Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 187 (2017) 1–8

Contents lists available at ScienceDirect

Spectrochimica Acta Part A: Molecular and Biomolecular

Spectroscopy
journal homepage: www.elsevier.com/locate/saa

Analytical evidences of the use of iron-gall ink as a pigment on

miniature paintings
Maurizio Aceto ⁎, Elisa Calà
Dipartimento di Scienze e Innovazione Tecnologica, Università degli Studi del Piemonte Orientale, viale T. Michel, 11, 15121 Alessandria, Italy
Centro Interdisciplinare per lo Studio e la Conservazione dei Beni Culturali (CenISCo), Università degli Studi del Piemonte Orientale, via Manzoni, 8, 13100 Vercelli, Italy

a r t i c l e i n f o a b s t r a c t

Article history: Iron-gall ink (IGI) has been used by scribes for writing since at least the 4th century CE. Another typical use of this
Received 22 March 2017 ink was for drawing: many Old Masters created beautiful sketches in brown-black hues. Despite its widespread
Received in revised form 6 June 2017 use to draw lines, it seems like IGI was hardly used for painting as well. In fact, the number of identification on
Accepted 13 June 2017
manuscripts is very low at present. This could be partially due to a lack of reliable diagnostic information. In
Available online 15 June 2017
this work we tried to better define the possibility of identifying IGI as a pigment on illuminate manuscripts, eval-
Keywords:
uating the pros and cons of three different techniques: UV–visible diffuse reflectance spectrophotometry with
Illuminated manuscripts optic fibres (FORS), Raman spectroscopy and XRF spectrometry. With concern to in situ non-invasive analysis,
FORS Raman spectroscopy has the best diagnostic power but FORS seems to provide the better compromise between
Iron-gall ink selectivity and ease of application. Moreover, new analytical evidences was given on the particular use of IGI by
Raman ancient illuminators: a non-invasive and micro-invasive diagnostic survey on Western manuscripts datable in
XRF the range 6–16th centuries was carried out showing that, apart from its widespread use as an ink for writing
and drawing, IGI was largely used as a pigment too. The large number of identification obtained allows us to
hypothesise that this pigment was used all through medieval Europe up to at least the Renaissance, where its
use is already documented in drawing.
The occurrence of IGI in miniature paintings older than 6th century or more recent than 16th century cannot be
excluded, as is its use beyond Europe; further measurements could instead widen the time range and the geo-
graphic area. Nevertheless, the present study allows shedding a new light on the use of this colourant all along
the period of medieval and Renaissance miniature painting art.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction
Iron-gall ink (thereafter termed IGI) is a well-known natural product
used since many centuries for writing [1]. It is produced by addition of
FeSO4, the so-called Roman vitriol, to an aqueous extract obtained
from galls, the typical nuts formed on plants upon infection by insects.
Other vitriols such as CuSO4 or ZnSO4 were also used. While the
Fe(II)-gallic acid complex is uncoloured, upon oxidation by air Fe(II)
turns to Fe(III) and the corresponding complex becomes insoluble and
brown-black coloured, which is the reason why it was used as ink for
writing. The reaction between Fe ions and gallnut extracts must have
been known at least since Roman times [2], because Pliny the Elder
cites it, even if with concern to a different context [3]: it was exploited
as an ante litteram chemical test to recognise the adulteration of
⁎ Corresponding author.
E-mail address: maurizio.aceto@uniupo.it (M. Aceto).
verdigris, used as a salve, with the less expensive ferrous sulphate,
which main use was instead as pigment for shoes (atramentum
sutorium).
It is not easy to say when IGI entered in use for writing, because the
recipes describing its preparation are absent in Antique and Early Mid-
dle ages until the appearance of Theophilus' treatise in the earlier 12th
century [4], after which many recipes were described. Therefore we
can only rely on diagnostic evidences to set up a chronology. At present,
the oldest occurrences of IGI were found by Moorhead et al. [5] by
means of multispectral imaging on the Codex Sinaiticus, a Biblical text
written in the middle of 4th century CE containing the earliest complete
copy of the Christian New Testament, and by Aceto et al. [6] by means of
Raman spectroscopy on the Vercelli Gospels, a Latin version of the Gos-
pels datable to the II half of 4th century.
To resume, we have analytical information evidencing the use of IGI
in texts at least from the middle of 4th century CE; this information is
certainly far from definitive and new measurements, performed on
older texts, could further date backwards the introduction of IGI. On

