You are on page 1of 20

1/2001

Information for users of


METTER TOLEDO thermal analysis systems

13

Dear Customer
We are pleased to report that 2000 was another successful year for METTLER TOLEDO. In
particular, we enjoyed above average growth in Thermal Analysis. In response to this, we
have added a new colleague to our market support group/editorial team.
We are delighted to see that more and more customers want to publish their own articles
in UserCom.

Interpreting TGA curves


Georg Widmann

Artifacts
Artifacts are effects observed on the measurement curve that are not directly caused by
the sample, i.e. effects that do not relate to the properties you want to measure.
The main types of TGA artifacts are:
Buoyancy effects caused by the density of the surrounding gas decreasing on heating.
Typically this results in an apparent gain in weight of 50g to 200g. Since buoyancy
effects are reproducible, the curves can be corrected by performing an automatic blank
curve subtraction. This also applies to buoyancy effects due to gas switching, which is
often used in TGA analysis. All the curves shown in this article are blank curve corrected.
Fluctuations of the purge gas flow rate can also affect the measurement curve. The
flow rate should therefore be kept as constant as possible and not be changed during a
measurement.
Sudden loss of mass arising from the ejection of part of the sample. This often occurs
when a sample decomposes with the formation of gas. You can prevent this by covering
the sample with coarse grain Al2O3 or a crucible lid with a hole.
Apparent weight gains caused by samples that foam and make contact with the furnace wall. You can avoid this problem by using smaller samples.

Contents
TA Tip
Interpreting TGA curves
New in our sales program
DMA/SDTA861e
Applications
Determination of the fiber content
of composite materials by
thermogravimetry
Investigation of the segregation
behavior of different mortar constituents with TGA/SDTA
Determination of polymer crystallinity with DSC measurements
DSC characterization of short
chain branching distribution in
polyethylene
Tips
TMA/SDTA840: Rapid cooling between measurements

Measurement conditions
In TGA, the exchange of materials between
the sample and the immediate surroundings must be possible. The crucible is
therefore either open or covered using a lid
with a hole.
The measurements are normally performed with a dynamic temperature program at rates between 0.5 and 50K/min,
often at 20K/min. A starting temperature
of 25C to 30C is used in order to detect
possible drying of the sample. The final
temperature is usually relatively high, e.g.
600C for organic substances or 1000C
for inorganic samples because in most
cases the decomposition of the sample is
also of interest. Organic samples are often
pyrolyzed under nitrogen. The carbon
black formed during the pyrolysis process
or present in the sample as an additive
(e.g. in polymers) is then burnt in an air
atmosphere.
A purge gas or at least a balance protection
gas must be used for all measurements.

Occasionally a gain in weight is observed. This can result from:


chemical reactions (reaction with gaseous substances in the purge gas such as
O2, CO2 with the formation of nonvolatile or hardly volatile compounds)
physical transitions (adsorption of gaseous substances on samples such as active charcoal).

Interpretation
In addition to the actual TGA curves, several other different curves are used to aid
interpretation:
the first derivative (DTG curve, rate of
change of mass),
the SDTA curve (measures exothermic or
endothermic effects in a similar way to
DSC) and possibly
EGA, Evolved Gas Analysis, i.e. FTIR or
MS measurements of evolved gases and
vapors.
The interpretation of SDTA curves is analogous to that of DSC curves and has already
been described in UserCom 10 and 11. The
next edition of UserCom will include an
article on EGA.

Besides the true mass changes mentioned


above, pseudo mass changes occur when
the magnetic force exerted on a sample
changes. This effect takes place at the socalled Curie transition point of ferromagnetic samples. A permanent magnet placed
below the furnace provides the magnetic
field needed to observe such effects. The
technique is good way to measure Curie
transition temperatures. Conversely, Curie
transitions can be used to calibrate and
adjust the temperature of thermobalances
that are not equipped with DTA or SDTA.
The visual examination of the sample after the measurement if possible use a
reflected-light microscope often provides
qualitative information on the residue
(type of ash, glassy, white or colored powder, soot particles).
The shape of the TGA curve is usually
characteristic of the effect measured and
can be used for identification purposes.
The schematic and measured effects
shown in this article illustrate many of the
typical curve shapes observed in TGA. All
the curves shown have been blank curve
corrected.
If two steps overlap, a different heating
rate (higher or lower) or a smaller sample
weight often allows a better separation to
be achieved. Higher heating rates shift
chemical reaction steps to higher temperatures.

Most TGA curves display weight losses.


These are typically caused by:
chemical reactions (decomposition and
loss of water of crystallization, combustion, reduction of metal oxides)
physical transitions (vaporization,
evaporation, sublimation, desorption,
drying).

Chemical reactions
Thermogravimetry is often used to investigate chemical reactions in which weight
changes occur. A typical weight loss step is
about 100 C wide (for 1% to 99% conversion). The step usually develops slowly
from the initially horizontal TG curve. The
point of inflection is at about 60% conver-

2
UserCom 1/2001

sion. The radius of curvature is appreciably smaller at the end of the reaction than
at the beginning (Fig. 1a).
With stoichiometric reactions, the molar
mass of the eliminated molecule can be
determined as shown in the following example: the molar mass of aspartame (Fig.
3) is 312g/mol, the step of 10.4% at
180C therefore corresponds to 32g/mol.
Possible decomposition products are therefore O2, S or CH3OH. The most probable
substance is however methanol because
aspartame is not a peroxide and does not
contain sulfur.
Diffusion controlled reactions, in which
the transport of starting materials or products is restricted, proceed at an almost
constant rate (i.e. the slope of the TG
curve is almost constant, Fig. 1c).
Several steps often occur (Fig. 1d).
Explosive substances sometimes decompose so rapidly that the force of the recoil
disturbs the TGA signal (Fig. 1e). This
problem can be avoided by using smaller
sample quantities or by diluting the
sample with an inert substance.

Fig. 1. TGA chemical reactions:


a: thermal decomposition with the formation of
gaseous reaction products.
b: corrosion, oxidation of metals (formation of
nonvolatile oxides);
c: combustion of carbon black on gas switching
d: multi-step decomposition;
e: explosive decomposition with recoil effect.

Fig. 2. Example of a chemical reaction showing a gain in weight: In air, iron powder takes up 40% oxygen and forms Fe3O4 and Fe2O3. The SDTA signal confirms the strongly exothermic reaction. Heating rate
20K/min, 150l Al2O3 crucible. The reaction requires an exceptionally wide temperature range (about
600K) to achieve 1% to 99% conversion.

Other gravimetric effects


Many gravimetric effects have nothing to
do with chemical reactions or melting processes. The most frequently observed are
drying steps. These usually occur at the
beginning of the temperature program and
are about 100K wide (Fig. 6a). Desorption
of other substances such as solvent residues or monomers proceeds in much the
same way. Low molecular weight organic
compounds often show a tendency to sublime, i.e. they pass directly from the solid
phase to the gas phase.
Liquids evaporate in open crucibles over a
wide temperature range below their boiling point. If, however, a sealed crucible
with a small hole in the lid is used, a socalled self-generated atmosphere is created
in the crucible. The molecules of vapor
remain in equilibrium with the liquid
phase until the boiling point is reached.

Gravimetric effects on melting


In theory, you do not expect to observe a
gravimetric effect when a substance melts.
The difference in buoyancy due to the
slight change in the density of the sample

Fig. 3. Aspartame, an artificial sweetener, demonstrates relatively complex decomposition behavior. After
losing its water of crystallization at 130C, methanol is eliminated at 180C with the formation of a
piperazine ring. This reaction takes place over an extremely narrow temperature range of only 20K for
1% to 99% conversion. The SDTA curve shows that the piperazine derivative formed melts at 250C.

