APPLICATION NOTE

Quality Management and

Production Control in Paper
Manufacturing
Confocal Raman Imaging and correlative Raman-AFM
for chemical and structural characterization
WITec GmbH, Lise-Meitner-Str. 6, 89081 Ulm, Germany
fon +49 (0) 731 140 700, fax +49 (0) 731 140 70 200
[email protected], www.WITec.de
Quality Management and Production Control in Paper Manufacturing
Confocal Raman Imaging and correlative Raman- Atomic Force Microscopy for chemical and structural characterization

Introduction Materials & Methods array was only a few milliseconds or 10s
of milliseconds, allowing the acquisition
Papermaking has an over 2000-year-long All experiments were performed with a of thousands of Raman spectra within
manufacturing history and launched a re- confocal Raman-AFM microscope of the minutes. The number of pixels and the in-
volution in communication and packaging. WITec alpha300 series. The alpha300 tegration times per spectrum are listed in
Over the centuries, paper’s contribution series is optimized for high-resolution the sections describing the corresponding
to civilization is difficult to overstate and imaging of samples and is able to combine Raman image throughout this document.
the refinement of its production processes several analytical measuring techniques in By rotating the microscope turret, it is
continues to benefit many industries. one instrument such as diffraction-limited possible to switch between confocal Ra-
Paper typically consists of plant fibers from confocal microscopy, chemically sensitive man microscopy and AFM without losing
wood. Those fibers are mainly composed Raman spectroscopy and cantilever-based the sample position. The samples were
of cellulose and lignin. The properties of scanning probe microscopy, including AFM imaged in AFM AC Mode using ArrowFM
paper are determined primarily by the and SNOM. Using these instruments it is cantilevers (Nanoworld, Germany) with
fiber composition. Processes and chemical possible to analyze the same sample po- a resonance frequency in the range of
treatments such as pulping, bleaching, sition with various measuring techniques 70–90 kHz and damping of r = 50 percent,
sizing and coating further influence the sequentially or even simultaneously (Ra- recording both topography and phase
characteristics of the produced paper man and AFM). Automated measurement images simultaneously.
including suitability for printing. procedures and large area scans are also
A thorough understanding of the micro- offered.
and nanostructures of the paper in its suc- The microscope can be equipped with a
cessive stages of production as well as the selection of excitation lasers for confocal
final paper composition at a sub-micro- Raman microscopy. For the measurements
meter level is essential for further impro- presented in this application note, either a
vement of its properties. With nano-ana- frequency-doubled 532 nm NdYAG laser or
lytical investigation methods such as a 785 nm diode laser (NIR) was used. The
confocal Raman microscopy and atomic excitation light was polarized horizontally
force microscopy (AFM) this detailed (in the x-direction) with respect to the
knowledge is readily accessible. Raman image. All measurements were
Confocal Raman imaging reveals valuable performed using an ultra-high-throughput
information about the spatial distribution spectrometer (UHTS) and back-illuminated
of the chemical compounds while AFM CCD cameras.
delivers high-resolution images of physical The distribution of chemical species on the
surface properties (e.g. roughness). In sample was determined in Raman spectral
combination, the techniques provide a imaging mode. In this mode a complete
comprehensive examination of the quality Raman spectrum is recorded at every
of the produced paper. image pixel, producing a 2D array of Ra-
man spectra. By extracting individual fea- alpha300 Raman AFM microscope
This application note presents results of
confocal Raman and AFM measurements tures such as Raman peak intensity, peak
from several stages of the paper manufac- width or peak position from the recorded
turing process followed by examples that spectra, Raman images can be generated
provide insight into printing and writing and fitting algorithms can be employed to
on paper. visualize the distribution of compounds
and characteristics. The illustration colors
are coded in the same way as the corres-
ponding spectra. The acquisition time for
a Raman spectrum within a 2D spectral
Results & Discussion paper, technical paper, decorative paper, array of spectra, two distinct spectra for
etc.) further treatments such as coating, cellulose were extracted and are shown
Paper Production laminating or impregnating are necessary in Fig. 1b. A significant difference bet-
Depending on the scope of use, paper has in order to achieve the required properties. ween the two spectra can be seen in the
to meet multiple requirements, leading In the following, the spectral differences intensity variation of the Raman bands
to variation in production methods. For of pure cellulose are first illustrated then at wavenumbers 1095 cm-1 and 1118 cm-1.
the production of packaging paper with followed by the analysis of residues on These two bands are characteristic for C-O
short fiber lengths and a yellowish color, cellulose fibers. and C-C stretching, respectively, and are
mechanical pulping is sufficient. In this laser polarization-direction dependent.
process, lignin is not removed from the The distribution of these two Raman
pulp, which offers a high production yield. Raman imaging of a pure cellulose paper spectra over the examined surface area is
High quality paper for writing and printing web presented in Fig. 1c. The color of the spectra
requires chemical pulping. In this process Fig. 1a shows the video image of a paper in Fig. 1b matches the color in the Raman
the chemical structure of lignin is destro- web consisting of pure cellulose fibers. The image 1c. Thus, pixels that are bright green
yed and dissolved in the pulp and can be red square in this video image indica- show a high spectral similarity to the
washed off of the cellulose fibers. Further- tes the area scanned in Raman spectral green spectrum, whereas bright red pixels
more, the fiber length is preserved for gre- imaging mode using green excitation and show a high spectral similarity to the red
ater strength. The finished paper is then a high numerical aperture air objective. An spectrum. The laser polarization direction
produced by feeding the pulp into a paper area of 50x50 µm2 was scanned and a 2D is indicated with the white arrow in Fig.
machine where it is formed as a paper web spectral array of 128x128 complete Raman 1c. Fibers oriented parallel to the laser
and the water is removed by pressing and spectra was recorded with an integration polarization direction show an increased
drying. For specialty paper (photographic time of 80 ms/spectrum. From this 2D C-O stretching band at 1095 cm-1 (red)