http://dx.doi.org/10.1016/j.saa.2017.06.017
1386-1425/© 2017 Elsevier B.V. All rights reserved.
2 M. Aceto, E. Calà / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 187 (2017) 1–8
the other side, according to the very large number of recipes for the
preparation of IGI that occurred from the 12th century onwards [7],
it is reasonable to think that this type of product was thoroughly
used by scribes all along Middle ages and Renaissance and further
on, until the introduction of synthetic inks; indeed, its diagnostic
identification on later documents has been reported in countless
instances.
As well as for writing, IGI has been widely used for drawing. Master
drawings executed with IGI are known by artists such as Leonardo da
Vinci, Rembrandt van Rijn, Guercino, Claude Lorrain, Eugene Delacroix
or Vincent van Gogh among others. In this case, however, the diagnostic
information is more limited with respect to the analysis of texts; among
the most recent works, the identification of IGI has been analytically ev-
idenced in studies on Italian Renaissance drawings [8] using imaging
techniques in combination with X-ray Fluorescence Spectrometry and
Raman Spectroscopy, on the collection of Prints and Drawings at the
British Museum [9,10], on a collection of drawings by Van Gogh [11]
and on 15th - 19th centuries drawings from the collection of Slovak Na-
tional Gallery [12].
A different matter is the use of this material as a colourant, i.e. not
only for the writing or drawing but also for painting, in which case,
being it a dispersion rather than a solution, it can be called a pigment.
The colour of IGI varies between brown and black so that it could be
suitable for a fairly wide range of hues. Despite this, its use for painting
seems unusually rare in the history of art. Among the few identification
present in the scientific literature, IGI was found on some early Medieval
Insular manuscripts kept at Trinity College Dublin [13]: the Book of Ar-
magh (ms. 52), the Codex Usserianus Primus (ms. 55), the Book of Durrow
(ms. 57), the Book of Dimma (ms. 59) and the famous Book of Kells (ms.
58) [14,15]. Other identification was reported by Aceto et al. [16] on a
12th century Italian illuminated manuscript, by Burgio et al. [17] on me-
dieval and Renaissance Italian manuscript cuttings and by Nastova et al.
[18,19] on medieval old-Slavonic manuscripts and on Byzantine and
post-Byzantine manuscripts [20]. Van Bos and Watteeuuw [21]
analysed grisaille illuminations in the Chroniques et Conquêtes de Charle-
magne, a 15th century manuscript held at Brussels Royal Library (ms.
KBR 9066), in which IGI was used not only to outline figures and details,
but also for darker shades. This low number of analytical evidences is
even more unusual if we consider that the other class of black
colourants, i.e. the carbon-based pigments (lampblack, vine black,
bone black, graphite, etc.) were used all along history both for writing
and for painting, so that many hundredths of identification have been
reported [22].
Rather than being due to a rare employment, the lack of evidences of
IGI in painting could be due to a difficulty in obtaining the diagnostic in-
formation. In fact, while most of the analyses carried out in the past on
black areas identified carbon-based pigments, it is reasonable to think
that measurements based on elemental techniques were not specific
enough in order to discriminate between IGI, carbon-based pigments
and other brown-black colourants such as bistre or sepia; the same
holds for polarised light microscopy and X-ray diffraction, two tech-
niques that were frequently used in the 80s, as IGI normally shows
amorphous character when attached to the pictorial support, as are car-
bon-based pigments.
The identification of IGI on artworks is usually carried out by means
of Raman spectroscopy [6,23]. This technique has possibly the best diag-
nostic power in the analysis of brown-black inks, as it provides a spec-
tral fingerprint for each of them, in particular when the Surface
Enhanced Raman Spectroscopy (SERS) mode is used [24,25]. Addition-
ally, Raman analysis of inks can be performed in totally non-invasive
way by means of portable systems. Similar information can be gained
by means of FT-IR spectrophotometry, although performing this analy-
sis with high efficiency requests an invasive or at least micro-invasive
approach [26].
While elemental techniques such as X-ray Fluorescence Spectrome-
try (XRF), Particle Induced X-ray Emission spectrometry (PIXE) or
Scanning Electron Microscopy - Energy Dispersive X-ray spectroscopy
(SEM-EDX) have been frequently used in the analysis of brown-black
inks (this topic was recently reviewed by Goltz [27]), the identification
they provide is not absolute and therefore they are not totally selective
in the discrimination between IGI, carbon-based inks, bistre and sepia.
In particular the content of iron is critical as this metal can be found in
all instances, due to an intentional addition (as for IGI), to a contamina-
tion from the processes used in the manufacture or to the support
(parchment or paper).
Further information can be obtained by means of chromatograph-
ic techniques: Chiavari et al. [28] were able to distinguish carbon
inks, sepia and IGI by means of Thermally assisted Hydrolysis Meth-
ylation (THN) coupled to GC-MS. It must be considered, however,
that this approach is intrinsically invasive and destructive. A similar
approach has been recently provided by Terahertz-Time Domain
Spectroscopy (THz-TDS) which, combined with ATR-FT-IR spectro-
photometry, allows discrimination of the different types of brown-
black inks [29].
When totally non-invasive methods are the only option allowed,
other techniques must be used. Together with the above cited Raman
and XRF techniques when used in portable configurations, good alterna-
tives are the use of multispectral and hyperspectral imaging [30] or of
UV–visible diffuse reflectance spectrophotometry with optic fibres
(FORS) [31,32].
In this work we firstly aim at evaluating the diagnostic power of
three different techniques, FORS, Raman spectroscopy and XRF spec-
trometry, in the selective identification of IGI on paintings, with specific
concern to illuminated manuscripts. Secondly we aim at evaluating the
effective use of IGI on miniature paintings along the history of art, veri-
fying whether the scarcity of identification corresponded to a real scarce
use or, rather, to lack of diagnostic information. Using both non-invasive
and micro-invasive techniques, a survey has been made on a large col-
lection of manuscripts covering a wide time span (6–16th centuries).
The manuscripts were mostly from Western Europe. FORS analysis
was used as a preliminary probe in order to identify the spectral features
of IGI; then Raman spectroscopy, both in conventional and in SERS
mode, was used to confirm the identification on selected samples. Final-
ly, XRF spectrometry was used to confirm the presence of iron and other
metals.
2. Materials and Methods
2.1. UV–visible Diffuse Reflectance Spectrophotometry with Optic Fibres
(FORS)
FORS analysis was performed with an Avantes (Apeldoorn, The
Netherlands) AvaSpec-ULS2048XL-USB2 model spectrophotometer
and an AvaLight-HAL-S-IND tungsten halogen light source; detector
and light source are connected with fibre optic cables to a 1.5 mm diam-
eter FCR-7UV200-2-1,5 × 100 probe which contains cables for both illu-
mination and detection; therefore incident and detecting angles were
respectively 45° and −45° from the surface normal, in order to exclude
specular reflectance. The spectral range of the detector was 200–
1160 nm; considering the range of emission of the light source, the op-
timal range of acquisition of spectra was 350–1100 nm. The best spec-
tral resolution of the system, calculated as FWHM, was 2.4 nm. Diffuse
reflectance spectra of the samples were referenced against the WS-2
reference tile, guaranteed to be reflective at 98% or more in the spectral
range investigated. The investigated area on the sample was 1 mm di-
ameter. In all measurements the distance between probe and sample
was 1 mm. To visualise the investigated area on the sample, the probe
contained a USB endoscope. The instrumental parameters were as fol-