Fig. 4. Gravimetric effects on melting:


a: sample with low vapor pressure (no TGA
effect);
b: volatile melt (the liquid sample evaporates);
c: moisture escapes on melting;
d: sample melts with decomposition.

on melting is usually less than 1g. Despite this, the melting of a substance often
causes a noticeable change on the TGA
curve. This has to do with effects such as
increasing vapor pressure or more rapid
decomposition of the sample in the liquid
phase (Figures 4 and 5).

The liquid then evaporates completely over


a short period of time producing a sharp
TGA step whose onset corresponds to the
boiling point (Fig. 6b).
Ferromagnetic materials are paramagnetic
above their Curie temperature and can
therefore give rise to a TGA effect. The

3
UserCom 1/2001

Fig. 6. Other gravimetric effects:


a: drying, desorption, sublimation;
b: boiling in a sealed crucible with a small hole in
the lid;
c: ferromagnetic Curie transition without a magnet:
no TGA effect;
d: the same sample with a permanent magnet
below the furnace.
Fig. 5. Acetylsalicylic acid melts at 140C with decomposition followed by the elimination of acetic
acid. A weight loss of 33.3% is therefore expected. The measured weight loss is clearly greater due to
the simultaneous vaporization of other substances.

Fig. 7. A permanent magnet below the furnace of the thermobalance attracts the ferromagnetic material
(apparent increase in weight). The force is no longer exerted when the Curie temperature of the sample is
exceeded, but takes effect again on cooling. The abscissa shows the sample temperature Ts.

4
UserCom 1/2001

magnetic field of the earth does in fact


have a vertical component away from the
equator but this is too weak to produce a
measurable effect (Fig. 6c). It is best observed by placing a strong permanent
magnet at a cool place below the furnace.
The magnetic field exerts a downward
force on the ferromagnetic sample. This
causes an apparent increase in weight that
is lost when the sample is heated above the
Curie temperature (Figures 6d and 7).
Final remarks
The information presented above on the
origin and physical and chemical background of typical TGA effects should help
you interpret your own measurement
curves. If the interpretation of the results
is still unclear, you might well consider
using other techniques such as:
DSC and TMA measurements
the analysis of any gaseous compounds
evolved (EGA)
the observation of the sample under the
hot stage microscope.

New in our sales program


DMA/SDTA861e
METTLER TOLEDO is proud to introduce
its completely new Dynamic Mechanical
Analyzer (DMA). The instrument is based
on new ideas developed at the Institute for
Dynamic Material Testing in Ulm, Germany. The DMA/SDTA861e module expands
the very successful STARe range of thermal analysis instrumentation and offers
unique new measurement possibilities.
The DMA allows you to determine the elasticity and the damping properties of a material to an accuracy of a few percent. Besides this, it can measure physical transitions (e.g. secondary relaxation phenomena in polymers) that cannot be detected
with conventional measurement techniques.

Deformation modes
The instrument provides four different deformation measurement modes:
tension
bending (3-point bending and dual cantilever)
shear
compression
The newly designed sample holder system
allows you to rapidly change from one deformation mode to the other. The interchangeable sample holder inserts that fix
the sample in position are very simple to
install.

Loading the sample


The new sample holder inserts allow you
to load and measure samples very quickly.
The samples are loaded in the appropriate
insert outside the DMA. This saves time
because the next sample can be prepared
while a sample is being measured.
Stiffness range
The exceptionally high force and displacement resolution of the instrument coupled
with the exceedingly wide force and displacement range enables you to measure
samples in a stiffness range of more than 6
decades. This means that you can for example measure L SBR in the shear mode
(Fig. 2).
Frequency range
The frequency range covers 6 decades you really can measure the sample from
the mHz to the kHz range. The most accurate method is to measure a sequence of
frequencies isothermally. This eliminates
the effects of temperature gradients that
otherwise arise from the heat capacity and
finite thermal conductivity of the sample.
This measurement technique is advantageous when high frequencies are available
because the measurements are then performed very quickly. Low frequencies of
course require correspondingly longer
measurement times.
Force and displacement
measurement
METTLER TOLEDO has packed many new
ideas into the design of the instrument.
The modulus of elasticity is determined
according to the following equation:
E or G = FA / LA g

Fig. 1. The new DMA/SDTA861e Dynamic Mechanical Analyzer from METTLER TOLEDO.

FA / LA:
FA:
LA:
g:

stiffness
force amplitude
displacement amplitude
geometry factor (depends on the
geometry of the sample and the
deformation mode)
5
UserCom 1/2001

Fig. 2. L SBR measured in shear mode

The DMA/SDTA861e in fact measures both


displacement and force. Up until now, this
could only be done with large and expensive instruments for material testing.
Temperature adjustment
The very successful SDTA technology
used in the TGA and TMA instruments allows the temperature scale to be calibrated
and adjusted using the melting temperatures of pure metals and substances. The
sample temperature can also be detected
with great accuracy.
Software
The V7.0 STARe SW provides the facilities
for:
creating single frequency, multifrequency and frequency series methods
logarithmic display
evaluations in the various log displays
(log-log, lin-log, log-lin)
master curve construction (time temperature superposition)

6
UserCom 1/2001

Overview of the DMA/SDTA861e specifications:


Temperature range
Force measurement
Force range
Length measurement
Dynamic length range
Frequency range
Maximum sample geometry
Bending, dual cantilever
Tension
Shear
Compression
External sample mounting
Options
Force range
Maximum frequency
Stiffness range
Gas controller

-150C ... 500C


Piezoelectric crystal force transducer
1 mN ... 40 N
LVDT
10 nm ... 2 mm
1 mHz ... 1000 Hz
5 x 15 x 100 mm
2 x 7 x 18 mm
3 x 18 mm (diameter)
15 x 15 mm (diameter)
In removable sample holder inserts

12, 18, 40 N
200, 1000 Hz
4, 6 decades
Gas switching and/or controlled gas flow

Applications
Determination of the fiber content of composite
materials by thermogravimetry
Dr. B. Benzler, Applications laboratory, METTLER TOLEDO, Giessen, Germany

Introduction
Plastics are frequently filled or reinforced
with very different types of material in order to improve their mechanical and thermal properties. Organic fillers and reinforcing materials (e.g. wood flour) increase the toughness of a plastic. The addition of fibers can result in a major increase in stiffness and structural strength.
Besides natural organic fibers such as jute
and sisal, synthetic inorganic fibers (e.g.
glass and carbon fibers) and organic fibers
such as aramid are widely used for reinforcement purposes. Aramid fibers consist
of poly-p-phenylene terephthalamide and
are remarkable because of their high tensile strength and their relatively high decomposition temperature of about 550C.
Fiberglass-reinforced thermoplastics are
being increasingly used for technical products: they can be processed by injection
molding or extrusion and exhibit excellent
mechanical properties. This allows them to
be used in very diverse fields (car manufacturing, precision machines, electrical
engineering etc.).
Some examples of products reinforced with
aramid fibers are high pressure flexible
hoses, belts and bulletproof jackets.
The quality assurance of such composite
materials consists primarily of checking
the desired fiber content. This is very easily
done using thermogravimetry, as will be
shown in the following two examples.
Thermogravimetry (TG) or
thermogravimetric analysis (TGA) measures the mass of a sample that is heated
at a constant rate (usually linear) in a
controlled atmosphere. The examples de-

scribed here deal with the loss of mass resulting from the pyrolytic decomposition
of the plastics under nitrogen.
Example 1: Fiberglass-reinforced
polyamide PA 6
Samples of about 11 mg each of PA 6
(without fiberglass) and fiberglass-reinforced PA6 were examined by TGA under
the following experimental conditions:
Instrument: TGA/SDTA 851e STARe System thermobalance
Heating rate: 10 K/min
Temperature
range:
25 C to 800 C
Atmosphere: Nitrogen, 50 ml/min.
The resulting measurement curves and
their evaluations are shown in Figure1.

