RSC Advances

PAPER

Detection of L-band electron paramagnetic

Cite this: RSC Adv., 2020, 10, 17311
resonance in the DPPH molecule using impedance
measurements
Ushnish Chaudhuri and R. Mahendiran *

Detection of electron paramagnetic resonance (EPR) using a microwave cavity resonating at a fixed
frequency (between 9 and 10 GHz) remains the most popular method to date. Here, we report a cavity-
less technique which makes use of only an impedance analyzer and a copper strip coil to detect L-band
EPR (f ¼ 1–3 GHz) in the standard EPR marker 2,2-diphenyl-1-picrylhydrazyl (DPPH). Our method relies
on measuring the magnetoimpedance (MI) response of DPPH through a copper strip coil that encloses
DPPH. In contrast to commercial EPR which measures only the field derivative of power absorption, our
method enables us to deduce both absorption and dispersion. Changes in resistance (R) and reactance
(X) of the copper strip while sweeping an external dc magnetic field, were measured for different
frequencies (f ¼ 0.9 to 2.5 GHz) of radio frequency current in the coil. R exhibits a sharp peak at a critical
value of the dc magnetic field, which is identified as the resonance field and X shows a dispersion at the
Received 12th April 2020
Accepted 27th April 2020
same frequency. The data were analyzed to obtain line width and resonance field parameters. The
resonance field increased linearly with frequency and the obtained Landé g factor of 1.999  0.0197 is
DOI: 10.1039/d0ra03285a
close to the accepted value of 2.0036, measured in the X-band. The simplicity of this technique can be
rsc.li/rsc-advances exploited to study paramagnetic centers in catalysis and other materials.

microwave synthesizers working in a broad frequency range

Introduction (50 MHz to 50 GHz) to deliver microwave power to CPW/MSR
The rst electron paramagnetic resonance (EPR) spectrum was to create intense local microwave magnetic eld on the surfaces
recorded by Evgeny Konstantinovich Zavoisky almost 70 years of CPW or MSR upon which a paramagnetic metallic sample is
ago using a simple home-built spectrometer.1 Over the years, placed. However, with increasing frequency of the microwave
EPR has evolved into a sophisticated instrument and become an signal, penetration of electromagnetic eld inside a conducting
essential tool to detect unpaired electrons in solids which in sample also decreases due to skin effect. An EPR spectrometer
turn provides information about electronic structure of the working L-band (1–3 GHz) or still lower frequencies is prefer-
paramagnetic centers and their chemical environments.2 EPR is able to ensure a deeper penetration of electromagnetic waves
also used to probe local defects in Si/SiO2,3 to study structural into a conducting or a biological sample. The EPR spectra at low
phase transitions,4 kinetics of chemical reactions,5,6 electron frequencies can exhibit better resolution in certain cases, e.g.
and spin transfer in catalysis,7,8 structure of proteins, organic the EPR spectra of Cu2+ complexes such as Cu(DOPA)2 and
free radicals, etc.7 Thus, applications of EPR span all the Cu(carnosine)4 were better resolved at 2.62 GHz than at 9.30
branches of science. To date, the most popular method to detect GHz.19 Eaton and Eaton have given an overview of spectrome-
EPR makes use of a xed frequency microwave cavity resonator ters developed for frequencies below X-band.20 In this paper we
in the X band (9 GHz) or Q band (35 GHz) frequency regime.9 describe a previously unreported technique for free radical
However, metallic samples with high conductive losses as well compounds which can aid researchers investigate EPR from
as aqueous biological/chemical samples with large dielectric a fresh perspective.
losses prove to be limiting for commercial EPR spectrometers In this article, we present a simple method to detect EPR in
operating in X band since they cause frequency shi due to the L-band frequency region, which makes use of only an
dispersion of permeability.10 Recently, coplanar waveguides impedance analyzer and a copper strip coil. In a closely related
(CPW)11–14 and microstripline resonators (MSR)15–18 have been work, Kitagawa studied EPR of DPPH placed above a 50 ohm
exploited to investigate smaller sized lossy samples over a broad impedance matched microinductor fabricated with CMOS-25
frequency range. These methods take advantage of advances in nanotechnology and used a vector network analyzer to extract
EPR signal in frequency sweep mode.21 A vector network
Department of Physics, National University of Singapore, 2 Science Drive 3,
analyzer measures the ratio of a transmitted or reected power
Singapore-117551, Republic of Singapore. E-mail: [email protected] to the incident power from a device under test. Impedance is

This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 17311–17316 | 17311
RSC Advances Paper