lows: 10 ms integration time, 100 scans for a total acquisition time of
1 s for each spectrum. The system was managed by means of AvaSoft
8 software running under Windows 7.
 M. Aceto, E. Calà / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 187 (2017) 1–8
2.2. XRF Spectrometry
XRF measurements were performed with a handheld EDXRF Thermo
(Waltham, USA) NITON spectrometer XL3T-900 GOLDD model, equipped
with a Ag tube (max. 50 kV, 100 μA, 2 W), large area SDD detector, energy
resolution of about 136 eV at 5.9 keV. The analysed spot had a diameter of
3 or 8 mm and was focused by a CCD camera, with a working distance of
2 mm. The total time of analysis was 240 s and the instrument was held in
position with a moving stage allowing millimetric shifts, in order to reach
the desired probe-to-sample distance; the stage was laid on a tripod. The
obtained spectra were processed with the commercial software WinAxil,
derived by the academic software QXAS from IAEA.
2.3. Raman Spectroscopy
Raman spectra were collected with a high-resolution dispersive
Horiba (Villeneuve d'Ascq, France) LabRAM HR model spectrophotom-
eter coupled with a confocal microscope. The instrument is equipped
with a 633 nm excitation laser, two (600 and 1800 lines/mm) dispersive
gratings, an 800 mm path monochromator and a Peltier cooled CCD de-
tector. The optical arrangement gave a spectral resolution of about
2 cm−1. Spectra were taken placing samples on the microscope stage
and observing them with long working distance 20×, 50× and 80× ob-
jectives. The sampled area was identified and focused using either a
video camera or the microscope binoculars. Laser power at the sample
was initially kept low (b100 μW) by means of a series of neutral density
filters, in order to prevent any thermal degradation of the molecules,
then gradually increased up to the optimal signal-to-noise ratio. Expo-
sure time was 1–120 s according to needs. The system was managed
with LabSpec 5 software running under Windows XP.
Fig. 1. Examples of analysed areas containing IGI. Top left: ms. CLXXIX, BCVC (Biblioteca Capitolare in Vercelli, Italy), f. 73r, area labelled with a yellow circle; top right: ms. J.II.17, BNU
(Biblioteca Nazionale Universitaria in Torino, Italy), f. 156v, preparatory sketch; bottom left: ms. C, BCVC, f. 25v, area labelled with a yellow circle; bottom right: ms. F.II.10, BNU, f.
232v, preparatory sketch.
3
2.4. Preparation of SERS Substrates
For SERS analysis, a colloidal paste of silver nanoparticles was pre-
pared. Reagents and solvents (nitric acid, hydrochloric acid, methanol,
silver nitrate and sodium citrate dihydrate) were purchased from
Carlo Erba reagents (Arese, Italy); Ultra high quality (UHQ) water was
obtained by means of a Millipore (Darmstadt, Germany) Direct-q 3 sys-
tem. The preparation of Ag colloidal pastes was carried out following the
procedure described elsewhere [33] which is based on the Lee and
Meisel reduction of silver nitrate [34]. The analysis of the fragments
from the manuscript was performed directly on the sample, by pouring
on it 1 μl of Ag colloidal paste and allowing to dry before exposing the
sample to the laser beam.
3. Results and Discussions
Firstly we will discuss the potential of the three techniques cited
(FORS, Raman spectroscopy and XRF spectrometry) in the selective
identification of IGI on paintings. Then focusing the attention on illu-
minated manuscripts, we will present the results of an analytical sur-
vey we carried out on a large number of manuscripts containing
brown and black figurative details. This included simple painted
areas, borders of initials or boxes, drawings or preparatory drawings,
etc. Some examples of the analysed areas are shown in Fig. 1. We
tried to include in the survey manuscripts from the widest possible
time range, which resulted to be 6th–16th centuries. After collecting
a significant number of identification by means of FORS, we evaluat-
ed the similarity of the spectra and performed Raman and XRF anal-
ysis on selected samples, in order to obtain a fingerprint identification
that could be extended to all samples.
4 M. Aceto, E. Calà / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 187 (2017) 1–8
Fig. 2. FORS spectra of IGI, a carbon-based pigment, bistre and sepia.
3.1. FORS Analysis
The UV–visible-NIR reflectance spectrum is useful to distinguish the
different classes of brown-black colourants, as already shown in previ-
ous works [31,35]. The FORS spectra of IGI, carbon-based pigments, bis-
tre and sepia are shown in Fig. 2. While the other spectra are straight
lines at nearly 0% reflectance all through the acquisition range, the spec-
trum of IGI is different in that it shows a marked rising after the red re-
gion, which accounts for its brownish tone. Note that deep black IGIs
also show a similar feature; in those cases the rising occurs in the NIR,
which is the reason why the ink/paint is perceived more black than
brown.
It is not easy interpreting the FORS spectrum of IGI, even when
turned into apparent absorbance coordinates. In fact, the only apparent
spectral feature present in the spectrum is an inflection point occurring
beyond ca. 730 nm. The inflection point varies widely and this variation
explains why the relative paints are perceived with hues in the range
from light brown to black, respectively. It must be considered that the
chromatic appearance of IGI, and therefore the exact position of the in-
flection point in the FORS spectrum, depends on several variables. The
large variety of recipes and the natural origin of the involved materials
mean that several different components (but also impurities) are pres-
ent in historical IGIs; in addition, various degradation mechanisms are
possible, among which those causing changes in colour [36]. According
to Jembrih-Simbürger et al. [37] there could be a correlation between
the colour of IGI and its metal content, but other authors [21] only par-
tially acknowledged this hypothesis; in fact in many cases carbon could
be added to reinforce a dark hue. Krekel [38] found that the colour is
rather related to the acidity of the mixture, while Sistach [39] investigat-
ed the effect of additional organic substances that could enter in the
original recipes. In the end, the composition of an IGI can strongly influ-
ence its spectral features and it may seem hard relying on the sole posi-
tion of the inflection point to accomplish its identification; nevertheless
it can be safely stated that the overall shape of the FORS spectrum, with
its typical rising in the NIR, is characteristic enough to allow its discrim-
ination from the other brown-black colourants used in medieval minia-
ture painting art.
In artworks later than 17th century, IGI could not be distinguished
from Van Dyke brown or Kassel earth, a colourant which FORS spectrum
is almost identical [40]. This pigment has an organic composition for
N90% with a small percentage of iron and manganese oxides. Indeed,
since Van Dyke brown was introduced in painting since the late Renais-
sance [41], the similarity of its spectral behaviour with that of IGI is not a
real issue as far as medieval manuscripts are concerned; instead, its
spectrum can be helpful in order of interpreting the reflectance spec-
trum of IGI. Van Dyke brown is considered as composed mainly of
humic acids, i.e. complex macromolecules characterised, among other
features, by several phenolic groups. Taking into account the tannin
composition of galls, if we consider the structures of IGI proposed
with some differences by Krekel [38] and recently by Ponce et al. [42]
in which central Fe3+ ions are coordinated by phenolic groups, we can
hypothesise that similar interactions are at the base of the spectral fea-
tures of both IGI and Van Dyke brown, characterised by an intense ab-
sorption band throughout the visible region but without any selective
absorption bands in such spectral region. It is reasonable excluding
the main electronic transitions reported by several previous studies on
iron oxides, due to Fe3+ − ligand field, magnetically coupled Fe3+ cat-
ions in adjacent sites and ligand-to-metal charge transfer [43].
One limit in FORS identification is when a highly dilute IGI pigment
is used to obtain a faintly brown hue. The more will be diluted the IGI
