The preparation for the zero-field Tc measurements was as
follows. After a batch of KHg-GIC's was removed from the furnace in its
intercalation ampoule, the individual GIC's were transferred inside an
inert-gas-filled glovebox to 4mm Pyrex tubes. When the specimens were
intended for critical field studies, they were mounted in special sample
holders before placement in the glass tubes (see Section ).
Once the specimens were inside the glass tubes, the tubes were attached to a
stopcock via a Cajon quick-connect vacuum fitting. They were then taken
out of the glovebox and attached to a pumping station, where they could be
evacuated to pressures on the order of 10-5 torr with a diffusion
pump. Once the ampoules were evacuated, they were filled with about 200
torr of He gas which had been passed through a liquid-nitrogen cold trap to
remove water. This gas was intended to provide thermal contact of the
sample to the glass tube, thus making the cryogenic temperature
measurements more accurate. The He-filled tubes were then sealed off with
a gas/oxygen torch.
Once the GIC's were in individual tubes, they were characterized
for staging fidelity with (00l) x-rays (see Section ). Those
that were well-staged and uniform in color were taken to the National
Magnet Laboratory for zero-field Tc and
critical field measurements. The zero-field Tc measurements
were performed using the same cryogenics, electronics, and thermometry that
were employed in the critical field studies. This apparatus is described
in Section
.
Tc was determined by measuring the voltage from an inductance bridge as a function of temperature from 4.2 K down to 0.4 K. For materials in which a higher Tc than 4.2 K seemed plausible, higher temperatures were achieved by bleeding warm He gas into the liquid helium cryostat. The transition was easily recognized as a large reproducible decrease in the inductive voltage. Signal-to-noise ratios of 1000 were not uncommon.
Tc was defined to be the intersection of a line drawn
tangent to the most linear part of the transition with the level upper part
of the transition. This definition, which is a standard one for inductive
measurements,[120] is illustrated by Figure . The
transition width, Delta Tc, was defined as the interval between the
10% and 90% completion temperatures of the transitions. The
dimensionless figure of merit used to characterize sample quality is Delta Tc/Tc. This ratio is a Tc-independent measure of how
narrow a specimen's superconducting transition is. Delta Tc/ Tc
ranged from about 0.2 for poorly ordered samples to about 10-2 for
well-ordered specimens.
Figure: Experimental definition of
Tc.
Tc and Delta Tc/Tc are reported for a number of
samples in Table . Delta Tc/Tc mostly increases
with decreasing Tc, which suggests that the low- Tc GIC's are
less homogeneous than the higher- Tc ones. The observation of
broader transitions in lower- Tc samples is in keeping with the idea
that the lower- Tc specimens contain both the alpha and beta
phases, while the higher- Tc specimens contain only the alpha phase.
If the beta phase is responsible for Tc suppression in
C4KHg, one might expect that samples might sometimes have two
separate superconducting transitions, the first at 1.5 K and a second at
0.8 K. Yet two separate transitions have never been observed in a C4KHg sample, although very broad transitions have sometimes been seen.
An interpretation for the lack of two transitions is suggested by the TEM
work performed by Timp[246]. The TEM micrographs consistently showed
that the in-plane phases were intermixed on a microscopic scale. If the
regions of beta phase are small and are surrounded by alpha-phase,
they may well be proximity-coupled to the alpha-phase regions.
Proximity coupling between the lower- and higher- Tc regions of the
specimen can give a single transition, just as is observed. Caution is
necessary when discussing the intermixture of phases, though, since a
high-enough degree of admixture along the c-axis would give broadened
(00l) x-ray lineshapes.[109] As the spectra in
Section show, these broadened lineshapes have not been observed.
Table: Transition temperatures Tc and reduced widths Delta Tc/Tc for a sampling of C4KHg specimens.
A sceptic might claim that the color of the C4KHg samples is
a property only of the surface layers, and is not really linked to the
superconductivity. This is not an unreasonable idea since the (sqrt3 × sqrt3)R30° superlattice has been tentatively identified
as a surface phase in Section . However, the gold color is
strongly correlated with the presence of the beta phase in the
(00l) x-rays. Furthermore, the color is as good a predictor of Tc as the x-rays are, as Table
shows. Lagrange and
coworkers have also reported the existence of the gold color in association
with the beta phase.[147] They report that repeated cleaving
of a ``gold'' sample will expose both pink and gold surfaces in turn.
From these considerations and the evidence in Table ,
it is clear that gold samples have lower Tc's than pink ones.
Furthermore, this Tc depression seems to be linked to the presence
of the beta phase. The possible effect of the two phases of C4KHg on superconductivity measurements is discussed further in
Chapter
. The differences in the critical fields of the two
types of C4KHg are detailed in Chapter
.