CITATIONS

SEE PROFILE SEE PROFILE

Francis Willis Matthew Robert Schwarze

Empa - Swiss Federal Laboratories for Materi…
129 PUBLICATIONS 2,091 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

3D imaging View project

X-ray studies on polymers and composites: the combination of two-dimensional WAXS, SAXS and X-
ray imaging techniques View project

All content following this page was uploaded by Francis Willis Matthew Robert Schwarze on 24 November 2017.

The user has requested enhancement of the downloaded file.

Appl. Phys. A (2016)122:260
DOI 10.1007/s00339-016-9670-1

INVITED PAPER

Relationship of vibro-mechanical properties and microstructure

of wood and varnish interface in string instruments
Marjan Sedighi Gilani1 • Johanna Pflaum2 • Stefan Hartmann3 • Rolf Kaufmann3 •

Michael Baumgartner4 • Francis Willis Mathew Robert Schwarze1

Received: 25 January 2016 / Accepted: 26 January 2016

 Springer-Verlag Berlin Heidelberg 2016

Abstract Wood varnish coatings not only are aestheti- sound radiation along the grain, but increase it in the per-
cally important, but also preserve the musical instrument pendicular direction. Changes in the vibrational properties
from wear and fluctuations in the ambient humidity. were discussed together with results of 3D images of wood
Depending on the thickness, extent of penetration into the and varnish microstructure, obtained from a customized
wood and the physical and mechanical properties after tabletop X-ray microtomographic setup. For comparison,
hardening, varnishes may change the mechanical and also the microstructure of the interface of the varnished wood in
vibro-acoustical properties of the coated wood. Contrary to the laboratory and of specimens from two old violins was
studies on the chemistry of the varnish and primer used for analyzed with the same X-ray tomography setup. Labora-
old and contemporary musical instruments, the physical tory varnishes with various compositions penetrated dif-
and mechanical properties of the varnished wood in rela- ferently into the wood structure. One varnish of a master
tion to the geometry of their interface have been poorly grade old violin had a higher density and was also thicker
studied. We implemented non-destructive test methods, and penetrated weaker into the wood, which is more likely
i.e., vibration tests and X-ray tomography, to characterize related to a more sophisticated primer and varnish appli-
the hardening-dependent change in the vibrational proper- cation. The study demonstrates the importance of the vibro-
ties of master grade tone wood specimens after coating mechanical properties of varnish, its chemical composition,
with four different varnishes. Two were manufactured in thickness and penetration into wood.
the laboratory, and two were supplied from master violin
makers. For a controlled accelerated hardening of the
varnish, a UV exposure method was used. It was demon- 1 Introduction
strated that varnishes increase wood damping, along and
perpendicular to the grain directions. Varnishes reduce the The term varnish is generally referred to as ‘‘a liquid
which, when coated over a solid surface, dries to a trans-
parent film’’ [1]. Wooden musical instruments are covered
& Marjan Sedighi Gilani by varnish layers often made by different organic com-
[email protected] pounds. They protect the musical instruments from changes
1
in the relative humidity (RH) and direct contact with the
Applied Wood Material Laboratory, (EMPA) Swiss Federal
Laboratories for Materials Science and Technology, 129
player. Varnishes are aesthetically important for the
Uberlandstrasse, 8600 Dübendorf, Switzerland instrument, and they also influence the acoustics of the
2
Albert-Ludwigs-Universität Freiburg, Bertoldstrasse 17,
instrument by modifying the wood properties, e.g., mass,
79085 Freiburg im Breisgau, Germany stiffness and damping.
3 Wood is the preferred material for most music instru-
Center for X-ray Analytics, (EMPA) Swiss Federal
Laboratories for Materials Science and Technology, 129 ments. It is a hierarchically structured cellular composite,
Uberlandstrasse, 8600 Dübendorf, Switzerland with various anatomies and biochemical compositions
4
Atelier für Geigenbau, Totengässlein 7, 4051 Basel, depending on the selected species. The density, hygro-
Switzerland scopicity and vibro-acoustical properties vary between

123
260 Page 2 of 11 M. Sedighi Gilani et al.

