Papers in Physics, vol. 5, art. 050009 (2013)

Received: 30 October 2013, Accepted: 4 November 2013
Edited by: S. A. Grigera
Licence: Creative Commons Attribution 3.0
DOI: http://dx.doi.org/10.4279/PIP.050009

www.papersinphysics.org

ISSN 1852-4249

Reply to the Commentary on “Graphite and its hidden
superconductivity”

P. Esquinazi1∗

I appreciate very much the time our colleague
Forgan took to read my manuscript and write his
comment on the exposed Physics and interpreta-
tion of experimental results. It is surely not easy
to find somebody that provides such a detailed re-
port. There is no doubt that the subject of the
manuscript remains highly controversial and most
of the community does not simply believe in the
existence of high or low temperature superconduc-
tivity in non-intercalated graphite. After working
for nearly 13 years with this material, I had the op-
portunity to deal with all kind of transport, mag-
netic and band structure data and revise part of the
interesting history of this material. My personal
experience related to the defect-induced magnetic
order in graphite discovered more than 10 years ago
(a phenomenon that one finds nowadays in a large
number of compounds) and the (over)skepticism
the whole community had at that time showed me
(once more) that in natural sciences one should not
always accept the opinion of the majority.

Before replying to Forgan’s comment, I would
like to tell you a short story. Six years ago and
after independent colleagues proved using different
experimental methods that one can have magnetic
order in graphite due to defects or non-magnetic

∗E-mail: esquin@physik.uni-leipzig.de

1 Division of Superconductivity and Magnetism, Institute
for Experimental Physics II, Fakultät für Physik und Ge-
owissenschaften, Universität Leipzig, Linnéstrasse 5, D-
04103 Leipzig, Germany.

ad-atoms like hydrogen, a speculative explanation
on the magnetic data published in the year 2000 [1]
(59), I decided that it was time to check whether
the claim of superconductivity in graphite at ex-
traordinarily high temperature could be real or not.
The first unexpected hint came by chance, when I
took one HOPG sample from Advanced Ceramics
company and asked the responsible of a dual beam
microscope to take a look at the interior with the
low-energy transmission electron microscope option
that machine had. What we saw at that time were
well defined quasi two-dimensional interfaces inside
the samples. I had no idea about the origin of
those interfaces and whether they could influence
the transport properties or not. Years later I found
that these interfaces had been recognized before [2]
(8) but nobody apparently payed attention on their
possible influence on measurable properties. It was
the systematic change in the absolute value of the
resistivity as well as in its temperature dependence
with the thickness of that kind of HOPG samples
[3,4] (7,10) that helped me clarify some inconsisten-
cies found in the graphite literature and provided
us with a hint of where superconductivity, if at all,
could be hidden.

In the reply, I will try to describe and empha-
size important details of the experimental evidence
that were not, apparently, taken into account or
simply just overseen in Forgan’s comment. I reply
to Forgan’s comment in the order of appearance,
copying at the beginning of each issue part of the
corresponding paragraph to help the reader. With

050009-1



Papers in Physics, vol. 5, art. 050009 (2013) / P. Esquinazi

the same purpose, I have included citations at the
end of the reply. In case the citation was already
in the manuscript I wrote it with its corresponding
number in parentheses, at least the first time I cite
it.

(1) “This bulk property and many others, such
as the de Haas van Alphen effect in large samples
[1] have been understood in general terms [2] as a
consequence of a semi-metallic band-structure [3]
since 1960.”

Reply: I believe the graphite story is an exam-
ple that shows that sometimes a “democratic” ex-
perimental fact and its possible “democratic inter-
pretation” do not assure correctness. In this case,
such a “democratic consent” can have a rather
negative influence in the development of science,
more dangerous than a wrong interpretation or an
experiment carried out incorrectly. The cited de
Haas - van Alphen as well as the Schubnikov - de
Haas (SdH) oscillations in the magnetoresistance of
graphite were taken as evidence for the existence
of a finite Fermi surface and understood in terms
of an anisotropic 3D band structure with coupling
constants between C-atoms obtained from fitting
experimental data. Those data, e.g., those SdH
oscillations, were assumed to represent the ideal
defect-free graphite structure. However, system-
atically done measurements on thin graphite flakes
[5] (11), as well as the influence of irradiation in
the SdH oscillations [6] (9), already indicate that
the oscillations found in some bulk samples do not
correspond to ideal graphite. Graphite samples,
independently of their size but without interfaces,
do not show any oscillations in the magnetoresis-
tance (even in its derivative) at low temperatures.
For such samples, these SdH oscillations can be ob-
served after producing defects (i.e., after increasing
the carrier density in some regions of the sample)
[6] or by applying a large enough electric field [7].