mixture, the lower will be its hiding power and the higher will be the
contribution of the support (parchment, paper or papyrus) in the
resulting FORS spectrum; on those cases the position of the inflection
point will be hardly predictable. In Fig. 3 the spectra of IGIs laid on
parchment at different pigment-to-medium ratios are shown: it can
be clearly seen that the inflection point is shifted towards the UV region
as long as IGI becomes more diluted, until the spectrum of the pigment
cannot be distinguished from the one of parchment.
One more aspect to consider is the possible addition of carbon pig-
ment to IGI in order to darken its tone [1]. Our experiment of mixtures
at different IGI/carbon ratios demonstrates that the typical shape of
IGI with the raising in the NIR region is maintained up to a ratio of 1:1
(spectra shown in Fig. 4); a major effect, however, is a marked decrease
of the reflectance.
Also, it is important to note that FORS can provide only the plain
identification of IGI: any information concerning the presence of other
materials (i.e. additives such as carbon, honey, vinegar, beer, etc. as sug-
gested by medieval recipes) in most cases would request a sample or at
least more powerful techniques such as Raman spectroscopy or FT-IR
spectrophotometry.
In the end, the identification of IGI in a non-invasive way by means
of FORS analysis showed to be selective and reliable enough, except
for extreme cases (i.e. very light brown paints). Considering the advan-
tages of the FORS technique (availability of truly portable devices, ease
of use, speediness of analysis, etc.), this seems to be the proper choice
in a delicate situation such as the analysis of ancient manuscripts. This
approach is in fact of the highest relevance as far as such artworks are
concerned, because these valuable items are extremely fragile and usu-
ally their inspection is allowed only in situ because sampling is rarely
allowed.
Fig. 3. FORS spectra of standard IGIs at different pigment-to-medium ratios.
 M. Aceto, E. Calà / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 187 (2017) 1–8
Fig. 4. FORS spectra of IGI/carbon mixtures.
3.2. XRF Analysis
XRF spectrometry is used here as an example of elemental technique
applied to the identification of IGI. It is however the most plausible
among those existing, because it can be used in situ with ease, due to
the availability of truly portable systems.
First at all, it must be considered that, while the absence of iron is a
clear indication that the pigment under analysis is not IGI, the presence
of iron must be taken as a clue only and not as a confirmation. In fact,
iron can come from other sources apart from being intentionally
added in the pigment as vitriol. The high penetration capability of X-
rays can surely provide elemental contributions from the support;
small amounts of Fe are usually present in the parchment as a conse-
quence of the processes involved in its preparation. Contamination
could arise from tools used in working raw materials. Finally, environ-
mental contamination must be considered also, since iron is perhaps
the most widespread metal in the atmospheric particulate matter [44].
Of course, care must be taken in choosing measurements points free
from unspecific contribution from the reverse side of the analysed
folios.
Some works in the scientific literature suggested that care must
be taken in basing the identification of IGI on the presence of iron
only [45]. Roldan et al. [24,25] found Fe and Ca as the most abundant
elements in sepia extracts and in bistre, along with minor amounts of
K, S and other transition metals, including Cu and Zn. The metal con-
tent could be related to the chelating ability of melanin in sepia and
of carboxylic/hydroxyl functions in bistre. These authors concluded
that elemental analysis cannot provide a reliable, selective charac-
terisation of brown-black colourants. Similarly Bicchieri et al. [46]
found a relatively small amount of iron in the ink of an 18th century
manuscript, and by means of complementary molecular techniques
identified the ink more correctly as logwood added with an iron
salt. On the other side, Delange et al. [47] hypothesised the use of
iron-based inks on some demotic papyri dating to 3rd-1st centuries
B.C. and used by the Greeks in Egypt, according to the high iron con-
tent as determined with Particle-Induced X-ray emission (PIXE)
analysis. The authors suggested that the inks were of the iron gall
type, despite the fact that sulphur was found to be absent. However,
their results have never been confirmed by other authors; as an ex-
ample, Scott [48] in his recent review on Egyptian pigments, did
not mention IGI as a material used by Egyptians, neither for writing
nor for painting.
It is hard to give typical values for the elements usually found in IGI
paints, as it is for inks. As stated before, the elemental content is not
automatically linked to the colour, so that, while a light brown IGI is
usually characterised by a low Fe content, the contrary is not necessarily
5
true, i.e. a dark IGI could owe its hue partially to the addition of carbon
or other organic compounds.
One element which must be always present is sulphur, even if it is
not exclusive of IGI. The content is highly variable but it is usually higher
than iron because it can comes also from other vitriols (CuSO4, ZnSO4),
from extenders (BaSO4, CaSO4), from additives of organic nature or from
parchment itself. A high S/Fe ratio could be taken as a more reliable clue
for the identification of IGI in comparison with sepia, bistre and carbon
pigments, which have lower amounts of sulphur.
Cu and Zn can yield additional diagnostic information. Despite the
fact that sepia extracts and bistre can contain these metal ions at trace
levels, in cases where their amount is of the same order of magnitude
as Fe, this is a clear indication of the intentional use of other vitriols in
addition to FeSO4 and therefore of the presence of IGI.
To resume, XRF analysis can be used in the identification of IGI as a
preliminary survey but it can hardly be definitive in the identification
of IGI.
On the other side, however, elemental analysis performed by means
of XRF (as by means of PIXE or SEM-EDX) can yield additional informa-
tion on the element distribution which is useful to compare different
IGIs. This information can be used, for example, to evaluate whether
the same IGI preparation had been used on a manuscript for writing,
for painting or for drawing preparatory sketches. This comparison can-
not be done with molecular techniques such as FORS or Raman
spectroscopy.
3.3. Raman Analysis
In Fig. 5 the spectrum of IGI is compared with those of carbon-based
ink, bistre and sepia; the most characteristic Raman bands of the differ-
ent brown-black inks are summarised in Table 1. While bistre and sepia
can hardly be distinguished from carbon-based pigments in conven-
tional Raman mode, the use of SERS allows obtaining a more detailed
spectrum [24,25,49] so that they acquire more selective features. At
any rate, the Raman spectrum of IGI, both in conventional and in SERS
mode, is relatively simpler in the discrimination, according to the typical
peaks occurring at 560, 1320, 1482 and 1579 cm−1, in agreement with
the figures reported in previous studies [6,23]. Moreover, the combina-
tion of Raman spectral information and multivariate analysis allows
distinguishing different types of tannic sources in iron-based inks [50].
The identification of IGI by means of Raman analysis is by all means
the most selective among non-invasive techniques, because Raman
spectroscopy provides a spectral fingerprint for each colourant; it allows
also distinguishing IGI from Van Dyke brown or Kassel earth, a feature
not possible with FORS or XRF analysis. It must be considered, though,
that the application of Raman spectroscopy in the analysis of ancient
manuscripts is usually more difficult than that of FORS and XRF. The
Fig. 5. Raman spectra of IGI in conventional mode, a carbon-based ink in conventional
mode, bistre in SERS mode and sepia in SERS mode. Spectra are offset for clarity.
6 M. Aceto, E. Calà / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 187 (2017) 1–8