wood species. At the submicron scale, e.g., cell wall sub- and mechanical properties of contemporary and old var-
layers and middle lamella, wood can be seen as a fibrous nishes and their impact on vibro-mechanics and hygro-
composite material, whereas the cellulose crystalline scopicity of the coated wood are not well studied. Our
microfibrils represent the reinforcing components in a knowledge on changes in wood properties after coating is
matrix consisting of hemicelluloses and lignin [2]. Briefly mainly based on studies by Haines [15], Ono [16, 17],
speaking, cellulose is mainly responsible for stiffness, Schleske [18, 19] and Obataya et al. [20] that examined the
while the matrix is more sensitive to humidity variations vibrational behavior of small wood strips or plates before
and contributes to the viscoelasticity and swelling of wood. and after coating. They reported changes in the stiffness
Microfibrils and matrix are linked by hydrogen bonds and and internal damping of the varnished wood and discussed
covalent cross-linkages, which enforce the hygroscopicity problems regarding the orthotropicity of wood properties,
and dimensional instability of the bulk material. Physical i.e., different stiffness and damping properties along the
and mechanical properties of wood are different along and grain (longitudinal) and perpendicular to the grain
perpendicular to the longitudinal cells, i.e., tracheids in (transversal). Considering that the varnish composition is
softwood. Moreover, Norway spruce ‘tone wood’ used for different for each violin maker, an understanding of the
soundboards of violin families has different properties, i.e., relationship between the chemistry of varnish and changes
lower density and narrower annual rings than timber used in the physico-mechanical properties of the coated wood is
for construction. Thus, it is safe to assume that treating of still a challenge. The problem is even more complex, when
the surface of this fine material, with a solid, e.g., varnish, we note that the criteria used by different violin makers and
will influence the delicate vibro-acoustic performance of a tone wood retailers for selecting master grade tone wood,
music instrument. To date, we do not exactly know how the even for the most commonly referred properties of tone
microstructure and physical properties of varnish will wood, e.g., internal damping or density, are different.
change the vibration and damping of the treated wood and The main objective of the presented study was to
influence the vibration and acoustics of string instruments. improve our understanding on the changes in vibro-me-
It is generally assumed that the varnish of old Cremona chanics of coated wood, in relation to the varnish compo-
violins have unique properties which are partly responsible sition and geometry of wood and the varnish interface. Our
for their superior sound quality. Even though experts agree hypothesis is that changes in mass, stiffness and
that the good contemporary violins might sound as good as microstructure of the interface layer influence damping,
the esteemed Cremona violins [3], the history and aes- stiffness and sound radiation of coated wood strips. Twin
thetical features of the old varnishes provides an excep- specimens of Norway spruce tone wood were excised from
tional added value to old instruments. This encourages sapwood and heartwood (the outer and inner part of the tree
further analytical studies to help understand the details of trunk), and their natural frequencies and internal damping
varnishing craftsmanship used for esteemed Cremona were measured before and after coating with different
instruments. Microscopy of semi-thin sections from his- varnishes at different hardening states. Certain simplified
torical instruments showed that the wood coating has two recipes were used for preparation of varnishes in the lab-
principal layers that consist of inorganic or organic com- oratory. Changes in the vibro-mechanical properties were
pounds: the primers, which are a transparent or lightly discussed in comparison with X-ray microtomographic
colored layer containing mineral particles and then several sections, showing the thickness of varnish and its depth of
layers of varnish. In the case of the violin, the varnish can penetration into wood. Moreover, results were compared to
be slightly colored, containing organic dyes, finely dis- the X-ray microtomographs of specimens from two old
persed of organic lakes or inorganic pigments [4]. The violins, one from a cello made in Turin, Italy, in 1640/60
vibrational properties of the varnished wood depend on and the other from a violin manufactured in Saxony,
(a) composition and thus the time-dependent physico Germany, around 1920.
(thermo-hygro)-mechanics of varnish and (b) the extent of
penetration of varnish into wood which makes the inter-
face. The chemistry of the layered structure of varnish in 2 Materials and methods
historical instruments, in terms of mineral elements and
organic compound, has been studied in in-depth in the past 2.1 Wood
[5–7]. Also, the geometry of the interface has been inves-
tigated using different microscopy techniques [8, 9] or by Twin wood plates with dimensions of 120 mm (R, radial
applying synchrotron [10, 11] methods. Nevertheless, non- direction) 9 2.5 mm (T, tangential direction) 9 420 mm
destructive X-ray imaging techniques have a larger (L, longitudinal direction) were excised from a plank of
potential for documentation and diagnosis of cultural her- Norway spruce wood (Picea abies L.), with narrow and
itage objects such as old violins [12–14]. Also, the physical even annual rings and perfectly quartered with minimal run