The unfilled PA 6 sample loses about 2.5%


of its mass at about 200 C. This effect is
due to the loss of moisture and is well
known for polyamides. Between 400C
and 500C the sample decomposes almost
completely, i.e. it undergoes almost 100%
degradation
As expected, the fiberglass-reinforced
sample behaves similarly except that the
effects are reduced in proportion to the
filler content. The residue of 24.5% that
remains can be assigned to the fiberglass
content of the sample.
A quantitative content determination is
therefore a simple matter because the filler
(fiberglass) does not decompose under the
experimental conditions used, but remains
behind as a residue while the plastic undergoes complete degradation.

Fig. 1. Thermogravimetric analysis of Polyamide 6 with and without glass fibers

7
UserCom 1/2001

Fig. 2. TGA of an epoxy resin aramid fiber composite (middle): above: pure aramid fibers, below: pure
resin.

Example 2: Aramid fiber-reinforced


synthetic resin
The experimental conditions for the
thermogravimetric analysis of the individual components, i.e. the resin and aramid fibers, and the composite material
were similar to those used in Example 1:

DB in the last column means, dry basis,


i.e. with respect to the dry content at
300C.
The resin content of the composite dry
substance is therefore:

Instrument: TGA/SDTA 851e STARe System thermobalance


Heating rate: 10 K/min
Temperature
range:
25 C to 800 C
Atmosphere: Nitrogen, 200 ml/min.

Recalculated with the actual moisture


content of the composite:
Moisture: 2.1%, resin: 49.8% and as the
difference to 100%, aramid fiber: 48.1%.
The aramid content could also be calculated from the step between 520C and
640C, or from the residue at 640C. The
result is however not so accurate; the pyrolysis of the aramid is probably different
in the presence of the epoxy resin.

Aramid undergoes decomposition in contrast to the glass fiber used in Example 1.


The TGA curves in Figure 2 can be interpreted as follows:
The pure resin first loses moisture and
then pyrolyzes in a single step process between 300C and 520C.
The aramid fibers initially lose about 3.1%
moisture. Between 300C and 520C the
weight loss is only 1.3%. Decomposition
takes place above 520C.
The composite material first suffers a loss
of about 2% due to moisture and then, as
expected, decomposes in two steps.
Sample
Aramid fiber
Composite
Pure resin

8
UserCom 1/2001

Conclusions
These examples show that
thermogravimetry is an excellent method
for the rapid and accurate determination
of the resin and fiber content in composite
materials. The decomposition temperatures provide additional qualitative information concerning the identity of the
components.

Thermogravimetric steps in %
RT...300C
300...520C
3.12
1.31
2.07
48.91
0.94
96.05

300...520C, DB
1.35
49.94
96.96

Investigation of the segregation behavior of different


mortar constituents with TGA/SDTA
Roger Zurbriggen, Elotex AG, Industriestr. 17a, CH-6203 Sempach Station, Switzerland

Introduction
The mineral constituents of mortar mixtures can be divided into two main groups
- reactive binders and inert fillers. By
mineral binders, we usually mean cement,
e.g. Portland cement or aluminous (highalumina) cement, anhydrite and/or hemihydrate (calcium sulfate) that hydrate
when mixed with water. The hydrates form
a solid matrix that binds the mortar together. Quartz or carbonate-rich sand and/
or their finely crushed powder are used as
fillers. Other possible mortar constituents
are latent hydraulic binders, glasses, lightweight fillers and a large number of organic additives such as cellulose ethers
and redispersible powders.
Fresh mortar is made by adding water and
mixing well. The particular working consistency of the mortar depends on the application. When used as an adhesive for
wall tiles, the shear strength and viscosity
should be on the high side. For self-leveling flooring, the mass may segregate to
certain degree. The mortar must remain
homogeneous, i.e. the coarse mineral constituents should not settle, and the water
with the dissolved and dispersed fine components should not rise to the surface too
much.
This article describes how the segregation
behavior of freshly made mortar was
quantitatively analyzed using TGA/SDTA. A
simple tile adhesive recipe was used as an
example.
Sample preparation
The tile adhesive consists of 40% Portland
cement (CEM I 52.5 R, JCF, Wildegg),
59.6% quartz sand (0.1-0.3 mm, Zimmerli
Mineralwerke, Zrich) and 0.4%
methylhydroxyethyl cellulose (MHEC
15000 PFF from Aqualon). 100 g of the dry
mixture were stirred with 23 g of water and
transferred to a concrete slab with a suitable tool. After five minutes, a 50x50 mm
earthenware tile was placed on the mortar
and a load of 2 kg applied. This compressed the mortar layer to 1.6 mm.

After setting, the tile was broken off and


samples of the mortar were taken sequentially along a vertical profile from the tile
surface to the concrete slab. The weight of
each sample compared with the total
weight of all the samples allowed the distance along the 1.6mm long profile to be
approximately estimated.
Measurements with the TGA/SDTA851e
The TGA curves show three significant
weight loss steps. The first at about 100C
has to do with the drying (capillary pore
residual water) and/or with the dehydration of ettringite. The first step is not
evaluated here because of the uncertainty
in the interpretation. The second step at
about 450C is due to the dehydration of
Ca(OH)2 (portlandite). Notice the reduced
portlandite content in the recipe containing MHEC. This indicates a reduced degree
of hydration, probably in connection with
the effect of the cellulose ether, which retards setting. The third weight loss step at
about 700C can be attributed to the decarbonation of CaCO3. The carbonate is

however not present in the original mixture and must therefore arise from a
carbonisation reaction. This interpretation
is supported by the fact that the carbonate
content correlates with the portlandite
content and has a tendency to increase
toward the concrete slab.
The SDTA curves show that the three
weight loss steps each correspond to endothermic processes. In addition there is an
exothermic process at about 260C and
another endothermic process at about
570C. These two peaks correspond to the
strongly exothermic decomposition of
MHEC and the solid-solid transition of
quartz. The MHEC and quartz content can
therefore be quantitatively determined by
evaluating the SDTA signal. The decomposition of MHEC in fact also causes a weight
loss. However, with low MHEC concentrations (in this case 0.4%), this weight loss
step is completely hidden by the drying
and dehydration processes of the ettringite.
In this case, only the quantitative evaluation of the exothermic SDTA peak allows
reliable information on the MHEC content

Fig. 1. The TGA and SDTA curves of three samples. The dotted curve shows the results of a TGA analysis
of a reference mortar sample consisting of Portland cement and quartz sand that did not contain any
MHEC. The uppermost sample in contact with the tile (continuous line) and the sample in the middle of
the mortar bed (dashed line) are shown as examples of the 5 individual samples.

9
UserCom 1/2001

to be obtained. To do this, it is of course


necessary to measure samples containing
different known amounts of MHEC and to
construct a calibration curve. A similar
calibration curve allows the weak endothermic - to -quartz transition at 573C
to be used for quantitative purposes.
Quantification of the MHEC and
quartz content using calibration
curves
To obtain the necessary calibration curves,
samples containing different known
amounts of constituents in question have
to be analyzed beforehand. As an example,
Figure 2 shows the SDTA curves of 6
samples containing different percentages
of quartz.
A plot of the quartz content as a function
of the SDTA peak area normalized to the
sample weight yields the calibration function shown in Figure 3 needed for the determination of the quartz content from
SDTA data. Frequently, in the first heating
measurement, organic residues also oxidize whose combustion enthalpy is many
times larger than the approximately 8 J/g
liberated in the solid-solid quartz transition. Even quartz contents over 50% are
then difficult to determine with the SDTA
peak area of the a- to b-quartz transition.
Each sample should therefore be measured
a second time at the same heating rate
used to obtain the quartz calibration
curve. In the example described here, each
sample was first measured from 30C to

1000C at 10 K/min in air (50ml/min).