Table 1
– List of the most characteristic Raman bands. (w) = weak; (m) = medium; (s) = strong.

Ink Wavenumbers

IGI 405 (w), 560 (m), 962 (m), 1320 (s), 1482 (s), 1562 (w),
1579 (w)
Carbon-based 1302 (s), 1558 (s)
pigment
Bistre 998 (w), 1095 (w), 1168 (w), 1244 (w), 1334 (m), 1544 (m)
Sepia 883 (m), 949 (w), 1262 (m), 1335 (m), 1363 (w), 1538 (w),
1572 (m)

sample analysed must be kept in absolute stability conditions in order to

have reliable results (this is typical of all analytical techniques bearing a
microscope system); an additional issue is given by the fact that parch-
ment, or paper as well, tends to crisp under laser irradiation, so that it is
hard to keep the sample constantly at the focal distance. In addition,
touching the manuscripts or keeping them open at angles wider than
90° is usually prohibited; all these features strongly limit the possibility
of stabilising the support. Finally, it must be considered that manu-
scripts are fragile items that cannot be subjected to long diagnostic ses-
sions but rather must be analysed with fast methods. These issues are
overcome when a micro sample is available, because both conditions
of stability are addressed. Considering this, Raman analysis seems to
be most suitable in cases where sampling is allowed.
3.4. Survey on Illuminated Manuscripts
The survey we carried out on 6–16th centuries illuminated manu-
scripts by means of FORS analysis resulted in the identification of IGI
used as paint in several instances. The list of identification is reported
in Table 2. As far as the overall shape is concerned, the relative spectra
were quite similar in all instances, apart from the positions of the inflec-
tion points which were found to vary in the range 730–900 nm,
Fig. 6. Examples of FORS spectra obtained in the analysis of the manuscripts listed in Fig. 1.
corresponding to hues ranging respectively from light brown to dark
black. This range, of course, must not be taken as exhaustive, given the
previously discussed sources of variability; further measurements on
other manuscripts could extend the limits. Examples of FORS spectra
can be appreciated in Fig. 6; these spectra were taken in the black-
brown areas shown in Fig. 1. In the case of XRF analysis supported this
identification: apart from the obvious occurrence of iron, we always
found a percentage amount of sulphur higher than iron in all cases
when FORS identified IGI. S/Fe ratio ranged between 3 and 50 on a
semi-quantitative base. Potassium also was generally higher than iron,
ranging from 1 to 7 times the % amount.
Inside the group of painted areas in which IGI was identified by
means of FORS, a few of them were chosen on which to confirm the

Table 2
– List of manuscripts in which IGI was detected. BCVC: Biblioteca Capitolare in Vercelli (Italy); BDS: Biblioteca Diocesana in Susa (Italy); BNF: Bibliothèque nationale de France in Paris
(France); BNU: Biblioteca Nazionale Universitaria in Torino (Italy); KB: Kungliga Biblioteket in Stockholm (Sweden); ÖNB: Österreichische Nationalbibliothek in Vienna (Austria);