123
Relationship of vibro-mechanical properties and microstructure of wood and varnish interface… Page 3 of 11 260

out according to the criteria for ‘master grade’ tone wood Germany. The main ingredients of the utilized varnishes
for violin making. The tree was felled in autumn 2010, in were:
the Berguen region, Switzerland. Forty quarter-sawn strip
• Laboratory_1 (lab_1): 50 g colophony, 30 g linseed oil
specimens of 12 (R) 9 2.5 (T) 9 150 (L) mm3 were pre-
and 30 g terpentin oil targeting a 77 % resin and a
pared for characterization of the vibro-mechanical prop-
23 % oil composition
erties in the longitudinal direction and also 20 strip
• Laboratory_2 (lab_2): 112,5 g burgundy resin, 37.5 g
specimens of 120 (R) 9 12 (T) 9 2.5 (L) mm3, for
amber and 150 g linseed oil targeting a 50 % resin and
transversal direction. From the 40 longitudinal specimens
a 50 % oil composition
above, 20 were prepared from sapwood (SW_L), the
• Luthier_1 (lut_1): 750 g amber, 200 g dammar, 60 g
outermost portion of a tree trunk, which is used for making
manila copal and 50 g aloe targeting a 62 % resin and
the central part of violin top plate and 20 from the heart-
38 % oil composition
wood (HW_L). Specimens were stored and conditioned in
• Luthier_2 (lut_2): 1.5 l linseed oil, 150 g aloe, 500 g
a climate chamber at 30 % relative humidity (RH) at
amber, 26 g dragons blood, 33 g gamboge targeting a
20 C. The density at equilibrium moisture content (EMC)
60 %resin and 40 % oil composition
was 439.4 (±8.3) kg/m3 for HW_L, 444.1 (±9.5) kg/m3 for
SW_L and 439.9 (±4.7) kg/m3 for transversally excised We used a mixture of egg white, linseed oil and water as
specimens (TR). a basic primer before varnishing the wood. All specimens
In Fig. 1, the selected material illustrated at three dif- were coated first, with a similar quantity of primer, 0.16 g
ferent spatial scales, (a) macroscale: sound board, (b) me- for the longitudinal and 0.12 g for the radial specimens. A
soscale: test specimens for the vibrational tests and drop of primer solution was placed with a pipette onto the
(c) microscale: microtomography image of the cellular wood and spread over the surface with a thin spatula. After
structure of Norway spruce wood, is presented. In the lat- 5 days of drying, the same method was implemented for
ter, the orthotropic directions and some important applying one layer of varnish. Our main aim was not to
anatomical feature of Norway spruce wood are highlighted. obtain a good quality varnish in the laboratory or to apply it
professionally on the wood surface, but to keep the pro-
2.2 Varnish and accelerated hardening with UV cessing method structured, documented and repeatable for
light our analytic studies.
To speed up hardening of varnish, all varnished speci-
We selected four varnishes, two prepared in the laboratory mens were simultaneously exposed to five and two incre-
and the other two were collected from two professional mental steps of 3 h (a total of 5, 8 and 11 h) in a custom-
luthiers in Germany. The luthiers made their varnishes in made UV box. As part of the hygroscopic moisture is
their own workshops, but shared the major components of desorbed after UV exposure, a waiting time of up to 7 days
their varnished and their volume fractions. The prepared was required between measurements until the specimens
and laboratory-made varnishes were designed by Johanna reached EMC at 30 % RH. Thus, the specimens were kept
Pflaum, a qualified luthier who studied at Mittenwald in in the controlled 30 % RH chamber between each UV

Fig. 1 a Norway spruce tone wood plank, b strip-shaped wood specimens and c 3D microstructure of Norway spruce wood acquired by
synchrotron X-ray tomography

123
260 Page 4 of 11 M. Sedighi Gilani et al.

exposure, and their mass measured until gaining EMC. first vibration mode were measured using a laser triangu-
Table 1 shows the mass of varnish layer measured after lation displacement sensor. Vibration emission and acqui-
treating the wood specimens and after 5-, 8- and 11-h sition of data were performed in a custom-developed
exposure to UV light and equilibrium at 30 % RH. algorithm in labview (described previously in [21]). A
Although we intended to coat the wood with a similar broadband frequency scan was used to detect the first
quantity of varnish, the total mass of coating was different resonance frequency, on which a second, narrow-band scan
depending on the varnish type, e.g., significantly higher for allowed the determination of the quality factor Q (band-
lab_2 and reduced after UV exposure. width at half-power). Then, the excitation (fixed at fR) was
Moreover, two 1.5 9 1.5 9 1.5 mm3 wood specimens stopped and the logarithmic decrement (k) of amplitudes
were excised from old violins, for comparison of the spatial was recorded to compare the tan d in both frequency (1/Q)
geometry of varnish and the wood interface with the var- and time (k/p) domains. Given the high aspect ratio of the
nished specimens treated in the laboratory. One specimen samples, the influence of shear and rotary inertia was
was excised from the spare edge of a Cello made in ignored. Thus, the specific modulus of elasticity was cal-
1640-60, after being trimmed by luthiers to a more man- culated by using the fR in Euler–Bernouilli equation:
ageable size in recent years. The instrument had been E 48p2 l4 2
crafted in Turin, Italy, most probably by Henricus Casner, ¼ 4 2 fRn ð1Þ
q mn h
also known as Enrico (also the specimen label in the
study), and is still in an excellent condition and sounds where l is the length of the sample, h is the thickness, fRn is
wonderful. The other specimen was from a violin manu- resonance frequency of the mode n and m is a constant
factured in Markneukrichen, Saxony, Germany, made equal to 4.730 at the first vibration mode. Also, E is
around 1920. It is a violin of average quality and not worth Young’s modulus along the specimen length, and q is the
restoration and thus dedicated for destructive tests. density of the material; thus, the unit of the specific mod-
ulus of elasticity E/q is MPa m3/kg. All measurements
2.3 Vibrational test setup were taken inside a humidity- and temperature-controlled
chamber, at 30 % RH and 20 C, while specimens were at
We used a free–free resonance flexural vibration test to EMC. The primer and varnish were implemented inside the
determine the first resonance frequency (fR) and internal same climatic chamber, and measurements were repeated
friction (tan d) in beam elements [16, 20]. Non-contact immediately after varnishing, and also after specimens
forced-released vibration was enforced through an electro- reached EMC after 5-, 8- and 11-h exposure to UV light.
magnet device, in the face of a thin metallic plate of Each measurement was repeated three times, and the mean
\20 mg weight. Metallic plates were glued to one end of values calculated. Subsequently, the sound velocity, c,
each wood specimen, and displacements at anti-node of the sound radiation coefficient, R, and characteristic