For the quartz determination, the residue
from the first measurement was cooled
and then measured again from 500C to
620C at 10K/min under N2 (50ml/min).
It is important to note that each type of
quartz leads to a different, but in each case
very reproducible, calibration curve. For
this reason, a calibration curve obtained
with one type of quartz should never be
used to quantify a different type of quartz.
Figure 4 shows the calibration curve for
mortar with different MHEC contents. The
calibration curve was obtained in the same
way as that for the quartz content. Both
calibration curves show correlation coefficients of 0.99.
The segregation behavior of
individual mortar constituents
If the contents of MHEC, portlandite, calcium carbonate, quartz sand and Portland
cement evaluated from the 5 TGA/SDTA
measurement curves are plotted as a function of the position of the sample along
the vertical section through the 1.6 mm
thick sample of tile adhesive, the relationships shown in Figures 5, 6 and 7 are obtained.
The distribution curve for MHEC (Fig. 5)
shows a clear enrichment effect towards
the tile and the concrete slab. In principle,
there are several fractionating mechanisms: the surface activity of MHEC could

Fig. 2. SDTA curves of six different dry mixtures with known quartz/cement ratios. The peaks are integrated over the same temperature range.

10
UserCom 1/2001

Fig. 3. Calibration curve for the determination of


the quartz content using SDTA peak areas

Fig. 4. Calibration curve for the determination of


the MHEC content from SDTA measurements

cause enrichment on the surface of the


mortar even before laying the tile. On the
other hand, MHEC is dissolved in mortar
water and migrates with the pore water at
the evaporation front (mortar surface before laying the tile) or towards a porous
tile and the concrete slab, both of which
are highly porous materials that absorb
well.
The concentrations of quartz sand and
Portland cement show a variation that is
complimentary within a range of about 6
wt %. No clear enrichment effect is noticeable (Figure 6). The viscosity of the fresh
mortar apparently prevents sedimentation
of the quartz sand. Since the ignition residues of all the samples vary within a small
range of 92.7 to 93.6 wt %, an enrichment
of quartz sand has a direct effect on the
concentration of the second main constituent, the cement.
The distribution profiles for Ca(OH)2 and
CaCO3 (Fig. 7) show a slight increase of
the content of both phases toward the concrete slab. Ca(OH)2 and CaCO3 enrichment
can on the one hand result from primary
cement enrichment and on the other hand
from a different degree of hydration and
carbonate formation. An additional complication in the interpretation is that carbonate formation occurs at the expense of
Ca(OH)2, but this does not occur everywhere to the same extent. In general, how-

dehydration reactions of different hydrates


(C-S-H phases?). In the case described,
however, the relatively strong exothermic
decomposition of the small amount of
0.4% MHEC produces a clear SDTA signal.
Using a calibration curve set up beforehand, the MHEC content can be determined with sufficient accuracy even for
low MHEC contents using the SDTA peak
area.
Fig. 5. Depth profile for MHEC in a tile adhesive

Fig. 7. Depth profiles for Portlandite and calcium


carbonate in a tile adhesive

ever, one notices a parallel relationship


between the concentrations of Ca(OH)2
and CaCO3, which means that hydration
and carbonate formation accompany one
another. The fact that the distribution

mixture. The rest is free pore water and


bound H2O and CO2 from the various hydrates and carbonates. Corresponding to
the known composition of the dry mixture,
the mortar after setting should contain
0.38% MHEC, 38% cement and 56% quartz
sand. The comparison with the values
determined from the TGA/SDTA measurements shows a relative deviation of 16% in
the case of MHEC and deviations of about
10% for the cement and quartz contents.
Taking the low content of MHEC into account, the results can be regarded as being
sufficiently accurate.

Fig. 6. Depth profile for the quartz and Portland


cement content in a tile adhesive

curves of Ca(OH)2 and CaCO3 do not exactly follow the cement distribution is evidence for the presence of different local
degrees of hydration.
The concentrations by weight of the different constituents of the mortar in the original mixture can be determined by averaging the concentrations of the different
mortar constituents along the depth profile. The ratio of the sample weight to the
total weight of all the samples is used as
the weight factor for the calculation of the
mean value for each sample. On the basis
of the TGA/SDTA measurements described
here, one finds values of 0.44% MHEC,
41.5% cement and 51.7% quartz sand for
the composition of the original mortar

Summary and conclusions


To investigate the segregation behavior in
the fresh mortar state, several samples
were taken along a depth profile of the
mortar after setting. The composition of
the samples was determined quantitatively
using TGA/SDTA, which allowed the heterogeneous distribution curve of each
mortar constituent to be determined over
the entire length of the profile.
Hydrates (e.g. portlandite) and carbonates
can be stoichiometrically quantified via
each weight loss step of their dehydration/
carbonate decomposition. In principle
organic additives such as
methylhydroxyethyl cellulose (MHEC) can
also be quantified via the weight loss step
of their exothermal combustion reaction.
The relative amounts of organic additives
are frequently so small that the corresponding weight loss steps are overlaid by

Quartz fillers and cements are thermally


stable in the temperature range of the
measurements and together make up the
ignition residue. The quartz content can
also be quantified using the endothermic
transition from - to -quartz at 573C
by means of a calibration curve. The cement content is obtained from the difference between the ignition residue and the
quartz content.
In addition the inhomogeneous distribution of substances such as the enrichment
of MHEC toward the tile and concrete slab
can be clearly shown using the depth profiles for different mortar constituents obtained from TGA/SDTA measurements.
All in all one can say that TGA/SDTA is a
very good method for the analysis of mortar. If the mortar systems are not too complex, all the important constituents can be
quantitatively or semi-quantitatively determined. In addition the inhomogeneous
enrichment of substances can be detected,
which in turn yields information on different fractionization processes in the fresh
mortar state (sedimentation, surface activity and migration with the pore water) and
also after setting (different local degrees of
hydration and carbonate formation).
Acknowledgements
I would like to thank Thomas Armbruster
(University of Bern) for many valuable
technical discussions. The measurements
were performed with a METTLER TOLEDO
TGA/SDTA851e in the F&E laboratory of
Elotex AG in Sempach Station, Switzerland.

11
UserCom 1/2001

Determination of polymer crystallinity with DSC


measurements
Dr. Markus Schubnell

Introduction
A large number of polymers are able to
form crystalline structures in which parts
of the macromolecules are orientated parallel to one another. Unlike low molecular
weight substances, the degree of crystallinity that can be achieved by polymers is
much lower than 100 percent and depends
on the molecular structure. Besides crystallites, amorphous regions can also be
formed in these materials. In practice one
distinguishes between two kinds of amorphous regions in partially crystalline polymers based on the different types of molecular mobility possible. The mobile
amorphous regions are between the crystallites. They also determine the step
height of the glass transition. At the surface of the crystallites there are rigid amorphous regions that exhibit an amorphous
structure but despite this do not take part
in the glass transition because of their reduced mobility (see UserCom 11). The
temperature at which crystallization occurs, how rapidly it takes place and the
degree of crystallinity achieved, depend on
the molecular structure of the sample.

known, the crystallinity of a sample can be


calculated from the area of the melting
peak. Table 1 summarizes the typical values of the melting enthalpies for a number
of completely crystalline polymers.
The crystallinity of partially crystalline
polymers can be investigated by DSC. A
simple method for doing this is described
in the following article.