a) laser polarization direction c)

1095 b)

1118

relative wavenumbers (cm-1)

Fig. 1: Confocal Raman imaging of raw cellulose

fiber bundles. Video image (a), orientation-de-
pendent Raman spectra of cellulose (b) and
color-coded Raman image of cellulose fibers (c).
Image area 50x50 µm2, 128x128 pixels, integration
time: 80 ms/spectrum.
 a) c)

cellulose b) d) Fig. 2: Confocal Raman study of resin

distribution around cellulose fibers. Video
resin image (a), Raman spectra of resin (green)
and cellulose (red) (b), color-coded Raman
image of a surface scan, image area 45x17

z-direction
µm2, 100x100 pixels, integration time:
112 ms/spectrum (c), and a depth profile,
image area 45x40 µm2, 100x100 pixels,
integration time: 112 ms/spectrum (d).

9 µm
relative wavenumbers (cm-1)

compared to fibers oriented perpendicular bution on cellulose fibers. ted in Figs. 2c and 2d clearly show that
to the laser polarization direction (green). Fig. 2 shows an example of a study of resin the resin surrounds the cellulose fibers
Thus, confocal Raman microscopy easily distribution in paper. For this experiment without penetrating into them.
reveals the orientation of cellulose fibers a sheet of paper was soaked in water to
in the paper web even if this cannot be reduce the strong fluorescence of the
determined optically. sample. The alpha300 confocal Raman
microscope was equipped with an NIR ex-
Raman and AFM imaging of residues on citation laser and a back-illuminated deep
cellulose fibers depletion CCD camera. A water immersion
As mentioned before, lignin is washed off objective (NA =1.0) was used for these
of the cellulose fibers during the chemical measurements. The video image in Fig.
pulping. Furthermore, small quantities of 2a shows several cellulose fibers oriented
extractives such as resin acids, triglyceri- almost perpendicular to the polarization
des, steryl esters, fatty acids, and sterols direction of the laser. The red square deno-
can be separated from wood by using tes the in-plane (x-y) scanned area of 45x17
solvents such as acetone or dichloromet- µm2. The blue line indicates where the
hane. During the papermaking process, depth (x-z) scan was performed. For the
these extractives are partly released into in-plane scan an array of 100x100 Raman
the circulation water. Today, efforts are spectra was recorded with an integration
made to decrease the water consumption time of 112 ms/spectrum. Fig. 2b shows the
in paper mills. Hence, water is re-circulated characteristic Raman spectra extracted
in the process, which leads to a build-up of from the spectral array. The distribution
dissolved and colloidal substances. The en- of cellulose (red) and resin (green) on the
richment of extractives in the circulating analyzed sample area is shown in Fig. 2c.
water in particular leads to problems such Fig. 2d shows a depth (x-z) scan of the
as precipitates, paper strength decreases sample. In this case an array of 100x100
and foaming. The following two examples Raman spectra was recorded over a range
show studies of residues and their distri- of 45x40 µm2. The Raman images presen-
A further survey was performed using the AC mode. Fig. 3c shows the topography
combination of confocal Raman and AFM image of a fiber bundle consisting of small,
to detect traces of wood extractives on oriented fibrills, characteristic for cellu-
pure cellulose fibers. An area of 20x40 µm2 lose fibers. The simultaneously recorded
was scanned in Raman spectral imaging AFM phase image (Fig. 3d) reveals in the a)
mode by acquiring a spectral array of central part an additional contrast which
80x160 complete Raman spectra. From can be attributed to the hexane extractive.
this spectral array three distinct spectra High-resolution AFM AC images recorded
were extracted as shown in Fig. 3a. The from this region are shown in Figs. 3e (to-
green and blue spectra are the orientati- pography image) and 3f (phase image). The
on-dependent Raman spectra of cellulose, hexane extractives show a clear contrast in
whereas the red spectrum represents he- the phase image, revealing a material with
xane extractives. The distribution of these different mechanical properties than the
relative wavenumbers (cm-1)
three compounds on the analyzed sample cellulose fibrills underneath.
area is shown in the color-coded Raman hexane
image in Fig. 3b. The white rectangle in Fig. cellulose I
3b marks the area that was imaged in AFM cellulose II