Table 1 Mass of varnish layer

Hardening state Mass of varnish layer (mg)
(in mg) after treating the surface
of wood specimens, and after 5-, Wet 5 h UV 8 h UV 11 h UV
8- and 11-h exposure to UV
light HW_L
Lab_1 93.2 (±1.6) 86.9 (±1.1) 84.0 (±1.4) 82.3 (±0.9)
Lab_2 101.1 (±1.1) 93.0 (±2.0) 92.1 (±2.0) 92.0 (±2.0)
Lut_1 84.84 (±1.7) 71.6 (±1.8) 72.9 (±2.8) 70.82 (±2.9)
Lut_2 83.2 (±0.6) 70.9 (±1.1) 72.66 (±4.1) 69.0 (±0.6)
SW_L
Lab_1 97.1 (±1.1) 90.22 (±2.1) 87.1 (±2.1) 85.4 (±2.3)
Lab_2 102.7 (±1.4) 92.6 (±1.9) 91.2 (±0.9) 90.2 (±1.7)
Lut_1 87.12 (±1.9) 73.7 (±1.6) 72.6 (±2.1) 70.9 (±2.4)
Lut_2 86.1 (±1.8) 72.7 (±1.7) 71.9 (±1.6) 70.4 (±1.8)
TR
Lab_1 72.8 (±0.7) 63.7 (±1.2) 60 ( (±0.8) 57.8 (±0.9)
Lab_2 80.8 (±1.5) 67.4 (±1.7) 65.8 (±1.7) 64.2 (±1.9)
Lut_1 66.1 (±1.5) 53.2 (±2.0) 51.6 (±1.6) 48.2 (±2.0)
Lut_2 65.9 (±0.6) 50.1 (±0.4) 48.4 (±0.5) 47.0 (±0.2)

123
Relationship of vibro-mechanical properties and microstructure of wood and varnish interface… Page 5 of 11 260

impedance, z, in the longitudinal (_L) and transversal (_RT) primer and varnish. The relationship has the form of a
directions were calculated as: power law, confirming the suggested dashed-line trend by
sffiffiffiffi sffiffiffiffiffi Ono and Norimoto [16]. The average longitudinal E/q of
E c E pffiffiffiffiffiffi
c¼ ; R¼ ¼ ; z ¼ cq ¼ Eq ð2Þ heartwood was 35.89 (±0.4) MPa m3/kg, of sapwood was
q q q 3
34.0 (±0.8) and of transversal specimens 1.93 (±0.14).
According to Eq. 2, the sound velocity in heartwood was
5991 (±35) m/s, slightly higher than in the sapwood, which
2.4 X-ray microtomography was 5832 (±72). Also, sound velocity was significantly
lower in transversal direction, 1390 (±51) m/s. This is due
X-ray tomographies were performed at the center for to the curtail-like acting of the growth rings for sound
X-ray analytics at Empa [22] by a microfocus X-ray tube waves, resulting in a slower speed in the transversal
from Viscom XT9160-TXD operating at an acceleration direction. The same argument explains the higher damping
voltage of 80 kV and a tube current of 120 lA. Thin in the transversal direction of 0.0168 (±0.0006), while it
sections were extracted with a razor blade, with a was 0.0056 (±0.0001) in the heartwood along the grain and
dimension of approximately 1.5 9 1.5 9 1.5 mm3 from 0.0057 (±0.0002) in the sapwood.
the coated sapwood with different varnishes and also from In Fig. 3, the box-and-whisker plots of longitudinal E/q,
old violins. All sections were scanned at ambient RH and and in Fig. 4, the longitudinal tan d, before and after
temperature and reconstructed with the same settings to coating with different varnished are presented. Also, Fig. 5
allow a direct inter-relative comparison. The acquisitions shows that the transversal E/q and tan d. E/q in heartwood
were corrected for artifacts and reconstructed using was higher and less scattered than in the sapwood. After
Octopus reconstruction software [25]. Full-field cone- coating of wood with all the varnish types, E/q was
beam tomographic acquisitions were performed which decreased and subsequently increased as a result of varnish
limited the spatial resolution to a voxel size of hardening after UV exposure. Nevertheless, it never
2.5 9 2.5 9 2.5 lm3. In addition to 3D visualization of reached the initial value of E/q before coating, even after
the microstructure, dependency of the X-ray absorption on 11 h of UV exposure. Also, coating increased the tan d, in
the density of material enables certain analysis of the both the longitudinal and transversal directions. UV
density distribution in the wood cell walls and varnish by exposure decreased the tan d, though it never declined to
comparing the relative gray values. the low damping values measured in the material before
coating. However, a moderate increase in damping can be
beneficial for the sound of violins [23, 24], as high notes
3 Results and discussion are dampened and instruments sound mellow and warm. In
the transversal direction, E/q was increased after varnish-
Figure 2 shows the specific modulus of elasticity, E/q ing and also after UV exposure as indicated in Fig. 5. The
versus internal damping, tan d of all wood specimens, 20 origin of the dissimilar change in the E/q of longitudinal
(HW_L), 20 (SW_L) and 20 (TR) before coating with the and transversal specimens is related to differences between
the magnitude of stiffness in the longitudinal and
transversal directions. In the longitudinal direction, E/q of
0.020
raw wood varied between 34 and 36 MPa m3/kg, while
value of only 1.93 MPa m3/kg was obtained in the
HW_L SW_L TR transversal direction (Fig. 2). Coating a media with another
0.015
stiffer layer improves stiffness of the composite, which is
internal friction tan δ