The size of the crystallites formed during


crystallization depends on how easily the
polymer chains fit into the crystal structure. In general, polymer chains are less
mobile at lower temperatures and only
small, less stable crystals are produced.
These crystals have a low melting point.
The mobility of the molecules is greater at
higher temperatures so that larger, more
perfect crystals are formed that melt at
higher temperatures. The melting curve of
a partially crystalline polymer therefore
contains information on the size distribution of the crystallites present in the material. If the melting enthalpy of a 100 percent crystalline material (Hf 100 %) is

As explained above, to determine the initial crystallinity the measured melting enthalpy, Hf, is compared to the value for a
100 percent crystalline sample, Hf100%.

12
UserCom 1/2001

Determination of crystallinity from


DSC curves
Figure 1 shows the melting curves of
samples of polyethylene (PE-HD,
Lupolen), polyamide 6 (PA 6, Durethan)
and polyethylene terephthalate (PET). The
melting behavior of these three samples is
very different. The melting peak of PA 6,
which begins at about 190C, covers a
range of about 40K. In comparison, PEHD shows a very broad melting peak that
begins below 60C and ends at about
150C. Unlike the other two curves, the
measurement curve of PET exhibits an
exothermic peak before the endothermic
melting peak. This first peak is due to the
crystallization of amorphous material.

Material
PE-LD, polyethylene, low density
PE-HD. polyethylene high density
PET, polyethylene terephthalate
PP, polypropylene
PA 6, polyamide 6
PA 66, polyamide 66
PTFE, polytetrafluoroethylene
Table 1. Melting enthalpies of different thermoplastics

The crystallinity, , is given by the equation

The problems that arise are:


the value of Hf100% has to be known
where to set the limits for the integration of the area under the melting curve
in order to determine the melting enthalpy, Hf
baseline type for the determination of
Hf
possible structural changes during the
measurement
The melting enthalpy, Hf100%, is the difference between the enthalpy curves of the
completely amorphous material and the
pure crystalline material. Hf100% cannot
be obtained directly in an experiment but
is calculated from the structural data of
the crystallites that have been determined
using X-ray diffraction. In general, Hf100%
is a function of temperature. The literature
data for PET are shown in Figure 2 to illustrate the temperature dependence of the
melting enthalpy.
The temperature dependence of the melting enthalpy in the observed melting
range is however often so low that it can
be neglected in comparison to the meaHf 100 %(J/g)
293
293
140
207
230
255
82

ference significantly deviates from the


baseline and leads to peaks depends on the
molecular structure and the
macrostructure of the material that is
already present.

Fig. 1. Melting curves of PET, PE-HD (Lupolen) and PA 6 (Durethan).

The melting peak also has to be evaluated


with materials that recrystallize in order to
determine the initial crystallinity. The
thermal effects that occur as a result of the
recrystallization must of course be subtracted from the total melting enthalpy.
This is best done by integrating over the
entire crystallization and melting range.
The linear extrapolation of the measurement curve in the region of the melt is
used as baseline (i.e. at temperatures
above the melting peak). The method is
illustrated in Figure 3.
Conclusions
The crystallinity of a sample can be estimated from its melting behavior. The
melting enthalpy of a pure fictive 100 percent crystalline reference sample presents
a problem. In addition, the result is greatly
influenced by the choice of the baseline
and the integration limits. To obtain reproducible results, it is best to extrapolate
the baseline in the melt to the region immediately after the glass transition.

surement accuracy. In practice a constant


value for Hf100% is normally used (see
Table 1).
To determine the melting enthalpy, it is
necessary to measure the entire melting
range. This means that the integration
limits have to be set accordingly. This is
not always a simple matter for polymers
such as polyethylenes that melt over a
wide temperature range. In this case, the
sample should be measured over the widest
possible temperature range and the integration limits set from immediately after
the glass transition temperature (if one
can be detected) to a temperature at which
the polymer has completely melted. Besides the integration limits, the baseline
type also plays an important role. Difficulties arise here because recrystallization
and cold crystallization processes take
place that lead to exothermic peaks over-

lapping the endothermic melting process.


Amorphous material can crystallize immediately after the glass transition. It is also
possible that small crystals melt at these
low temperatures and, depending on the
conditions used, recrystallize again as
shown in the example in Figure 3. The
heat flow measured by the DSC corresponds in any case to the difference between the crystallization and melting processes. The temperature at which this dif-

Fig. 2. Calculated melting enthalpies for PET as a


function of temperature.

Fig. 3. Determination of the initial crystallinity of partially crystalline materials. Small crystals melt immediately after the glass transition and recrystallize to form larger, more stable crystals that then melt at
higher temperatures. Before the measurement, the sample (PET) was completely melted and annealed at
120C for 40 minutes.

13
UserCom 1/2001

DSC characterization of short chain branching


distribution in polyethylene
D. DAOUST, J.J. BIEBUYCK, V. STEPHENNE and R. LEGRAS, Universit Catholique de Louvain (UCL), Laboratoire des hauts polymres, Place Croix
du Sud, 1, B 1348 Louvain-la-Neuve, BELGIUM

Introduction [1 9]
The following acronyms are used in this
article:
MMD Molar Mass Distribution
SCBD Short Chain Branching Distribution
SCB Short Chain Branching
TREF Temperature Rising Elution Fractionation
SIST Stepwise Isothermal Segregation
Technique
SSA Successive Self-nucleation / Annealing
The properties of the copolymers of ethylene and higher -olefines (1-butene, 1hexene, 1-octene, 4-methyl-pentene, etc.)
are closely related to the molar mass distribution (MMD) and the short chain
branching distribution (SCBD) within the
macromolecule.
The MMD and the type and content of
monomers (or branching) can be determined by Size Exclusion Chromatography
(SEC) and Nuclear Magnetic Resonance
(NMR). The determination of the SCBD
requires the use of fractionation techniques. For instance, the monomer distribution along the molecular chains can be
determined by Temperature Rising Elution
Fractionation (TREF). This technique
fractionates the heterogeneous polyethylene (PE) by stepwise crystallization in solution onto an inert support. The support
temperature follows the crystallization
temperatures of the different molecular
structures of the PE chains. The method
involves dissolving the polymer in a suitable solvent at high temperature at 130C
to 140C. The dilute polymer solution is
then cooled slowly to room temperature.

14
UserCom 1/2001

Afterward the temperature is increased in


steps. Fractions are thereby formed that
can then be analyzed with the standard
methods used for the determination of the
characteristics of molecular structures.
Although TREF has been successfully applied, the main disadvantages of this fractionation technique are the relatively high
costs involved and the time needed to perform an experiment. Several research
groups have therefore tried to develop simpler and more rapid methods for the
analysis of polyethylenes and -olefinecopolymers using Differential Scanning
Calorimetry, DSC, even though the results
of these methods are not as accurate as
those from TREF analysis.
This study compares two DSC methods for
estimating the SCBD of copolymers following appropriate thermal treatment. The
first method is known as the Stepwise Isothermal Segregation Technique (SIST).
This involves the stepwise fractionation of
the sample through crystallization by
cooling the sample in a series of isothermal steps from the melt analogous to the
TREF technique. SIST is mainly used for
polyethylenes (PE). Up until now no sigDensity (g/cm )
SCB content b(/10000C)
Mnc
Mwc
Mzc
Hd
a

PE1
0.958
e
26000
161300
745600
6.2

nificant results have been obtained with


polypropylenes (PP). The second method
is based on a superposition of self-nucleation and annealing cycles and is known
as Successive Self-nucleation / Annealing
(SSA). SSA is a more recent method than
SIST and is also used for the characterization of polyethylenes. In this paper we
compare conventional DSC melting analysis with the SIST and SSA methods for the
characterization of three polymers: a
homopolyethylene and two ethylene /1hexene copolymers with different proportions of butyl groups.
Experimental details
Materials
Three different polyethylenes were investigated: a homopolyethylene, PE1, (essentially linear PE) polymerized using a
Ziegler-Natta catalyst and two ethylene-1hexene copolymers, PE2 and PE3,
(branched PE) polymerized using supported Cr catalysts. The density, SCB content and the average molar masses of the
materials are given in Table 1.
PE2
0.943
25
17900
214200
1417600
12.5