(2) “I now turn to the various sections of the pa-
per. In section II, there is an account of strong
magnetoresistance effects. Similar effects have also
been observed in bismuth [4] and have a very inter-
esting explanation [4] in terms of the semi-metallic
properties of graphite and bismuth, so there is no
need to propose a superconducting explanation for
this.”

Reply: One can start arguing against the ex-
planation of the (magnetic field driven) Metal-
Insulator-Transition (MIT) proposed in that paper

[8] (26) in Forgan’s report taking into account the
large (about 6) number of free parameters and the
inconsistency of the proposed explanation with the
nearly linear field dependence of the resistance ob-
served at low temperatures in ordered graphite. On
the other hand, one could argue that those crit-
ical points are of rather technical nature that do
not touch the main idea of the proposed expla-
nation. However, there is a more serious prob-
lem with the “interesting explanation” based on
the assumed electronic band structure of Bernal
graphite: As I wrote in my review, the MIT is not
observed when the graphite sample has no inter-
faces. That is to say, the “democratically observed”
metallic-like temperature dependence of graphite
samples is simply not intrinsic of ideal graphite
[3, 4, 9] (7,10,12). Therefore, the explanation used
in [8] to understand apparent “intrinsic properties”
of Bernal ideal graphite is actually not applicable.
The MIT as well as the metallic-like temperature
dependence of graphite samples is related to the
existence of interfaces, or other lattice defects, and
this matches the forthcoming explanations of other
effects in the review. If one does not realize or
accept this fact discussed at the beginning of the
review, one loses an important part of evidence.

Regarding the MIT found also in bismuth in
that paper [8]: it is important to remark that in
that work [8], no information about the internal
structure of the measured Bi sample was given and
whether interfaces were there or not. This is impor-
tant because, as graphite, interfaces in Bi samples
can have superconducting properties at tempera-
tures even above 10 K [10–14] (-,-,90,-,91), although
pure Bi bulk is not a superconductor. This inter-
face effect in both semimetals seems to be more
than a simple coincidence.

(3) “In section III, a tiny hysteresis in magnetore-
sistance is described. Two comments are relevant
here: the author notes that the sign of the hys-
teresis is opposite to that expected for a supercon-
ductor and limits himself to stating that the data
provide “striking hints that granular superconduc-
tivity is at work in some regions of these samples”.
This is hardly definitive proof.”

Reply: Certainly, a definitive proof for the ex-
istence of granular superconductivity only through
an anomalous magnetoresistance hysteresis loop is
not. However, if we put all the pieces of the puz-
zle together, the proposed interpretation does not

050009-2



Papers in Physics, vol. 5, art. 050009 (2013) / P. Esquinazi

appear like a simple coincidence. The evidence of
an anomalous hysteresis in the magnetoresistance
measured in graphite flakes and its enhancement
with constrictions leave not so many possibilities of
interpretation. Note that the hysteresis appears in
the temperature region where the resistance shows
a metallic-like temperature dependence. Coinci-
dentally, it is in this temperature region where we
observe the field driven MIT. This experiment can
be easily reproduced in all graphite flakes that show
such a maximum in the resistance vs. temperature,
if one has a measurement system with a 10−5 rel-
ative error in the measurement of the resistance
and good temperature stability. A hysteresis in
the magnetoresistance is evidence for magnetic en-
tities that remain pinned or for example, magnetic
anisotropy observed in ferromagnets. Typical ferro-
magnetic hysteresis loops in the magnetoresistance
have been observed in graphite flakes [15] (46) as
well as in bulk graphite after proton irradiation [16]
(22). These facts do not speak for an origin of the
anomalous hysteresis observed in [17, 18] (38,42)
in terms of ferromagnetism. Similar anomalous
hysteresis in the magnetoresistance were observed
in granular superconductors [19–21] (39,40,41) and
explained in terms of Josephson-coupled supercon-
ducting grains [19].