Raman b) AFM Topography c) AFM Phase d) AFM Topography e)

AFM Phase f)

Fig. 3: Confocal Raman AFM images of traces of wood extractives on cellulose fibers: Raman spectra of cellulose (green and blue spectrum) and hexane
extract (red spectrum) (a), color-coded Raman image of a cellulose fiber. Image area 20x40 µm2, 80x160 pixels, integration time: 100 ms/spectrum. (b),
high-resolution AFM topography (c) and phase image (d) of the fiber and zoom-ins (e: topography; f: phase).
 Table 1: Roughness Parameters
a) b)

Fig. 4: (a) The distribution of calcium carbonate

(blue) on cellulose paper (red and green color
show the different orientations of the cellulose
fibers), image area 150x150 µm2, 150x150 pixels,
integration time: 90 ms/spectrum. (b) The distri-
bution of starch (red), calcium carbonate (blue)
and a polymeric material (green) on cellulose
paper (yellow), image area 50x50 µm2, 100x100
pixels, integration time: 512 ms/spectrum.

Paper Finishing and an array of 150x150 Raman spectra 5a, revealing the fibrillar fine structure
The pure cellulose paper introduced in was recorded with an integration time of of cellulose fibers with diameters in the
the previous section is a yellowish, rough 90 ms/spectrum. Red and green in Fig. 4a nm-range. This surface is rather rough and
paper, which can hardly be used for writing denote the different orientation of the cannot be used for writing or printing. In
and printing. To provide a paper surface cellulose fibers, whereas blue shows the Figs. 5b and 5c the topography of lamina-
more suitable for e.g. high-resolution distribution of calcium carbonate on the ted paper is shown after applying one and
halftone screening, the raw paper is coated coated paper. In Fig. 4b additional additives two laminations, respectively. The roug-
with a thin layer of calcium carbonate and used in paper coating were investigated. hness parameters calculated from these
other additives. Confocal Raman imaging The confocal Raman image of 50x50 µm2 differently processed paper surfaces using
and AFM can provide valuable informati- reveals that the coating of this paper the WITec software suite for data analysis
on on their distribution and effect as the sample contains, along with calcium car- and evaluation are summarized in table
following examples will illustrate. bonate (blue color), starch grains (red ) and 1. From these results it is evident that at
a polymer matrix (green ). The cellulose is least two layers of lamination should be
Fig. 4a shows an investigation of a calcium shown in yellow. applied in order to decrease the surface
carbonate coating on paper. This image AFM can additionally contribute to the roughness. The combined Raman-AFM stu-
was obtained with a laser excitation wa- analysis of paper finishing with high-re- dies give insight into the chemical nature
velength of 532 nm and a 40x air objective solution images that describe the sur- of the additives and laminates as well as
(NA = 0.6). An area of 150x150 µm2 was face roughness of the paper (Fig. 5). The their effect on the surface roughness.
scanned in Raman spectral imaging mode topography of raw paper is shown in Fig.

a) b) c)