Table 1. Main characteristics of the three PE samples.


a: at 23C
b: butyl content measured with 13C-NMR
c: number-average, weight-average and average molar masses (g/mol)
d: dispersion (H = Mw/Mn)
e: below the NMR detection limits

PE3
0.933
68
14800
205500
1460200
14.1

Differential Scanning Calorimetry


(DSC)
Samples of about 10 mg were measured in
40 l aluminum crucibles with pierced
lids using a METTLER TOLEDO DSC821e
equipped with a liquid nitrogen cooling
system. The instrument was purged with
high purity nitrogen.
First procedure: conventional DSC
analysis:
To eliminate thermal history, each
sample was heated to 220C and the
melt held at this temperature for 10
minutes. The sample was then cooled
from 220C to -20C at 10K/min, annealed at 20C for 2minutes and the
DSC curve measured to 220C at
10K/min.
Second procedure: Stepwise Isothermal
Segregation Technique (SIST) (Fig. 1).
Each PE sample was heated to 220C,
the melt held at this temperature for 10
minutes and then cooled to the first crystallization temperature (Tc = 130C) at
50K/min. The sample was then allowed
to crystallize isothermally for 60 minutes at this temperature. Afterward it was
rapidly cooled at 50 K/min to the next
crystallization temperature 5K lower
than the preceding crystallization temperature and again allowed to crystallize
for 60 minutes. This procedure was repeated for successive 5K intervals down
to a temperature of 40C. Each sample
thereby underwent a series of 19 isothermal steps. Finally the sample was cooled
to -20C at 20K/min and the DSC curve
measured from -20C to 220C at
5K/min.
Third procedure: Successive Self-nucleation / Annealing (SSA) (Fig. 2)
Each sample was heated to 220C, the
melt held at this temperature for 10
minutes and then cooled at 10K/min to
-20C to obtain the initial reference
state. The sample was then heated at
10K/min to the first annealing temperature (Ta=130C). It was held isothermally at this temperature for 60
minutes and then cooled to -20C at
10K/min. After this first thermal cycle

the sample was again heated to 125C


(Ta -5K) at 10K/min and held at this
temperature for 60 minutes before cooling it again to -20C at 10K/min. This
procedure was repeated lowering the
annealing temperature successively in
steps of 5K for each cycle down to
40C. Each sample thereby underwent a
series 19 annealing steps. Finally the
sample was cooled to -20C at 20K/min
and the DSC curve measured from
-20C to 220C at 5K/min.
Results
The DSC curves of the three PE samples
obtained using the first procedure are
shown in Fig.3a (heating run) and Fig.
3b (cooling run). The melting and crystallization temperatures (Tm and Tcc) and the
corresponding enthalpies are listed in
Table 2.
An analysis of the data in Table 1 and Figures 3a and 3b leads to some interesting
conclusions:
The crystallization and melting peaks
are asymmetric.
The differences in the thermal behavior
of the materials are due to their short
chain branching (SCB) content and
their short chain branching distribution
(SCBD). The presence of butyl side

chains in the linear PE chain corresponds to a structural defect that is not


incorporated in the crystal lattice. The
melting temperatures Tm and the crystallization temperatures Tcc observed for
PE2 and PE3 are therefore lower than
the values for the linear polymer PE1.
The depression of the melting temperature is generally independent of the
length of the side chain if the side chain
consists of ethyl or longer groups [10].
If the SCB content increases, the crystallization and melting enthalpies decrease.
All three PE samples have a broad melting and crystallization range. In addition, PE2 and PE3 actually exhibit two
crystallization peaks. The second crystallization peaks of PE2 and PE3 are
however very small compared to the
main peak. This low temperature peak
(T = 78C for PE2 and T = 72C for
PE3) is always found in polymers containing ethylene groups and depends on
the SCB content: the higher the SCB
content the more this peak is shifted to
lower temperature.
The shapes of the PE1 crystallization
and melting peaks indicate the presence
of defects in the chain (short or long
side chains), although at a very low
level.

Fig. 1. Temperature program used for SIST

15
UserCom 1/2001

The SIST and SSA treatments were performed on all three PE samples. The resulting melting curves are shown in Figures 4 and 5. Table 3 summarizes the
melting enthalpies obtained following the
three different sample pretreatments (conventional DSC, SIST and SSA). The use of
SIST and SSA pretreatments allows the
three PE samples to be classified into two
groups. With PE2 and PE3 copolymers several more or less resolved peaks appear in
the DSC curves. As expected, this is not the
case with the PE1 homopolymer.

Fig. 2. Temperature program used for SSA

Fig. 3a. DSC measurements of the three different PE samples (heating experiment at 10 K/min)

The differentiation of the different PE


types following the SIST and SSA treatments is based on their differing ability to
crystallize due to the differences in the side
chain branching distribution (SCBD). The
multiple peaks observed after the SIST and
SSA treatments therefore reflect the different crystals that occur at the different crystallization temperatures due to the distribution of the butyl groups. In fact, monomolecular chains with several butyl groups
contain shorter ethylene sequences that are
able to crystallize. These chains therefore
form thinner lamellae and consequently melt
at lower temperatures. In contrast, chains
with less butyl groups or longer ethylene sequences melt at higher temperatures.
Of the different experimental conditions
used in this work, SSA leads to a better
separation than SIST.
Conclusions
The results of the work presented in this
article show that the SIST or SSA thermal
treatments of PE with different SCB contents followed by DSC measurements are
able to provide qualitative information on
the molecular architecture of these copolymers.
Of the different experimental conditions
used in this work, the SSA treatment gives
the best fractionation on crystallization.
In any case, SIST and SSA analyses with
DSC allow polymers to be more rapidly
characterized than TREF.

PE1
PE2
PE3
Fig. 3b. DSC measurements of the three different PE samples (cooling experiment at 10 K/min)

16
UserCom 1/2001

Conventional DSC
216
188
174

SIST

SSA

234
186
178

224
192
166

Table 3. Melting enthalpy, Hm, in J/g of PE1, PE2


and PE3 determined after the different thermal
pretreatments (DSC, SIST and SSA).

Tm (C)
PE1
PE2
PE3

138.0
129.0
126.0

Tcc (C)
First peak
115.0
115.0
113.5

Tcc (C)
Second peak
80.5
74.0

Hm (J/g)

Hcc (J/g)

216
188
174

216
183
172

Table 2. Thermal characteristics of samples PE1, PE2 and PE3 determined by conventional DSC analysis
(first procedure)

Fig. 4. DSC curves after SIST treatment (heating rate: 5K/min).

Acknowledgement
The authors gratefully acknowledge the
help given by Mrs. Batrice van der Heyden
in performing the DSC measurements.
Literature
[1] T. Kamiya, N. Ishikawa, S. Kambe, N.
Ikegami, H. Nishibu and T. Hattori,
Antec90, 871-873.
[2] D.R. Parikh, B.S. Childress, and G.W.
Knight, Antec91, 1543-1545.
[3] E. Adisson, M. Ribeiro, A. Deffieux,
and M. Fontanille, Polymer, 1992,
volume 33, number 20, 4337-4342.
[4] M.Y. Keating, and E.F. Mccord,
Thermochimica Acta 243, 1994,
129-145.
[5] A.J. Muller, Z.H. Hernandez, M.C.
Arnal, and J.J. Sanchez, Polymer bulletin, 1997, 39, 465-472.
[6] A. Hanyu, and R. Wheat, Antec98,
1887-1891.
[7] C. Vandermiers, J.F. Moucin, P.
Damman, and M. Dosiere, Polymer,
2000, 41, 2915-2923.
[8] G. Glokner, Journal of applied polymer science: Applied polymer symposium 45 ,1990, 1-24.
[9] V. Stephenne, D. Daoust, G. Debras,
M. Dupire, R. Legras, and J. Michel,
submitted in journal of applied polymer science.
[10] R.G. Alamo, L. Mandelkern,
Thermochimica Acta 238, 1994, 155.