(4) “ Section IV is headed “Direct evidence for
Josephson behaviour”. This quotes data ... and the
fact that magnetic fields could increase, decrease or
have no effect on the voltages observed, also cast
great doubt on the Josephson interpretation.”

Reply: Indeed, the currents used here were small
simply because high currents shift the observed
transitions systematically to lower temperatures
[22] (45) and this fact should be taken as evidence
in the direction of superconductivity, actually. It
is also correct that no measurement of a strictly
zero resistance state has been shown in [22] just
because in such measurements and due to the fi-
nite sensitivity one cannot measure zero resistance,
this is actually obvious. However, the estimate of
the minimum measured resistance given in the com-
ment is not quite correct. Upon sample and at low
enough temperature, the voltage noise around zero
voltage measured at 1 µA is ±5 nV, or ±5 mΩ for
samples with a resistance at high temperature of
the order of 100 Ω. It should be clear that upon
Josephson coupling and the characteristics of the
superconducting patches at the interfaces, thermal

fluctuations may affect the zero average value. To
verify that a zero resistance state is possible one
needs to show that currents remain for a sufficiently
large time by measuring, for example, the magnetic
moment of a ring where a superconducting current
flows, as it has been done recently using graphite
flakes embedded in alkanes [23] (94). Hopefully,
new experiments in this direction will clarify the
situation.

Such sharp transitions in the measured voltage
vs. temperature do not appear to be simply possi-
ble from wrong contacts or anisotropic current dis-
tributions. In particular, when the whole I − V
behavior is compatible with the one expected from
Josephson-coupled regions within the interfaces.
That an anisotropic current distribution can exist
in the graphite lamellae is especially true in the
case the contacts are localized at the corners of the
lamellae, i.e., in a Van de Pauw configuration, as
it has been clearly stated in the review and in [22]
(45). In this case, the simple model used to fit
the I − V curves takes explicitly into account the
anisotropic arrangement. The negative resistance
behaviour was observed only in that case but not
for the usual linear electrode arrangement [22], as
expected. It seems to be more than a simple coinci-
dence that the same equation with only the critical
Josephson current as free parameter is sufficient to
interpret the I−V curves measured in very different
configurations. Finally, one can convince oneself
about the relationship between interfaces and the
observed transitions simply by measuring a lamella
without interfaces but with the same anisotropy of
graphite and with similar contacts.

Forgan correctly pointed out the striking mag-
netic field effects on the I − V and V (T) at con-
stant current. The effect of a magnetic field on the
I − V characteristics is as expected only for large
(thick) samples and in the same field region (a few
kOe) where the field suppresses the metallic-like
behaviour, i.e., the MIT. At fields higher than sev-
eral Tesla, one observes a reentrance in the I −V
curves at low currents, i.e., the resistance starts to
decrease with field [22]. This effect in the resistance
has already been reported in 2003 [24] (48) and its
interpretation remains open. Again, this effect is
related to the existence of the interfaces and it does
not appear to be a simple artefact of anisotropic
current distribution or contact problems.

(5) “In response to [6], a colleague repeated

050009-3



Papers in Physics, vol. 5, art. 050009 (2013) / P. Esquinazi

their measurements as an undergraduate project
[7]. Their clear conclusion was that if the correct
diamagnetic background slope (that obtained at
large fields) is subtracted, then the hysteresis cor-
responds to a tiny ferromagnetic component. How-
ever, if a slightly different background is chosen, the
hysteresis loops look somewhat like the response of
a granular superconductor.”

Reply: The arbitrariness of the diamagnetic
background subtraction has been taken into ac-
count in the description of the results in [25, 26]
(58,29) and, indeed, the appearance of the hys-
teresis can be changed subtracting different back-
grounds. Although I do not know the results from
the undergraduate project, I would like to empha-
size a few details and the difference of the results
in [25, 26] with those one expects from a simple fer-
romagnetic response.