more likely the case in transversally excised wood speci-

0.010
mens. Thus, we can assume that varnish is stiffer than
wood in the transversal direction, even if we did not
tan δ = 10-1.23x (E'/ρ)-0.68
directly measure stiffness of the varnish coating. The
Ono and Norimoto 1983
0.005 scattering of the measured data was lowest on raw wood
and increased after varnishing, especially varnish lab_2,
but often reduced after prolonged exposure to UV light.
0.000 Similar to tan d, E/q scattered stronger in varnished wood,
0 5 10 15 20 25 30 35 40 e.g., lab_2. However, apart from this slight difference, the
specific dynamic modulus E/ρ [MPa m3 kg-1]
box-and-whisker presentation did not indicate any other
Fig. 2 Relationships between tan d and E/q along the grain significant differences, i.e., between the changes in the
(HW_L and SW_L) and perpendicular to the growth rings (TR) properties in relation to the varnish type.

123
260 Page 6 of 11 M. Sedighi Gilani et al.

Fig. 3 Box-and-whisker plots (a) (b)

showing changes in longitudinal 38 38
E/q of a sapwood and

E/ρ [MPa m3 kg-1

E/ρ [MPa m3 kg-1
SW_L HW_L
b heartwood specimens treated 36 36
with varnish and after different
exposure times to UV light

Lab_1
34 34

32 32

30 30
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV
38 38

E/ρ [MPa m3 kg-1

E/ρ [MPa m3 kg-1

36 36
Lab_2

34 34

32 32

30 30
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV
38 38
E/ρ [MPa m3 kg-1

E/ρ [MPa m3 kg-1

36 36
Lut_1

34 34

32 32

30 30
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV
38 38
E/ρ [MPa m3 kg-1

E/ρ [MPa m3 kg-1

36 36
Lut_2

34 34

32 32

30 30
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV

Column bars in Fig. 6 show the change in the sound of a solid body is damped by sound radiation, a higher
radiation and acoustic impedance of wood, after imple- sound radiation correlates with the emission of a louder
menting the 4 different varnishes and hardening after 11-h sound. Also, the orthotropic changes in sound radiation,
UV exposure. The initial R and z data (averaged between 5 i.e., reduction in longitudinal direction and increase in the
raw specimens) are also labeled closed to each column bars transversal direction, have an important influence on the
which represent the change in these properties after resulting sound of the music instrument, as the wood
implementing the respective varnishes. Varnishing orthotropicity is reduced. One should note that in violins,
decreased the sound radiation, up to 6.5 %, along the grain. the wood carries a significant load in longitudinal, but also
In this regard, no significant difference was observed in the transversal direction, which is mechanically less
between the sapwood and heartwood. Among different robust. This fact plays a significant role in defining the
varnishes, changes were most significant after coating with arching of the violin soundboard by violin makers, for re-
lab_2 varnish, which is the one with the highest oil frac- distribution of forces in the desired directions. In this
tion. Changes in sound radiation were also higher for the regard, the stiffening impact of the varnish along the woods
varnish prepared in the laboratory, lab_1, compared to the weakest direction is likely beneficial. While varnishing has
professional makers’ varnishes. Contrary to the longitudi- a different influence on the longitudinal and transversal R,
nal specimens, varnishing increased sound radiation in the the acoustic impedance z always decreased, up to 2.7 %
transversal direction, and at a higher rate with the profes- longitudinally and 10.5 % perpendicularly. z relates to the
sional makers’ varnishes. As R describes how the vibration transmission of vibrational energy from the soundboard to