Fig. 5. DSC curves after SSA treatment (heating rate: 5K/min).

17
UserCom 1/2001

Tips
TMA/SDTA840: Rapid cooling between measurements
Dr. Markus Schubnell

Introduction
It takes a certain amount of time for the
TMA to cool down from the end temperature of an experiment to the initial temperature of the next measurement. To a
good approximation, the cooling behavior
can be described by an exponential function characterized by the cooling time
constant, . If the cooling time constant is
known, the time that the module needs to
cool down from a temperature of T1 to a
temperature of T2 is given by the following
equation:
tT1T2 = [ln(T1T0) ln(T2T0)]
where T0 is the lowest temperature that is
reached asymptotically. It lies between the
cryostat temperature and room temperature. The temperature of the cryostat
should be set to 22.0C as recommended
in the operating instructions. If the temperature of the cooling flange is below
16C or above 45C, the measurement is
automatically terminated and an error
message is displayed.
In principle, there are several ways to reduce the cooling time constant. These are
described in the following sections and are
discussed briefly in connection with typical
measurements.
Opening the furnace at the end of
the measurement
If the furnace is opened at the end of a
measurement, the surrounding air is free
to enter the furnace and reach the sample.
The cooling time constant is however only
reduced by about 20 s, which therefore
only has a small effect on the furnace
cooling time. On the other hand, the
sample itself cools much more rapidly
(cooling time constant about 50 s), which
may be desirable in certain circumstances.

18
UserCom 1/2001

So that the furnace can be opened immediately after it reaches the end temperature, the Furnace Open Permission must
have been previously activated in the experiment buffer.
Cooling with helium
At first sight, helium seems particularly
suitable for use as a purge gas because its
thermal conductivity is about six times
better than that of air or nitrogen. With a
helium flow rate of 30 ml/min, the cooling time constant is about 430s. This
cooling time constant is in fact only
slightly less than the cooling time constant obtained using a nitrogen flow rate
Purge gas Purge gas [ml/min]
air
0
air
0
air
2000
nitrogen
30
helium
30

space between the cooled jacket and the


furnace filled with helium. At low flow
rates, the gas itself removes only a relatively small amount of heat, so that the
cooling time constant is hardly affected
(compare also the cooling time constant
without gas: 470 s).
High gas flow at the end of the
measurement
Purging the TMA with a flow rate (purge
gas inlet) of about 2 l/min at the end of
the measurement reduces the cooling time
constant to about 320 s. A gas flow of
1-3 l/min can be obtained very conveniently and cheaply using an aquarium

Furnace condition T0 [C] Cooling time constant [s]


closed
23.5
470
open
23.5
450
closed
23.5
320
closed
23.5
435
closed
23.5
430

Table 1. Cooling time constants for different cooling conditions

of 30 ml/min (435 s). The reason for this


is that the purge gas does not improve the
thermal coupling between the furnace and
the water-cooled furnace jacket. To cool
the furnace efficiently, the TMA would
have to be evacuated beforehand and the

Fig. 1. TMA 840 cooling curves with and without


purge gas; purge gas flow rate 2 l/min

pump (costs about $30). We have also proven


that this simple cooling technique does not
have any adverse effects on the next measurement. Figure 1 shows the cooling curves obtained. The TMA was heated at 50 K/min to
1000C, held isothermally at this temperature for five minutes and then cooled
down. The furnace temperature during
cooling is shown with and without the use
of additional purge gas.
Summary
Table 1 summarizes the cooling time constants obtained for different cooling methods. The table shows that a high purge gas
flow rate is the most effective means of
cooling the TMA840. Tables 2 and 3, show,
in addition, the calculated cooling times

Initial temperature T1 [C]


End temperature T2 [C]
25
30
50
70
100

100

150

200

300

500

700 1000

21.7
16.3
7.7
3.7
0.0

25.5
20.1
11.5
7.5
3.8

28.1
22.6
14.0
10.0
6.4

31.5
26.1
17.5
13.5
9.8

35.8
30.3
21.8
17.7
14.0

38.5
33.1
24.4
20.4
16.8

41.3
35.9
27.3
23.3
19.6

Table 2. Cooling times in minutes for natural cooling (no purge gas); =7.83 min, T0=20C

Initial temperature T2 [C]


End temperature T2 [C]
25
30
50
70
100

100

150

200

300

500

700 1000

14.8
11.0
5.2
2.5
0.0

17.4
13.7
7.8
5.1
2.6

19.1
15.4
9.6
6.8
4.3

21.5
17.8
11.9
9.2
6.7

24.3
20.6
14.8
12.1
9.6

26.2
22.5
16.6
13.9
11.4

28.1
24.5
18.6
15.9
13.4

for different start and end temperatures


with no flow rate or a flow rate of air of
2 l/min. The following values were used:
for , 7.83 min (with no purge gas) and
5.33min (with 2 l/min purge gas flow);
and for T0 20C.
Remarks on the TGA/SDTA851e
The above results also apply in principle to
thermogravimetry. We have however noticed that, at high purge gas flow rates, hot
gas from the furnace penetrates the housing of the microbalance and causes the
balance to become warm. This in turn
leads to an unacceptable balance drift,
which adversely affects later measurements (about 0.1 mg/h).

Table 3. Cooling times in minutes with a purge gas flow rate of 2 l/min air; =5.3 min, T0=20C

Dates

Exhibitions, Conferences and Seminars - Veranstaltungen, Konferenzen und Seminare


Eighth International Conference on
Composites Engineering, ICCE8
ESTAC 8
222nd American Chemical Society
National Exposition
MEDICTA 2001
GEFTA/Eurostar-Science
29th Annual North American

August 5-11, 2001 Tenerife (ES)


Aug. 25-29, 2002 Barcelona (ES)
Aug. 27-29, 2001 Chicago, IL (USA)
Sept. 2-7, 2001 Corua (ES)
11-13.9.2001 Mnchen (DE)

Thermal Analysis Society


Symposium Thermische Analyse
The Belgium Polymer Group meeting
TAWN Meeting Benelux
RICH-MAC Mostra Internazionale
delle Apparecchiature Chimiche
K - Kunststoffmesse
Eastern Analytical Society

Sept. 24-26, 2001 St. Louis, MO (USA)


May 4, 2001 Diepenbeek (BE)
May 16-17, 2001 Mol (BE)
October 4-5, 2001 The Nederlands (NL)
October 2-6, 2001 Milano (IT)
25.10-1.11 Dsseldorf (DE)
Nov. 4-11, 2001 Atlantic City, NJ (USA)

TA Customer Courses and Seminars in Switzerland- Information and Course Registration:


TA-Kundenkurse und Seminare in der Schweiz - Auskunft und Anmeldung bei:
Helga Judex, METTLER TOLEDO GmbH, Schwerzenbach,
Tel.: ++41-1 806 72 65, Fax: ++41-1 806 72 40, e-mail: helga.judex@mt.com