Subject Manuscript Provenance Source Period Use of IGI

Vienna Genesis Theol. gr.31 Antiocheia (Syria) ÖNB 6th century figurative details
Carmen Paschale E.IV.42 Bobbio (Italy) BNU 7th century crosses
Hieronymus, in Isaiam Latin 11627 Corbie (France) BNF 8th century figurative details
Codex Aureus A 135 Canterbury (UK) KB 8th century figurative details
Liber quattuor Evangeliorum CXXXIV Bavaria (Germany) BCVC 9th century figurative details
Homiliae G.V.7 Bobbio (Italy) BNU 9th century figurative details
Regula, et alia G.V.4 Bobbio (Italy) BNU 9-10th century figurative details
Moralia sive expositio in job F.I.6 Bobbio (Italy) BNU 9-10th century figurative details
Evangeliarium Latin 1126 France BNF 9-10th century figurative details
Vita sancti Hieronymi presbyteri F.III.15 Bobbio (Italy) BNU 10th century large initials
Passionarium F.III.16 Bobbio (Italy) BNU 10th century letter outlines
Psalmi davidici LXII ? BCVC 10th century figurative details
Psalterium G.V.2 Bobbio (Italy) BNU 10-11th century letter outlines
Homilarium F.II.19 Bobbio (Italy) BNU 10-11th century letter outlines
Homilarium F.II.20 Bobbio (Italy) BNU 10-11th century letter outlines
Vita sancti Columbani discipulorum F.IV.12 Bobbio (Italy) BNU 10-12th century figurative details
Antiphonarium G.V.20 Bobbio (Italy) BNU 11th century letter outlines
Commentarius in Apocalypsim J.II.1 Catalunya (Spain) BNU 11-12th century figurative details
Breviarium monasticum Bobiense F.II.10 Bobbio (Italy) BNU 11-12th century preparatory sketch
Decretum Gratiani XXV Italy (Lombardy?) BCVC 12th century figurative details
Evangeliarium C Italy (Lombardy?) BCVC 12th century figurative details
Liber in quo continetur ordo… CLXXIX Vercelli (Italy) BCVC 13th century figurative details
Biblia sacra E.VI.1 BNU 13th century letter outlines
Liber psalmorum David XVI ? BCVC 14th century figurative details
Missale iuxta ritum eusebianum CCXVI Italy BCVC 14th century letter outlines
Psalmi et Hymni CCXXIII Italy BCVC 14th century letter outlines
Biblia sacra E.IV.44 ? BNU 14th century letter outlines
Missale Romanum E.II.4 Aosta (Italy) BNU 14th century figurative details
Vitae et acta sanctorum J.II.17 Sicily (Italy) BNU 14th century preparatory sketch
Officia in festo S. Colimbani F.II.22 Bobbio (Italy) BNU 15th century figurative details
Expositio in Aristotelis Physica XV.II.63 Venice (Italy) BNU 15th century figurative details
Breviarum Romanum E.III.7 Italy (Lombardy?) BNU 15th century figurative details
Indulgenza I-851 Roma BDS 16th century figurative details
 M. Aceto, E. Calà / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 187 (2017) 1–8
identification by means of Raman analysis. Such items were selected ac-
cording to the availability of micro samples taken from the painted
areas. In all cases, the samples yielded invariably the typical spectrum
of IGI in conventional Raman mode. Therefore, considering the good
similarity among the FORS spectra of the samples listed in Table 2, we
found reasonable extending the confirmation of the presence of IGI to
all items.
4. Conclusions
From the analytical point of view, the present work tried to better
define the possibility of identifying IGI as a pigment on illuminate man-
uscripts, evaluating the pros and cons of three different techniques:
FORS, Raman spectroscopy and XRF spectrometry. With concern to in
situ non-invasive analysis, Raman spectroscopy is in theory the best
choice according to its higher diagnostic power, but FORS seems to pro-
vide the better compromise between selectivity and ease of application,
in particular whenever difficult conditions (e.g. manuscripts that can be
hardly handled) are involved.
New analytical evidences have been given on the particular use of IGI
in the palette of ancient illuminators. We carried out in non-invasive and
micro-invasive way a diagnostic survey on manuscripts datable in the
range 6th–16th centuries, showing that, apart from its widespread use
as an ink for writing and drawing, IGI was largely used as a pigment
too. The large number of identification obtained allows us to hypothesise
that this pigment was used all through medieval Europe up to at least the
Renaissance, where its use is already documented in drawing.
It is more than obvious that the occurrence of IGI in miniature paint-
ings older than 6th century or more recent than 16th century cannot be
excluded, as is its use beyond Europe; further measurements could in-
stead widen the time range and the geographic area. Also, it must be
considered that it is not possible distinguishing between IGI and Van
Dyke brown on artworks from the 17th onwards. Nevertheless, the
present study allows shedding a new light on the use of this colourant
all along the period of medieval and Renaissance miniature painting art.
Acknowledgements
Authors would like to thank the persons who allowed analytical in-
vestigations on the manuscripts under their responsibility: Dr. T.
Leonardi (Biblioteca Capitolare in Vercelli, Italy), Dr. A. Zonato
(Biblioteca Diocesana in Susa, Italy), Dr. F. Porticelli (Biblioteca
Nazionale Universitaria in Torino, Italy), Dr. C. Denoël and Dr. C. Förstrel
(Bibliothèque nationale de France in Paris), Dr. C. Hofmann
(Österreichische Nationalbibliothek in Vienna) and Dr. C. Ahlgren