123
Relationship of vibro-mechanical properties and microstructure of wood and varnish interface… Page 7 of 11 260

Fig. 4 Box-and-whisker plots (a) (b)

showing changes in longitudinal 0.008 0.008
tan d of a sapwood and
b heartwood specimens treated 0.007 0.007
with varnish and after different
exposure times to UV light

Lab_1

tanδ
tanδ
0.006 0.006

0.005 0.005
SW_L HW_L
0.004 0.004
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV
0.008 0.008

0.007 0.007
Lab_2

tanδ
tanδ
0.006 0.006

0.005 0.005

0.004 0.004
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV
0.008 0.008

0.007 0.007
Lut_1

tanδ
tanδ

0.006 0.006

0.005 0.005

0.004 0.004
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV
0.008 0.008

0.007 0.007
Lut_2

tanδ

tanδ

0.006 0.006

0.005 0.005

0.004 0.004
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV

the air; thus, its reduction is more likely a drawback for the green to blue colors indicate material with a relatively
musical instrument. The lab_2 showed the highest reduc- lower density. In this context, all 4 varnished sections had a
tion in z. Luthiers who play their instruments before and lower density than wood cell walls. They filled the wood
after varnishing report that the instruments often produce a lumina, in a depth of a few hundred micrometers from the
‘sharper’ sound before varnishing. However, varnishing is surface. No significant differences were observed between
an inevitable step of violin making, for the beauty of the the depth of varnish penetration in earlywood and latewood
instrument and for protecting it from wear and moisture tracheids, even though they have different lumen size and
exchange with air, which in turn negatively changes the pit structure, and thus different permeability [26]. The
vibrational properties of wood in the short and long term. varnish lab_2 was different from the other varnishes, as it
In Fig. 7, X-ray tomography sections of wood speci- penetrated deeper into the wood. In the latter case, there
mens treated with 4 different varnish coatings are pre- were even lumina that were randomly filled up to 400 lm
sented. The thin-wall earlywood tracheids of Norway beneath the surface. This may partially explain the inferior
spruce wood are distinguishable with larger cell lumina properties of this varnish, regarding the impulse changes in
with a diameter of approximately 30–40 lm. The latewood the sound radiation and acoustic impedance. However,
tracheids have thicker cell walls, and their lumina were different varnishes may have different stiffness which may
smaller (8–15 lm) than the cells in the earlywood. Red to significantly influence the vibro-mechanical properties of
yellow colors correspond to the higher gray values, due to the varnished wood. The averaged profile of the gray val-
higher attenuation of X-ray in denser materials, while the ues over all sections confirms the above observation; gray

123
260 Page 8 of 11 M. Sedighi Gilani et al.

Fig. 5 Box-and-whisker plots (a) (b)

showing changes in the a E/q 2.5 0.040
and b tan d in perpendicular to

E/ρ [MPa m3 kg-1

the grain direction of wood 2.3 0.030
specimens treated with varnish

Lab_1
and after different exposure

tanδ
2.0 0.020
times to UV light
1.8 0.010

1.5 0.000
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV
2.5 0.040

E/ρ [MPa m3 kg-1

2.3 0.030
Lab_2

tanδ
2.0 0.020

1.8 0.010

1.5 0.000
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV
2.5 0.040
E/ρ [MPa m3 kg-1

2.3 0.030
Lut_1

tanδ
2.0 0.020

1.8 0.010

1.5 0.000
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV
2.5 0.040
E/ρ [MPa m3 kg-1

2.3 0.030
Lut_2

tanδ

2.0 0.020

1.8 0.010

1.5 0.000
raw 5h UV 8h UV 11h UV raw 5h UV 8h UV 11h UV

value of the lab_2 specimens is significantly different from Saxony, as makers always implement several layers of
the others, in 200 lm depth from the surface. As we varnish and not only one coating. The varnish also pene-
applied only one layer of varnish for simplification and trates weaker into the wood, which is likely due to the more
removing additional variables, e.g., discontinuities in sophisticated primer application in a real instrument com-
manually applying a multilayered varnish, the varnish on pared to the laboratory varnish, but also can be dependent
top is too small for quantification and further discussions. in the viscosity of the varnish itself. One interesting
In Fig. 8a, b, the 3D rendered geometry of the varnished observation is different densities of the varnishes, which
wood from 2 old violins and a middle section from each was significantly higher in Enrico (red–yellow colors), than
(inlet) are presented. The labeled specimen as Enrico is in the Saxony and all the contemporary varnishes. Wood
extracted from an instrument made in 1640, with a varnish density, especially in the latewood tracheids, was higher in
of highest quality, while the other varnish, from the the Enrico specimen, and the growth rings were narrower,
instrument made in 1920s, is of average quality with traces when compared to the Saxony and contemporary tone
of repair and restoration over the years. The varnish was wood. A layer of glue was observed in the middle of the
thick and cracked all over the instrument, the wood was Enrico specimen, with a density similar to the varnish and a
infected by wood worm, and the instrument was in a poor thickness of up to 80 lm. Microcracks, with an opening of
condition and cannot be restored or played anymore. In below 10 lm, were observed all around the surface of
both cases, the varnish thickness was greater in the labo- Saxony varnish, which were locally covered by another
ratory specimens, up to 40 lm in Enrico and 80 lm in thin layer with a higher density on top (Fig. 8c). Also,