There are two methods one can use without the
need of any background subtraction to obtain the
true magnetic response of the sample. The first one
is to measure the remanent magnetic moment (i.e.,
at zero field) after cycling the sample to a maximum
field strength, see Fig. 3 in [25] or Fig. 10 in [26].
In general, for a ferromagnetic sample and from the
measurements of the minor loops at low fields one
obtains a remanence Mr that increases following
the Rayleigh law, i.e., Mr ∝ Hnmax, with Hmax the
maximum applied field strength and n ∼ 1 . . . 1.5,
see [27]. The measurements done in graphite pow-
ders and HOPG samples with interfaces do show,
however, a within error nearly zero remanence up
to a temperature dependent critical field hJc1(T)
[25, 26].

One could still argue that a critical field may
also appear through magnetic domain pinning or
magnetic anisotropy response in any ferromagnetic
inclusions that exist in the graphite powder. In
this case, one can use a second method to measure
the difference between field cooled (FC) and zero-
field cooled (ZFC) temperature dependent mag-
netic moment m(T). A finite positive difference
between mFC −mZFC can be taken as due to pin-
ning of magnetic entities, superconducting vortices
or magnetic domains in ferromagnets, for exam-
ple. Let us assume that the hysteresis seen in [25]
is due to a ferromagnetic response with a satura-
tion field Hs ∼ 1 kOe. In this case, the difference
mFC −mZFC would increase with field reaching a
maximum at a field H . Hs, but then it should

decrease to zero sharply at higher fields. This ex-
periment can be easily carried out using, for exam-
ple, a real ferromagnet as micro or nano particles
of magnetite, which upon size and sample prepara-
tion can show saturation fields of this order. How-
ever, this is not the behaviour reported in [25] (see
Fig. 6 in the Supporting Information of that ar-
ticle). This difference, after reaching a minimum
or plateau at ∼ 1 kOe, increases again steadily up
to the maximum applied field of 7 T. It seems dif-
ficult to find a ferromagnetic material that shows
a hysteresis in field that does not saturate but in-
creases steadily with field of the order or higher
than 7 T. We have repeated this kind of experi-
ments in different graphite powders from the same
source as used in [25] and this behaviour is well re-
producible and it does not appear compatible with
a ferromagnetic response. Measurements with fer-
romagnetic small particles embedded in disordered
carbon show clearly different behavior from that
reported in [25, 26], as expected.

Regarding Forgans’s comment, and I quote,
“There are many possible reasons (both real and
due to experimental artifacts) why measurements
on a sample taken on heating and cooling might
disagree”, one can test the SQUID system and con-
vince oneself about its limits and sensitivity, using
different samples, as HOPG samples without in-
terfaces [26] or just amorphous carbon powder or
ferromagnetic particles in a non magnetic matrix,
etc., and check whether similar behavior for the
mFC − mZFC is observed using exactly the same
SQUID sequence. The results we obtain from all
those samples provide us with the necessary confi-
dence.

Let us assume now that the undergraduate stu-
dents cited by Forgan measured indeed a hysteresis
of ferromagnetic nature. If it is due to magnetic
impurities, then this can be firstly proved through
elemental analysis and the behaviour of the rema-
nence and the difference mFC − mZFC must be
different from the one of a granular superconduc-
tor. But, what if this ferromagnetic response is
not due to magnetic impurities but it is due to
graphite itself due to, for example, hydrogenation
through the water treatment? Note that hydrogen
may trigger magnetic order in graphite, as shown
recently through measurements using three differ-
ent experimental methods as XMCD [28], mag-
netization and the anisotropic magnetoresistance

050009-4



Papers in Physics, vol. 5, art. 050009 (2013) / P. Esquinazi

[29] (56). The hysteresis curves of ferromagnetic
graphite reported in those works as well as in sev-
eral other independent reports show, however, sat-
uration fields of the order of a few kOe, without
any opening of the hysteresis at higher fields, and
a small remanence in contrast to the nearly square
hysteresis loops found in [25]. Thus, the usual fer-
romagnetic behaviour of graphite does not seem
to be compatible with the reported observations in
[25, 26].