(Kungliga Biblioteket in Stockholm, Sweden).
References
[1] C. James, The evolution of iron gall ink and its aesthetical consequences, in: A.J.E.
Brown (Ed.), Postprints of the Iron Gall Ink Meeting, Conservation of Fine Art, New-
castle upon Tyne 2001, pp. 13–22.
[2] T.E. Edmonds, An indicator of its time: two millennia of the iron–gall-nut test, Ana-
lyst 123 (1998) 2909–2914.
[3] G. Plinius Secundus, Naturalis Historiae, Book 34, p. 112, in: H. Rackham (Ed.), Pliny
Natural History, Books XXXIII–XXXV, Loeb Classical Library, Harvard University
Press, Cambridge, MA, 1995.
[4] C.R. Dodwell, Theophilus: The Various Arts, Nelson, London, 1961.
[5] G. Moorhead, S. Mazzarino, F. Marzo, B. Knight, A physical perspective of Codex
Sinaiticus: an overview from British Library, in: S. McKendrick, D. Parker, A.
Myshrall, C. O'Hogan (Eds.), Codex Sinaiticus: New Perspectives on the Ancient Bib-
lical Manuscript, The British Library, London 2015, pp. 221–238.
[6] M. Aceto, A. Agostino, E. Boccaleri, A. Cerutti Garlanda, The Vercelli gospels laid
open: an investigation into the inks used to write the oldest gospels in Latin, X-
Ray Spectrom. 37 (2008) 286–292.
[7] Reconstructions of iron-gall ink recipes for the InkCor project, in: A. Stijnman, M.
Clarke, J.H. Townshend, A. Stijnman (Eds.), Art of the Past: Sources and Reconstruc-
tions, Archetype Publications, London 2005, pp. 125–134.
[8] G. Verri, S. Tanimoto, C. Higgitt, Inks and Washes, in: J. Ambers, C. Higgitt, D.
Saunders (Eds.), Italian Renaissance Drawings : Technical Examination and Analysis,
Archetype Publications/The British Museum, London 2010, pp. 57–75.
7
[9] An examination of iron gall ink drawings in the British Museum deacidified with
barium hydroxide, in: C. Barry, J. Fields, M. Clarke, J.H. Townshend, A. Stijnman
(Eds.), Art of the Past: Sources and Reconstructions, Archetype Publications, London
2005, pp. 83–88.
[10] M. Strlič, L. Cséfalvayová, J. Kolar, E. Menart, J. Kosek, C. Barry, C. Higgitt, M. Cassar,
Non-destructive characterisation of iron gall ink drawings: not such a galling prob-
lem, Talanta 81 (2010) 412–417.
[11] M. Shelley, S.A. Centeno, Technical studies: observations on the drawing materials
used by Van Gogh in Provence, in: C. Ives, S.A. Stein, S. Van Heugten, M. Vellekoop
(Eds.), Vincent Van Gogh: The Drawings, Metropolitan Museum of Art, New York
2005, pp. 348–359.
[12] M. Reháková, M. Čeppan, K. Vizárová, A. Peller, D. Stojkovičová, M. Hricková, Study
of stability of brown-gray inks on paper support, Heritage Sci. 3 (2015) 1–7.
[13] L. Burgio, S. Bioletti, B. Meehan, Non-destructive, in situ analysis of three early Me-
dieval Manuscripts from Trinity College Library Dublin (Codex Usserianus Primus,
Book of Durrow, Book of Armagh), in: J. Hawkes (Ed.), Making Histories: Proceed-
ings of the Sixth International Conference on Insular Art, Shaun Tyas, Donington
2013, pp. 42–49.
[14] S. Bioletti, R. Leahy, J. Fields, B. Meehan, W. Blau, The examination of the book
of Kells using micro-Raman spectroscopy, J. Raman Spectrosc. 40 (2009)
1043–1049.
[15] B. Doherty, A. Daveri, C. Clementi, A. Romani, S. Bioletti, B. Brunetti, A. Sgamellotti, C.
Miliani, The book of Kells: a non-invasive MOLAB investigation by complementary
spectroscopic techniques, Spectrochim. Acta A 115 (2013) 330–336.
[16] M. Aceto, A. Agostino, G. Fenoglio, M. Gulmini, V. Bianco, E. Pellizzi, Non invasive
analysis of miniature paintings: proposal for an analytical protocol, Spectrochim.
Acta A 91 (2012) 352–359.
[17] L. Burgio, R.J.H. Clark, R.R. Hark, Raman microscopy and X-ray fluorescence analysis
of pigments on medieval and Renaissance Italian manuscript cuttings, Proc. Natl.
Acad. Sci. U. S. A. 107 (2010) 5726–5731.
[18] I. Nastova, O. Grupče, B. Minčeva-Šukarova, S. Turan, M. Yaygingol, M. Ozcatal, V.
Martinovska, Z. Jakovlevska-Spirovska, Micro-Raman spectroscopic analysis of inks
and pigments in illuminated medieval old-Slavonic manuscripts, J. Raman
Spectrosc. 43 (2012) 1729–1736.
[19] I. Nastova, O. Grupče, B. Minčeva-Šukarova, M. Kostadinovska, M. Ozcatal, Spectro-
scopic analysis of pigments and inks in manuscripts. III. Old-slavonic manuscripts
with multicolored rubication, Vib. Spectrosc. 78 (2015) 39–48.
[20] I. Nastova, O. Grupče, B. Minčeva-Šukarova, M. Ozcatal, L. Mojsoska, Spectroscopic
analysis of pigments and inks in manuscripts: I. Byzantine and post-byzantine man-
uscripts (10–18th century), Vib. Spectrosc. 68 (2013) 11–19.
[21] M. Van Bos, L. Watteeuw, Composition of iron gall inks in illuminated manu-
scripts (eleventh-sixteenth century): the use by scribes and illuminators, in:
M.J. Driscoll (Ed.), Care and Conservation of Manuscripts 14, Proceedings of
the Fourteenth International Seminar, Museum Tusculanum Press, Copenhagen
2014, pp. 177–193.
[22] J. Winter, W. West Fitzhugh, Pigments based on carbon, in: B.H. Berrie (Ed.), Artists'
Pigments. A Handbook of their History and Characteristics, vol. 4, Washington & Ar-
chetype Publications, London, National Gallery of Art 2007, pp. 1–37.
[23] A.S. Lee, V. Otieno-Alego, D.C. Creagh, Identification of iron-gall inks with near-infra-
red Raman microspectroscopy, J. Raman Spectrosc. 39 (2008) 1079–1084.
[24] M.L. Roldan, S.A. Centeno, A. Rizzo, An improved methodology for the characteriza-