TMA/DMA (Deutsch)
10. September 2001 Greifensee
S6)Re SW Workshop Basic (Deutsch) 10. September 2001 Greifensee
TGA (Deutsch
11. September 2001 Greifensee
DSC Basic (Deutsch)
12. September 2001 Greifensee
DSC Advanced (Deutsch)
13. September 2001 Greifensee
S6)Re SW Workshop Advanced (Deutsch) 14. September 2001 Greifensee
TMA/DMA (English)
September 17, 2001 Greifensee
S6)Re SW Workshop Basic (English) September 17, 2001 Greifensee

TGA (English)
DSC Basic (English)
DSC Advanced (English)
S6)Re SW Workshop Advanced (English)

September 18, 2001 Greifensee


September 19, 2001 Greifensee
September 20, 2001 Greifensee
September 21, 2001 Greifensee

Workshop Tipps und Hinweise


fr gute Messungen
Workshop Kurveninterpretierung

19. November 2001 Greifensee


20. November 2001 Greifensee

TA-Kundenkurse und Seminare (Deutschland)

Fr nhere Informationen wenden Sie sich bitte an METTLER TOLEDO GmbH, Giessen:
Frau Ina Wolf, Tel.: ++49-641 507 404.
Kundenkurse und Workshops:
DSC-Kundenkurs
Workshop Kurveninterpretation
TG-Workshop
S6)Re-Workshop (Basic)

22/23. 10. 2001 Giessen/DE


24. 10. 2001 Giessen/DE
13./14. 11.2001 Giessen/DE
15.11.2001 Giessen/DE

Fachseminare:
Thermische Analyse in der pharmazeutischen Forschung und Produktion 21. 6. 2001 Basel (CH)
Thermoanalytische und spektroskopische Methoden an Kunststoffen
30. 10. 2001 Dsseldorf (DE)

Cours et sminaires dAnalyse Thermique en France

France: Renseignements et inscriptions par Christine Fauvarque, METTLER TOLEDO S.A., Viroflay, Tl.: ++33-1 30 97 16 89, Fax: ++33-1 30 97 16 60.
Cours clients :
TG et logiciel S6)Re
DSC et logiciel S6)Re
DSC avanc et logiciel S6)Re
TMA et logiciel S6)Re

16 octobre 2001 Viroflay (France)


17 octobre 2001 Viroflay (France)
18 octobre 2001 Viroflay (France)
19 octobre 2001 Viroflay (France)

Journes dinformation:
Journe dinformation
Journe dinformation
Sminaires:
DSC Alternative et ses applications

25 septembre 2001 Nancy (France)


20 novembre 2001 Nantes (France)
23 octobre 2001 Paris La Dfense (F)

19
UserCom 1/2001

Cours et sminaires dAnalyse Thermique en Belgium

Belgique: Renseignements et inscriptions par Pat Hoogeras, N.V. METTLER TOLEDO S.A., Zaventem, Tl.: ++32-2 334 02 09, Fax: ++32 2 334 0 334.
TA courses: ( day each topic)
Theory of polymers
17 September 2001 am Brussels (BE)
Basics of DSC & calibrations
17 September 2001 pm Brussels (BE)
Amorpous samples (Tg, workshops)
18 September 2001 am Brussels (BE)
Crystalline samples (Tm, Tc workshops) 18 September 2001 pm Brussels (BE)

S6)Re SW training (workshops)


Curve interpretation
TGA: principles, applications, calibration
TMA / DMA: principles, applications

19 September 2001 am Brussels (BE)


19 September 2001 pm Brussels (BE)
20 September 2001 am Brussels (BE)
20 September 2001 pm Brussels (BE)

TA Customer Courses and Seminars in the Netherlands

Voor verdere informatie kunt U kontakt opnemen met: Hay Berden Mettler-Toledo B.V. Tiel . Tel. 0344-638363.

Corsi e Seminari di Analisi Termica per Clienti in Italia

Per ulteriori informazioni prego contattare: Simona Ferrari, Mettler-Toledo S.p.A., Novate Milanese,
Tel.: ++39-2 333 321, Fax: ++39-2 356 2973 E-mail: simona.ferrari@mt.com
Corsi per Clienti:
DSC base
DSC avanzato

18 Settembre 2001 Novate Milanese


19 Settembre 2001 Novate Milanese

TGA
TMA

20 Settembre 2001 Novate Milanese


21 Settembre 2001 Novate Milanese

Cursos y Seminarios de TA en Espaa

Para detalles acerca de los cursos y seminarios, por favor, contacte con: Francesc Catala en Mettler-Toledo S.A.E., Tel: ++34 93 223 76 00 E-Mail: francesc.catala@mt.com
Seminarios de Anlisis Trmico:
Jornada TA de aplicationes a Polmeros 23-oct-01 Barcelona
16-oct-01 Madrid

Jornada TA para Usuarios del Sistema S6)Re 24-oct-01


17-oct-01
Jornada TA de aplicaciones a Farmacia y Qumica 25-oct-01
18-oct-01

TA Customer Courses and Seminars for Sweden and the Nordic countries

Barcelona
Madrid
Barcelona
Madrid

For details of training courses and seminars please contact: Catharina Hasselgren at Mettler-Toledo AB,
Tel: ++46 8 702 50 24, Fax: ++46 8 642 45 62 E-mail: catharina.hasselgren@mt.com

TA customer courses and seminars in the UK

For details of training courses and seminars please contact: Rod Bottom at METTLER TOLEDO Ltd Leicester, Tel 0116 234 5025, Fax 0116 236 5500
DSC basic training course

October 23, 2001 Leicester

TA Customer Courses and Seminars in the USA and Canada

Basic Thermal Analysis Training based upon the STARe System version 6 is being offered in California and at Columbus, Ohio Headquarters. Training will include lectures and hands-on workshops. For information contact Jon Foreman at 1-800-638-8537 extension 4687, fax: 614-438-4871 or by e-mail: jon.foreman@mt.com
TA course

October 10 11, 2001Columbus (OH)

TA information day

June13-14, 2001 San Francisco, CA

TA Customer Courses in the South East Asia Regional Office, Kuala Lumpur.
For information on dates please contact:

Malaysia: Jennifer Lee at ++ 603-7032773, fax : 603-7038773


Singapore: Lim Li/Clive Choo at ++ 65-8900011, fax : 65-8900013

Thailand: W.Techakasembundit at ++ 662-7196480, fax 662-7196479


Or SEA regional office: Soosay P. at ++ 603-7041773, fax 603-7031772

TA Customer Courses and Seminars in Japan

For details of training courses and seminars please contact: Yasushi Ikeda at METTLER TOLEDO Japan, Tel.: +81-3-5762-0606; Facsimile: +81-3-5762-0756
TA information day
October 25, 2001 Osaka
Advanced S6)Re course
September 14, 2001 Tokyo
Advanced S6)Re course
November 15, 2001 Osaka
For further information regarding meetings, products or applications please contact your local METTLER TOLEDO representative.
Bei Fragen zu weiteren Tagungen, den Produkten oder Applikationen wenden Sie sich bitte an Ihre lokale METTLER TOLEDO Vertretung.
Internet: http:/www.mt.com

Editorial team

METTLER TOLEDO GmbH, Analytical, Sonnenbergstrasse 74, CH-8603 Schwerzenbach, Switzerland

Dr. M. Schubnell,
Physicist

Dr. R. Riesen,
Chemical Engineer

J. Widmann,
Chemical Engineer

Dr. J. Schawe,
Physicist

e-mail: urs.joerimann@mt.com, Tel.: ++41 1 806 73 87, Fax: ++41 1 806 72 60

Layout and production

Promotion & Dokumentation Schwerzenbach, Walter Hanselmann ME-51710099


Printed on 100% chlorine-free paper, for the sake of the environement.

20
UserCom 1/2001

C. Darribre,
Chemical Engineer

U. Jrimann
Electrical Engineer