Nevertheless, if the ferromagnetic response is due
to the treatment the undergraduate students did
to the graphite powder, and vanishes after pressing
the powder (see Fig. 4(c) in [25]), that means that
it may come from interfaces between grains or grain
surface near region. In this case, it should be taken
seriously, spending more time on its characteriza-
tion. One should not take for granted that mag-
netic order in any graphite or carbon-based sam-
ple is due to impurities and not something intrin-
sic. There is enough evidence about defect-induced
magnetism in graphite, due to vacancies as well as
due to hydrogen (for a short review see [30] and refs.
therein). It is even possible that some interfaces
in graphite show magnetic order or even that both
phenomena, superconductivity and magnetism, ap-
pear and a mixture of both signals is observed. I
would like to cite one sentence written in the con-
clusion of the work in [31] (21) where the flat band
at the interfaces between Bernal and rhombohedral
graphite structures is proposed as the origin for the
high temperature superconductivity: “In general,
flat bands are susceptible to instabilities with re-
spect to some other ordered states; for example, a
magnetic state could also be possible.” Evidently, if
both phenomena occur at the interfaces or surface
of graphite, the situation will get more interesting
but difficult from the experimental point of view.
Unless one can prove that the ferromagnetic hys-
teresis is due to impurities, one should not take
this kind of evidence offhandedly.

(6) “We see in [6] that the hysteresis at 300 K
is essentially the same as that at 5 K. We bear in
mind that by assumption the superconductivity is
confined to an atomic layer, and that the higher
the Tc of a superconductor the shorter the coher-
ence length. These two together ensure that ther-
mal fluctuations (which are already very noticeable
at T ∼ 100 K in cuprate materials) would be huge
for any room temperature graphite superconductiv-

ity [9]. Thermal fluctuations would greatly reduce
vortex pinning and magnetic irreversibility at room
temperature, contrary to what is observed.”

Reply: A possible answer to this comment might
be that the superconductivity at graphite interfaces
has not only a ten times larger critical tempera-
ture but also ten times larger activation energies as
in cuprate materials. In fact, in [25, 26] not only
a ten times larger critical Josephson fields (within
the interpretation given in that paper), but also
a larger pinning potential barrier than in cuprates
have been estimated from the time relaxation mea-
surements. In this case, the effect of thermal fluc-
tuations should not affect substantially the sam-
ple response between 5 K and 300 K. It is rather
premature to speculate based on usual s-wave (or
d-wave) pairing equations. If it is true that the
magnetic field makes this interface superconductiv-
ity robust then the equations used for conventional
and cuprates superconductors will not be applica-
ble.

Forgan speculates that the coherence length
should be very small. Should it really be small?
The usual estimate of the coherence length is
based, in general, on the Ginzburg-Landau theory.
Whether this is applicable in the case of supercon-
ductivity at graphite interfaces has to be seen. We
note that a direct measurement of the coherence
length is not possible. The use of the proximity ef-
fect or the change of the critical temperature with
sample size are possible experimental methods one
can take to estimate the coherence length, if the
upper critical field cannot be measured. If one uses
the variable sample size method on the interfaces
found in graphite, one observes indeed a system-
atic decrease of the Josephson critical temperature
measured at a fixed current, decreasing the width of
the interfaces, i.e., the thickness of the TEM lamel-
lae, behaviour measured recently in more than eight
samples of the same HOPG batch (same interface
density) [32]. This size dependence may provide
also a hint to understand the differences in the be-
havior of large and small samples, i.e., for SQUID
or transport measurements.

It is still too early to answer whether recent the-
ories [31, 33] (21,93) can explain quantitatively the
observed behaviour. Note that, according to the
last theoretical work [33], the nature of granular su-
perconductivity that may exist at interfaces is not
as simple as in usual Josephson-coupled localised

050009-5



Papers in Physics, vol. 5, art. 050009 (2013) / P. Esquinazi

superconductor grains in a normal matrix. The size
of the effective region that influences the observed
Josephson behaviour may be larger than the intrin-
sic coherence length.

(7) “I cannot give an overriding simple expla-
nation for all the different results reported in Es-
quinazi’s paper, but neither can the author.”

Reply: This is obviously true (for both) and I
fully agree. Taking into account that the Physics
of interfaces in graphite is a subject starting just
now and that nobody has tried to make them sys-
tematically, it is natural that we need time. Taking
the example of the cuprates and the time needed to
clarify some, not all, open issues, nobody should be
surprised that we do not have “an overriding simple
(why simple?) explanation” at the moment.