tion and identification of sepia in works of art by normal Raman and SERS,
complemented by FTIR, Py-GC/MS, and XRF, J. Raman Spectrosc. 45 (2014)
1160–1171.
[25] M.L. Roldan, S.A. Centeno, A. Rizzo, Y. van Dyke, Characterization of bistre pigment
samples by FTIR, SERS, Py-GC/MS and XRF, MRS Proc. 1656 (2014)http://dx.doi.
org/10.1557/opl.2014.823.
[26] J. Senvaitiene, A. Beganskiene, A. Kareiva, Spectroscopic evaluation and characteriza-
tion of different historical writing inks, Vib. Spectrosc. 37 (2005) 61–67.
[27] D.M. Goltz, A review of instrumental approaches for studying historical inks, Anal.
Lett. 45 (2012) 314–328.
[28] G. Chiavari, S. Montalbani, S. Prati, Y. Keheyan, S. Baroni, Application of analytical py-
rolysis for the characterization of old inks, J. Anal. Appl. Pyrolysis 80 (2007)
400–405.
[29] T. Bardon, R.K. May, P.F. Taday, M. Strlič, Systematic study of terahertz time-domain
spectra of historically informed black inks, Analyst 138 (2013) 4859–4869.
[30] J. Havermans, H. Abdul Aziz, H. Scholten, Non destructive detection of iron gall inks
by means of multispectral imaging part 1: development of the detection system,
Restaurator 24 (2003) 55–60.
[31] L. Gál, Michal Čeppan, Milena Reháková, Vladimír Dvonka, Jarmila Tarajčáková, Jozef
Hanus, Chemometric tool for identification of iron–gall inks by use of visible–near
infrared fibre optic reflection spectroscopy, Anal. Bioanal. Chem. 405 (2013)
9085–9091.
[32] M. Aceto, A. Agostino, G. Fenoglio, A. Idone, M. Gulmini, M. Picollo, P. Ricciardi, J.K.
Delaney, Characterisation of colourants on illuminated manuscripts by portable
fibre optic UV–visible-NIR reflectance spectrophotometry, Anal. Methods 6 (2014)
1488–1500.
[33] A. Idone, M. Aceto, E. Diana, L. Appolonia, M. Gulmini, Surface-enhanced Raman
scattering for the analysis of red lake pigments in painting layers mounted in
cross sections, J. Raman Spectrosc. 45 (2014) 1127–1132.
[34] P.C. Lee, D. Meisel, Adsorption and surface-enhanced Raman of dyes on silver and
gold sols, J. Phys. Chem. 86 (1982) 3391–3395.
[35] J.G. Neevel, The development of in-situ methods for identification of iron–gall inks,
in: J. Kolar, M. Strlič (Eds.), Iron–Gall Inks: On Manufacture, Characterization, Degra-
dation and Stabilization, Narodna i univerzitetna knjižnica, Ljubljana 2006,
pp. 147–172.
8 M. Aceto, E. Calà / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 187 (2017) 1–8
[36] O. Hahn, W. Malzer, B. Kanngiesser, B. Beckhoff, Characterization of iron-gall inks in
historical manuscripts and music compositions using X-ray fluorescence spectrom-
etry, X-Ray Spectrom. 33 (2004) 234–239.
[37] D. Jembrih-Simbürger, V. Desnica, M. Schreiner, E. Thobois, H. Singer, K. Bovagnet,
Micro-XRF analysis of watercolours and ink drawings by Albrecht Dürer in the
Albertina in Vienna, Technè 22 (2005) 32–37.
[38] C. Krekel, Chemistry of historical iron gall inks, Int. J. Forensic Doc. Exam. 5 (1999)
54–58.
[39] M.C. Sistach, I. Espadler, Organic and inorganic components of iron gall inks, in: J.
Bridgland (Ed.), Preprints, 10th Triennial Meeting of the ICOM Committee for Con-
servation, vol. 2, ICOM Committee for Conservation, Paris 1993, pp. 485–490.
[40] Fiber Optics Reflectance Spectra (FORS) of Pictorial Materials in the 270–1700 nm
range, http://fors.ifac.cnr.it/index.php 2016 (accessed 25.11.16).
[41] R. Feller, R.M. Johnston-Feller, Van Dyke Brown, in: E.W. Fitzhugh (Ed.), Artists' Pig-
ments. A Handbook of Their History and Characteristics, vol. 3, Oxford University
Press, New York 1997, pp. 157–190.
[42] A. Ponce, L.B. Brostoff, S.K. Gibbons, P. Zavalij, C. Viragh, J. Hooper, S. Alnemrat, K.J.
Gaskell, B. Eichhorn, Elucidation of the Fe(III) gallate structure in historical iron
gall ink, Anal. Chem. 88 (2016) 5152–5158.
[43] J. Torrent, V. Barrón, Diffuse reflectance spectroscopy of iron oxides, in: A.T. Hubbard
(Ed.), Encyclopedia of Surface and Colloid Science, vol. 1, Marcel Dekker, New York
2002, pp. 1438–1446.
[44] H. Puxbaum, A. Limbeck, Chemical compounds in the atmosphere, in: E. Merian, M.
Anke, M. Ihnat, M. Stoeppler (Eds.), Elements and Their Compounds in the Environ-
ment, 2nd editionWiley-VCH, Weinheim 2004, p. 21.
[45] J. Colbourne, A survey of methods used in the technical examination and analysis of
brown inks, in: A.J.E. Brown (Ed.), Postprints of the Iron Gall Ink Meeting, Conserva-
tion of Fine Art, Newcastle upon Tyne 2001, pp. 37–46.
[46] M. Bicchieri, M. Monti, G. Piantanida, A. Sodo, All that is iron-ink is not always iron-
gall! J. Raman Spectrosc. 39 (2008) 1074–1078.
[47] E. Delange, M. Grange, B. Kusko, E. Mènei, Apparation de l'encre métallogallique en
Égypte à partir de la Collection de Papyrus du Louvre, Revue d'Egyptologie 41
(1990) 213–217.
[48] D. Scott, A review of ancient Egyptian pigments and cosmetics, Stud. Conserv. 61
(2016) 185–202.
[49] E.P. Tomasini, E.B. Halac, M. Reinoso, E.J. Di Liscia, M.S. Maier, Micro-Raman spec-
troscopy of carbon-based black pigments, J. Raman Spectrosc. 43 (2012)
1671–1675.
[50] G. Piantanida, E. Menart, M. Bicchieri, M. Strlič, Classification of iron-based inks by
means of micro-Raman spectroscopy and multivariate data analysis, J. Raman
Spectrosc. 44 (2013) 1299–1305.