H. Flandorfer et al. / Thermochimica Acta 459 (2007) 34–39
35
paths and mechanisms. The purpose of this study is to determine
the enthalpy of formation of intermetallic compounds formed
due to interfacial reaction between Ag–Cu–Sn solders and Ni
substrates at low temperatures. As there is no evidence for the
formation of ternary IMCs in the quaternary Ag–Cu–Ni–Sn sys-
binary IMCs stable at low temperatures
Selected literature data for the limiting partial enthalpy of
mixing of Ag, Cu or Ni with Sn are collected in Table 1. Like-
wise, Table 2 contains literature values for the enthalpies of
formation of various intermetallic compounds in the three binary
systems Ag–Sn, Cu–Sn and Ni–Sn.
Pieces of Sn with a total mass between 4 and 5 g were placed
into a BN (Boron nitride) crucible (8 mm in diameter, 68 mm
in height) which was put inside the quartz tube that has to
be introduced in one of the calorimeter cells. Before starting
the experiment, the quartz tube was flushed several times with
high purity argon (99.999% with Oxisorb cleaning system).
Finally, an argon flow of 30 ml/min was kept constant during
measurement. Little pieces (8–15 mg) of pure transition met-
als or samples were dropped from the automatic drop device at
drop temperature (DT ≈ 298 K) into the liquid Sn bath main-
tained at the measurement temperature (MT = 773, 873, 973, or
1073 K, respectively). Attheendofeachseriesofmeasurements,
the calorimeter was calibrated by five additions (approx. 30 mg
each) of standard ␣-Al2O3 supplied by NIST (National Institute
of Standards and Technology, Gaithersburg, MD, USA).
2. Experimental procedure
2.1. Sample preparation and analysis
The drop temperature (DT) and the measurement tempera-
ture (MT) for each drop were determined using thermo-resistors
and thermocouples. The accuracy is usually better than 1 ◦C.
Generally, the average values of DT and MT for the different
drops in one run were used for the data evaluation. From the
scatter of the ∞experimental results it is estimated that the val-
Samples of the intermetallic compounds Ag3Sn, Ag4Sn,
Cu3Sn(), Cu41Sn11(␦), Cu6Sn5, Ni3Sn-LT, Ni3Sn2-HT, and
Ni3Sn4 with the nominal compositions as given in Table 2
were synthesized from high purity materials: Ag shot (99.98%,
¨
¯
OGUSSA, Vienna, Austria), heated in a carbon crucible at
ues of ΔSolH as well as the final values of the enthalpy of
700 ◦C for 10 min to remove surface impurities; Cu (99.98%
Goodfellow, Cambridge, UK) treated under H2 flow at 200 ◦C
for 2 h to remove oxide layers. Ni (99.98%) and Sn (99.9985%,
both Alfa Johnson Matthey, Karlsruhe, Germany) were used as
received and without further treatment.
formation should be accurate within 1 kJ/mol.
3. Results
¯
3.1. Determination of ΔSolH in liquid tin bath
Weighed amounts of the pure elements were arc melted in
an inert argon atmosphere. The reguli were inverted several
times and the arc melting process was repeated to ensure com-
plete homogenization of alloys. The alloys were encapsulated
in quartz tubes and sealed under vacuum (≤1 Pa), then annealed
at appropriate temperatures listed in Table 2. After an annealing
time of 21 days, the alloys were removed from the furnace and
quenched in cold water.
A Guinier-Huber film camera with Cu K␣1 radiation was
used in order to analyze the phase composition of the sam-
ples. The powdered alloys were fixed on a plastic foil and pure
Si (99.9999%) was used as an internal standard. The exposure
timewas6 h. Table2summarizesinformationregardingnominal
compositions, annealing temperatures and identified phases.
In the course of the present study, the molar enthalpies of
successiveadditionsof smallpiecesof thepureelements toacor-
responding bath of pure tin or Sn-alloy (after the first addition).
The corresponding results for 773 K are plotted versus the molar
fraction of the added transitions metal in Figs. 1–3. In a simi-
the intermetallic compounds Ag3Sn, Ag4Sn, Cu3Sn, Cu41Sn11,
Cu6Sn5, Ni3Sn-LT, Ni3Sn2-HT and Ni3Sn4. As an example, the
results for Ag3Sn, Cu41Sn11 and Ni3Sn2-HT phases are shown
in Figs. 4–6.
The values are the measured heat effects ꢀQ of the following
reactions:
2.2. Calorimetry
(A)DT + (Sn or Sn-alloy)MT
→
(Solution)MT
;
ꢀQ/nA = ꢀsolH(A)
(1)
The solution calorimetric measurements were carried out in a
Calvet-typemicrocalorimeter(HT1000Setaram, Lyon, France).
Each of the twin cells is surrounded by a thermopile with more
than 200 thermocouples. The calorimeter is suitable for tem-
peratures up to 1000 ◦C and heated by a wire wound resistance
furnace. A self-made automatic drop device serves for up to 30
drops, measurement control and data evaluation is performed
with the software LabView and HiQ, both supplied by National
the graphical programming language G and HIQ is used for eval-
uation and presentation of numerical data; the syntax is very
similar to MATLAB. The whole apparatus was described in
details by Flandorfer et al. [18].
(AxBy)DT + (Sn or Sn-alloy)MT
→
(Solution)MT
;
ꢀQ/nA B = ꢀsolH(AxBy)
(2)
x
y
(A)DT refers to the pure elements and (AxBy)DT refers to the
intermetallic compounds in solid state, both at drop temperature,
and nA, nA B refer to their molar amounts.
x
y
(Sn or Sn-alloy)MT refers to the liquid Sn bath at the mea-
surement temperature and ꢀQ is the measured heat effect for an
individual addition. ΔSolH is the total heat effect measured for
one mole of the pure elements and intermetallic compounds.
¯