[1] Y Kopelevich, P Esquinazi, J Tor-
res, S Moehlecke, Ferromagnetic- and
superconducting-like behavior of graphite, J.
Low Temp. Phys. 119, 691 (2000).

[2] M Inagaki, New Carbons: Control of Structure
and Functions, Elsevier (2000).

[3] J Barzola-Quiquia, J L Yao, P Rödiger,
K Schindler, P Esquinazi, Sample size ef-
fects on the transport properties of mesoscopic
graphite samples, Phys. Status Solidi A 205,
2924 (2008).

[4] N Garćıa, P Esquinazi, J Barzola-Quiquia,
S Dusari, Evidence for semiconducting behav-
ior with a narrow band gap of Bernal graphite,
New J. Phys. 14, 053015 (2012).

[5] Y Ohashi, K Yamamoto, T Kubo, Shubnikov
- de Haas effect of very thin graphite crystals,
In: Carbon’01, An International Conference
on Carbon, Pag. 568, The American Carbon
Society, Lexington, KY, United States, (2001).

[6] A Arndt, D Spoddig, P Esquinazi, J Barzola-
Quiquia, S Dusari, T Butz, Electric carrier
concentration in graphite: Dependence of elec-
trical resistivity and magnetoresistance on de-
fect concentration, Phys. Rev. B 80, 195402
(2009).

[7] A Ballestar, J Barzola-Quiquia, S Dusari,
P Esquinazi, R R da Silva, Y Kopelevich, Elec-

tric field induced superconductivity in multi-
graphene, arXiv:1202.3327 (2012).

[8] X Du, S W Tsai, D L Maslov, A F Hebard,
Metal-insulator-like behavior in semimetallic
bismuth and graphite, Phys. Rev. Lett. 94,
166601 (2005).

[9] B C Camargo, Y Kopelevich, S B Hubbard,
A Usher, W Böhlmann, P Esquinazi, Effect of
structural disorder on the quantum oscillations
in graphite, (unpublished). In this work the
authors show that in certain HOPG samples
(SPI) of high grade, the density of interfaces
is much lower than in, for example, Advanced
Ceramics HOPG ZYA samples. In this new
HOPG samples basically no SdH oscillations
are found and the temperature dependence of
the resistance shows a semiconducting behav-
ior with saturation a low temperatures. (2013).

[10] D V Gitsu, A F Grozav, V G Kistol, L I Lep-
orda, F M Muntyanu, Experimental obser-
vation of a superconducting phase with Tc '
8.5 K in large-angle bismuth bicrystals, JETP
Lett. 55, 403 (1992).

[11] F M Muntyanu, L I Leporda, Restructuring
of the energy spectrum in large angle bismuth
bicrystals, Phys. Solid State 37, 298 (1995).

[12] F Muntyanua, A Gilewski, K Nenkov,
J Warchulska, A Zaleski, Experimental mag-
netization evidence for two superconducting
phases in Bi bicrystals with large crystallite
disorientation angle, Phys. Rev. B 73, 132507
(2006).

[13] F M Muntyanu, A Gilewski, K Nenkov, A J
Zaleski, , V Chistol, Fermi-surface rearrange-
ment in Bi bicrystals with twisting supercon-
ducting crystallite interfaces, Phys. Rev. B 76,
014532 (2007).

[14] F Muntyanua, A Gilewski, K Nenkov, A Za-
leski, V Chistol, Superconducting crystallite
interfaces with Tc up to 21 K in Bi and Bi-
Sb bicrystals of inclination type, Solid State
Commun. 147, 183 (2008).

[15] J Barzola-Quiquia, P Esquinazi,
Ferromagnetic- and superconducting-like
behavior of the electrical resistance of an

050009-6



Papers in Physics, vol. 5, art. 050009 (2013) / P. Esquinazi

inhomogeneous graphite flake, J. Supercond.
Nov. Magn. 23, 451 (2010).