123
Relationship of vibro-mechanical properties and microstructure of wood and varnish interface… Page 9 of 11 260

Fig. 6 Sound radiation (left) lab_1 lab_2 lut_1 lut_2 lab_1 lab_2 lut_1 lut_2
and acoustic impendence (right) (a) -2 -1.5
after coating wood specimens
with different varnishes and -3 -1.8
after 11-h UV exposure in
a longitudinal sapwood, -4 -2.1

Δz (%)
ΔR (%)
b longitudinal heartwood and 2569
2605
c perpendicular to the grain -5 -2.4
direction 13074 13428
-6 -2.7 2572
13144 2613
12901
SW_L SW_L
-7 -3
lab_1 lab_2 lut_1 lut_2 lab_1 lab_2 lut_1 lut_2
(b) -2 -1.5

-3 -1.8

-4 -2.1

Δz (%)
ΔR (%)

2647
-5 -2.4
13717
13654 2640
(m4/skg) 2631
-6 13479 -2.7 (kgm-2s-1)
2610
13706 HW_L HW_L
-7 -3
lab_1 lab_2 lut_1 lut_2 lab_1 lab_2 lut_1 lut_2
(c) 5 -7
TR 3156
3217
4 -8

3 3298 -9
Δz (%)
ΔR (%)

2 2989 -10
588
1 -11 620 607
630
TR
0 -12

some high density particles with dimensions of few lm wood, which is correlated with the speed of sound and also
were observed on the wood surface and trapped inside the in the internal damping in the material, after coating with 4
varnish, which are parts of inorganic powders, e.g., Pumice different varnishes (professionally made or in the labora-
or chalk, used for filling the lumina on the wood surface to tory). It was demonstrated that all varnishes, independent
avoid the penetration of varnish deep into the wood. The from their original compositions, increased the internal
assessment of the detailed chemical composition of these damping of wood, along or perpendicular to the grain. The
components and also determining their mechanical prop- specific modulus of elasticity and sound radiation were
erties for upscaling the microstructure–property relation- decreased along the grain and increased in the perpendic-
ships in wood and varnish interface through a coupled ular direction. After exposure of varnished wood speci-
numerical and experimental approach are the topics of our mens to UV light, the specific modulus of elasticity
future studies. increased and the internal damping decreased, as a result of
hardening of the varnish. X-ray microtomographs showed
no significant difference between extent of varnish pene-
4 Conclusion tration into the earlywood and latewood tracheids. With all
coatings, the varnish penetrated approximately 100 lm
Varnish coating plays an important role for the long-term into the wood, with exception of the home-made laboratory
quality of sound of wooden musical instruments, but as varnish that had the highest oil fraction (lab_2) and pene-
applied on the wood surface, impulse changes in the vibro- trated deeper into wood. The latter varnish caused a larger
mechanical behavior of wood can be recorded. We exam- scattering for all measured vibrational properties, and a
ined the change in the specific modulus of elasticity of more significant reduction in the longitudinal sound

123
260 Page 10 of 11 M. Sedighi Gilani et al.

Fig. 7 a Reconstructed sections (a) (b)

of X-ray tomograms showing
the spatial uptake of varnish median gray value
(blue) in early and latewood 50 75 100 125 150

lab_1
tracheids and b profile of gray 0
value, average of 600 sections

distance from top surface (um)

100

lab_2
200

300
lut_1

400 lab_1

lab_2
500
Lut_1
lut_2

600 Lut_2
T

(a) (b)
L glue line

R latewood
T
Saxony
Enrico

earlywood
resin ducts
T
microcrack

covered crack with another layer open crack fillers

(c)

fillers

Fig. 8 Showing 3D rendered geometry of wood and a varnish coating on top in a Enrico specimen from Turin 1640/60 and b Saxony specimen
from 1920, with a higher magnification slice inlet. c Saxony cracks and fillers

radiation. However, the depth of varnish penetration was also the mechanical properties of varnish itself are an
not the only parameter that determined changes in the important factor, which we will demonstrate in a future
vibro-mechanical properties of the varnished wood. Thus, study. The orthotropic changes in the sound radiation or