Fig. 5: High-resolution AFM images of untreated paper (a) and laminated paper after applying one layer (b) and two layers (c).
Paper Coating a depth of 100 µm. Most of the coatings Fig. 7a shows the unique Raman spectra of
Depending on the final application of the used form a well-defined interface without the two polymers (red and green) and the
paper, its coating can consist of several lay- penetrating into one another. characteristic Raman spectra of cellulose
ers with different chemical compositions. Another example of modified paper (blue) and calcium carbonate (orange).
Confocal Raman imaging can contribute to is shown in Fig. 7. Two adhesives were The color-coded depth profile (Fig. 7b)
the analysis of paper layers in a nondest- applied to the paper surface to produce a shows that the two polymers are sharply
ructive way. Fig. 6 shows a depth scan (x-z sticky paper label. In this experiment an distinguishable (red and green) from one
scan) obtained from multi-colored paper. area of 50x40 µm2 was scanned in the x-z another as well as from the paper beneath
An area of 60x100 µm2 was scanned in the plane and an array of 100x80 Raman spec- (blue and yellow).
x-z plane and an array of 100x250 Raman tra was acquired with an integration time
spectra was recorded with an integration of 100 ms/spectrum. The measurements
time of 50 ms/spectrum. This experiment were performed with an excitation laser
was performed using an excitation laser wavelength of 532 nm and an air objective
wavelength of 532 nm and an oil immersi- (NA = 0.9).
on objective (NA = 1.25). The distribution of Fig. 7: Confocal Raman depth
the various materials in the layers can be
polymer I a) scan through a paper sticker.
polymer II (a) Raman spectra calculated
clearly determined and imaged as shown
cellulose from the spectral array. (b)
in Fig. 6 (information about the chemical color-coded Raman image
calcium carbonate
composition remains undisclosed due to of depth profile showing the
confidentiality). The image clearly reveals two polymeric layers (red
the presence of 10 different layers along and green) applied on paper
(yellow and blue). Image area
50x40 µm2, 100x80 pixels,
integration time: 100 ms/
spectrum.

relative wavenumbers (cm-1)

b)
z-direction

z-direction

Fig. 6: Depth scan through a multilayer coating

of a paper sample. Different colors in the image
represent different coating materials. Image area
60x100 µm2, 100x250 pixels, integration time: 50
ms/spectrum.
Writing and Printing on Paper An array of 128x128 complete Raman spec- Summary
Confocal Raman imaging can also provide tra was acquired over 90 µm in the x-direc-
information about the qualities of inks tion and 50 µm in the z-direction with an Confocal Raman imaging and atomic force
applied on paper. Fig. 8a shows a video integration time of 100 ms/spectrum. The microscopy can be applied in all steps of
image recorded at the intersection of two distribution along the depth profile of the paper production and printing to control,
different inks. The area marked with the inks is shown in Fig. 8d, using the same examine and improve the process. During
red frame (50x50 µm2) was scanned in color-code as Fig. 8c. This depth profile the pulping and pressing process confocal
Raman spectral imaging mode by recor- clearly shows that the red-marked ink was Raman imaging can provide information
ding an array of 128x128 complete Raman applied on top of the green-marked ink, on the orientation of cellulose fibers and-
spectra with an integration time of 100 thus showing the order in which these several chemical residues from circulating
ms/spectrum. From the array of spectra, inks were applied to the paper. water or wood extractives on cellulose
three different Raman spectra were calcu- fibers can be identified unambiguously. In
lated. At some surface areas only lumine- c) further production processes such as finis-
scence could be detected (not shown as a hing and coating, confocal Raman imaging
spectrum), whereas other areas showed and AFM can reveal valuable information
the unique Raman spectra presented about the distribution of the applied sub-
in Fig. 8b. These two spectra show only stances and the surface structure. Coating
very small differences of which some are layers can be visualized and investigated
highlighted with arrows in the figure. In with depth scans and the qualities of
the color-coded Raman image presented in printing can be described in a non-dest-
Fig. 8c, blue denotes the luminescent areas ructive manner, without laborious sample
while the distribution of the two inks are preparation.
represented in red and green. An area whe- Confocal Raman imaging combined with
re both inks overlap is clearly visible. The high resolution AFM contribute to a more
order in which these inks were applied to detailed understanding of the paper ma-
the paper can be examined with a depth nufacturing process.
scan (x-z scan), performed along the black
line in the video image (Fig. 8a). d)

a)
z-direction

paste I

paste II
10 µm

Fig. 8: Confocal Raman imaging of inks on paper. Video image (a), evaluated Raman spectra (b),
color-coded Raman image recorded from the paper surface, image area 50x50 µm2, 128x128 pixels,
integration time: 100 ms/spectrum (c), and color-coded depth profile, image area 90x50 µm2, 128x128
pixels, integration time: 100 ms/spectrum (d). The inks are shown in red and green in the images.

ink I b)
ink II

relative wavenumbers (cm-1)