[16] P Esquinazi, J Barzola-Quiquia, D Spemann,
M Rothermel, H Ohldag, N Garćıa, A Setzer,
T Butz, Magnetic order in graphite: Exper-
imental evidence, intrinsic and extrinsic dif-
ficulties, J. Magn. Magn. Mat. 322, 1156
(2010).

[17] P Esquinazi, N Garćıa, J Barzola-Quiquia,
P Rödiger, K Schindler, J L Yao, M Ziese,
Indications for intrinsic superconductivity in
highly oriented pyrolytic graphite, Phys. Rev.
B 78, 134516 (2008).

[18] S Dusari, J Barzola-Quiquia, P Esquinazi,
Superconducting behavior of interfaces in
graphite: Transport measurements of micro-
constrictions, J. Supercond. Nov. Magn. 24,
401 (2011).

[19] L Ji, M S Rzchowski, N Anand, M Thinkam,
Magnetic-field-dependent surface resistance
and two-level critical-state model for granular
superconductors, Phys. Rev. B 47, 470 (1993).

[20] Y Kopelevich, C dos Santos, S Moehlecke,
A Machado, Current-induced superconductor-
insulator transition in granular High-Tc Super-
conductors, arXiv:0108311 (2001).

[21] I Felner, E Galstyan, B Lorenz, D Cao, Y S
Wang, Y Y Xue, C W Chu, Magnetoresistance
hysteresis and critical current density in gran-
ular RuSr2Gd2−xCexCu2O10−δ, Phys. Rev. B
67, 134506 (2003).

[22] A Ballestar, J Barzola-Quiquia, T Scheike,
P Esquinazi, Evidence of Josephson-coupled
superconducting regions at the interfaces of
highly oriented pyrolytic graphite, New J.
Phys. 15, 023024 (2013).

[23] Y Kawashima, Possible room temperature
superconductivity in conductors obtained by
bringing alkanes into contact with a graphite
surface, AIP Advances 3, 052132 (2013).

[24] Y Kopelevich, J H S Torres, R R da Silva,
F Mrowka, H Kempa, P Esquinazi, Reentrant
metallic behavior of graphite in the quantum
limit, Phys. Rev. Lett. 90, 156402 (2003).

[25] T Scheike, W Böhlmann, P Esquinazi,
J Barzola-Quiquia, A Ballestar, A Setzer, Can
doping graphite trigger room temperature su-
perconductivity? Evidence for granular high-
temperature superconductivity in water-treated
graphite powder, Adv. Mater. 24, 5826 (2012).

[26] T Scheike, P Esquinazi, A Setzer,
W Böhlmann, Granular superconductiv-
ity at room temperature in bulk highly oriented
pyrolytic graphite samples, Carbon 59, 140
(2013).

[27] S Kobayashi, S Takahashi, T Shishido, Y Ka-
mada, H Kikuchi, Low-field magnetic charac-
terization of ferromagnets using a minor-loop
scaling law, J. Appl. Phys. 107, 023908 (2010).

[28] H Ohldag, P Esquinazi, E Arenholz, D Spe-
mann, M Rothermel, A Setzer, T Butz, The
role of hydrogen in room-temperature ferro-
magnetism at graphite surfaces, New J. Phys.
12, 123012 (2010).

[29] J Barzola-Quiquia, W Böhlmann, P Es-
quinazi, A Schadewitz, A Ballestar, S Dusari,
L Schultze-Nobre, B Kersting, Enhancement
of the ferromagnetic order of graphite after sul-
phuric acid treatment, Appl. Phys. Lett. 98,
192511 (2011).

[30] P Esquinazi, W Hergert, D Spemann, A Set-
zer, A Ernst, Defect-induced magnetism in
solids, IEEE Transactions on Magnetics 49,
4668 (2013).

[31] N B Kopnin, M Ijäs, A Harju, T T Heikkilä,
High-temperature surface superconductivity in
rhombohedral graphite, Phys. Rev. B 87,
140503 (2013).

[32] A Ballestar, Superconductivity at graphite in-
terfaces, Ph.D. Thesis, University of Leipzig,
(unpublished).

[33] W A Muñoz, L Covaci, F Peeters, Tight-
binding description of intrinsic superconduct-
ing correlations in multilayer graphene, Phys.
Rev. B 87, 134509 (2013).

050009-7