123
Relationship of vibro-mechanical properties and microstructure of wood and varnish interface… Page 11 of 11 260

speed of sound makes the continuation of this study on 2D 9. C.Y. Barlow, J. Woodhouse, JCAS 1(4), 2–9 (1989)
elements such as plates or the real curved soundboard of 10. G. Latour, J.P. Echard, B. Soulierm, I. Emond, S. Vaiedelich, M.
Elias, Appl. Opt. 48(33), 6485–6491 (2009)
the instrument important. Characterizing the orthotropic 11. N. Sodini, D. Dreossi, R. Chen, M. Fioravanti, A. Giordano, P.
change in vibro-mechanical properties of wood plates after Herrestal, L. Rigon, F. Zanini, J. Cult. Herit. 13, 44–49 (2012)
varnishing and numerical simulation of the relation 12. M.P. Morigi, F. Casali, M. Bettuzzi, R. Brancaccio, V. D’Errico,
between these changes and the vibration of the violin body Appl. Phys. A 100, 653–661 (2010)
13. P. Reischig, L. Helfen, A. Wallert, T. Baumbach, J. Dik, Appl.
are the topic of our next study. Phys. A 111, 983–995 (2013)
14. P. Dredge, S. Ives, D.L. Howard, K.M. Spiers, A. Yip, S. Ken-
Acknowledgments We would like to acknowledge the financial derdine, Appl. Phys. A (2015). doi:10.1007/s00339-015-9455-y
support of the Walter Fischli Foundation and COST Action FP1302. 15. D. Haines, Catgut Acoust. Soc. Newsl. 33, 19–23 (1980)
We also acknowledge the support of Daniel Heer and Markus Heeb 16. T. Ono, M. Norimoto, Jpn. J. Appl. Phys. 22, 611–614 (1983)
for rigorous sample preparation at EMPA wood workshop and Iris 17. T. Ono, J. Acoust. Soc. Jpn. (E) 14, 397–407 (1993)
Brémaud for making it possible to use her developed vibrational 18. M. Schleske, Catgut Acoust. Soc. J. 3(6, 2), 27–43 (1998)
testing hardware and software. 19. M. Schleske, Catgut Acoust. Soc. J. 4(5, 2), 50–64 (2002)
20. E. Obataya, T. Ono, M. Norimoto, J. Mater. Sci. (2000). doi:10.
1023/A:1004782827844
References 21. I. Brémaud, Diversité des bois utilisés ou utilisables en facture
d’instruments de musique (Diversity of woods used or usable in
musical instruments making). (Ph.D. dissertation, University of
1. R. Mayer, The Artist’s Handbook of Materials and Techniques Montpellier II, France 2006)
(Viking Press, New York, 1940) 22. Empa Center for X-ray Analytics. http://www.empa.ch/web/
2. R.B. Adusumalli, R. Raghavan, R. Ghisleni, T. Zimmermann, J. empa/center-for-x-ray-analytics. Accessed Jan 2016
Michler, Appl. Phys. A 100, 447–452 (2010) 23. V. Bucur, Acoustics of Wood (Springer, Berlin, 1987), p. 181
3. C. Fritz, J. Curtin, J. Poitevineau, H. Borsarello, I. Wollman, F.C. 24. F.W.M.R. Schwarze, M. Spycher, S. Fink, New Phytol. 179,
Tao, T. Ghasarossian, PNAS 111(20), 7224–7229 (2014) 1095–1104 (2008)
4. M. Malagodi, C. Canevari, L. Bonizzoni, A. Galli, F. Maspero, 25. J. Vlassenbroeck, B. Masschaele, V. Cnudde, M. Dierick, K.
M. Martini, Appl. Phys. A (2013). doi:10.1007/s00339-013-7792- Pieters, L. Van Hoorebeke, P. Jacobs, Octopus 8: A high per-
2 formance tomographic reconstruction package for X-ray tube and
5. J.P. Echard, B. Lavédrine, J. Cult. Herit. 9, 420–429 (2008) synchrotron micro-CT, in Advances in X-ray Tomography for
6. F. Caruso, S. Saverwyns, M. Van Bos, D. Francesca, C. Martino, Geomaterials, eds by J. Desrues, G. Viggiani, P. Bésuelle (ISTE,
A.E. Ceulemans, J. de Valck, E. Caponetti, Appl. Phys. A (2012). London, 2010). doi:10.1002/9780470612187.ch1326.
doi:10.1007/s00339-011-6729-x 26. M. Sedighi Gilani, M. Griffa, D. Mannes, E. Lehmann, J. Car-
7. T. Rovetta, C. Canevari, L. Festa, M. Licchelli, S. Prati, M. meliet, D. Derome, Int. J. Heat Mass Trans. 55, 6211–6221
Malagodi, Appl. Phys. A (2015). doi:10.1007/s00339-014-8882-5 (2012)
8. G. Latour, J.P. Echard, M. Didier, M.C. Schanne-Klein, Opt.
Express 20(22), 24623–24635 (2012)

123

View publication stats