calculated from scattering parameter S12 and S11.21 In contrast, by two/four probe methods or vector network analyzers in the
the instrument used in the current work (Agilent E4991 RF frequency range 10 MHz to 20 GHz.24–26 For the later case,
impedance analyzer) measures impedance of a copper strip a microstripline is photolithographically fabricated on a dielec-
based on the radio frequency current–voltage technique and tric substrate.27,28 On the other hand, we have recently have
does not require 50 ohm impedance matching or CMOS tech- discovered that MI measured by passing microwave (MW)
nology. The copper strip is prepared manually by cutting and current directly through some electrically conducting Mn-based
folding a copper foil to desired shape. Our method is compact, perovskite oxides can detect not only paramagnetic but also
requires very less instrumentation and does not require multi- ferromagnetic resonance of exchange coupled t2g spins of
step device fabrication therefore, it can be easily incorporated in Mn3+:t32ge1g and Mn4+:t32ge0g ions.29,30 In manganites, short-range
teaching and research laboratories. Our technique provides correlation between localized t2g core spins (S ¼ 32) mediated
additional information pertaining to the absorptive and by hopping of eg electron (S ¼ 12) leads to intense ESR signal in
dispersive components of the high frequency magnetic the paramagnetic state. Resistivity of a manganite shows sem-
susceptibility whereas conventional EPR spectrometers are iconducting like behavior in the paramagnetic and its value is of
designed to provide information about the eld derivative of the the order few milliohm cm at room temperature. However,
power absorbed by the sample. Using our simple setup, we DPPH is an insulator and hence an indirect method is employed
demonstrate the detection of EPR due to free radicals in in the present work to measure the MI of DPPH.
a standard sample of 2,2-diphenyl-1-picrylhydrazyl (DPPH). As Our technique involves using a copper strip coil as an
a stable and well-characterized solid radical source, DPPH is the antenna and a radio frequency impedance analyzer (Agilent
most popular reference sample with Landé g-factor of 2.0036.22 model E4991A) as a microwave signal source and detector.
The intensity of EPR signals depends on the number of radicals DPPH powder obtained from Sigma-Aldrich™ was pressed into
for a freshly prepared sample and can be determined by a disc shaped pellet at room temperature using a hydraulic
weighing the DPPH sample. DPPH exhibits a single response press (pressure 5 ton per in2). Then, the disc was cut into
line in X-band with a small linewidth 1.5–4.7 Oe due to the a rectangular bar of dimension (4.5 mm  3.5 mm  0.5 mm).
presence of only one unpaired spin per 41 atoms. A 0.2 mm thick copper strip was folded in the shape of
a cuboidal coil of the same dimension as that of the sample. The
sample was tightly xed inside the coil whose inner surface was
Experimental details covered with a Kapton tape to electrically insulate the sample
from the copper strip. One end of the copper strip coil was
To detect EPR in DPPH sample we measure the magneto-
soldered to the signal line while the other end was soldered to
impedance (MI) of a copper strip surrounding the sample. MI
the ground of a subminiature A type (SMA) coaxial connector.
refers to the variation of electrical impedance (Z(f,H) ¼ R(f,H) +
The radio frequency (rf) current from the impedance analyzer
iX(f,H)) of a material in presence of an applied dc magnetic eld
ows through the strip coil and terminates at the ground of the
(Hdc) at different frequencies of alternating current (f). It
SMA connector creating an rf magnetic eld in the interior of
consists of measuring the magnetic eld dependence of the
the strip coil along the axial direction as shown in Fig. 1. Hence,
magnitude of impedance (Z) alone or resistance (R) and reac-
the DPPH sample experiences an rf axial magnetic eld. An
tance (X) of the sample i.e., magnetoresistance and magneto-
electromagnet is used to apply dc magnetic eld perpendicular
reactance, respectively.23 Magnetoimpedance in so
to the axial rf eld. The resistance (R) and reactive (X) compo-
ferromagnetic metallic conductors has been investigated many
nents of the electrical impedance of the copper strip were
researchers using either an inductance–capacitance–resistance
simultaneous measured at different frequencies of rf current
(LCR) meter in the frequency range (10 kHz to few tens of MHz)

Fig. 1 (a) Schematic diagram of a copper strip coil soldered to a SMA connector which is connected to a radio-frequency impedance analyzer. irf
is the radio-frequency current in the strip coil, hrf is the radio-frequency magnetic field and it is perpendicular to the dc magnetic field Hdc. (b)
Resistance (R) and reactance (X) of DPPH molecule as a function of dc magnetic field (Hdc) as measured by the strip coil at room temperature with
excitation frequency f ¼ 2 GHz.

17312 | RSC Adv., 2020, 10, 17311–17316 This journal is © The Royal Society of Chemistry 2020
Paper RSC Advances

while sweeping the dc magnetic eld. The electrical impedance ðDHÞ2

Vf 1 df R or X ¼ Ksym
of the copper strip is Z ¼ ¼ ¼ R þ iX; where f is the ðHdc  Hr Þ2 þ ðDHÞ2
I I dt
rf magnetic ux passing through the strip coil given by f ¼
ðDHÞðHdc  Hr Þ
m0mrHrfA. Here, Hrf is the magnetic eld inside the strip-coil and þ Kasym þC (1)
ðHdc  Hr Þ2 þ ðDHÞ2
A is the cross-sectional area of the strip coil. Since the high
frequency permeability is m ¼ m0  im00 where m0 is the in-phase Here, DH and Hr are the line widths and resonance elds cor-
and m00 is the out-of phase of the permeability which describe responding to a particular frequency. Ksym and Kasym are the
dispersion and absorption or loss in the sample, respectively. By frequency dependent magnitudes of the absorptive and
00
substituting the complex permeability, we obtain R ¼ Gðum0 mr Þ dispersive components present in the signals and C is
0
and X ¼ Gðum0 mr Þ; where G is a constant depending on the a constant offset. To understand the line shape, we look at the
geometry of the strip coil. Since the high frequency permeability Ksym/Kasym ratio for R and X as obtained from the DPPH sample.
of the paramagnetic DPPH sample is affected by the application On tting R at 2 GHz it was found |Ksym/Kasym| ¼ 4.5658, indi-
of dc bias magnetic eld, resistance and reactance of the strip cating R is dominated by the symmetric component while for X,
coil also changes. R and X were recorded without and with the |Ksym/Kasym| ¼ 9.6943  104 indicating the line shape is
sample inside the strip coil and data for each frequency and dominated by the dispersive component. In conventional EPR
magnetic eld were subtracted to obtain only the sample spectrometers, the derivative of the power absorbed is usually
contribution. As the impedance analyzer measures the radio- measured and the line shape is tted to the Dysonian equation
frequency current–voltage characteristics of the strip coil, given by eqn (2):35
there is no need extra analysis to extract the impedance from " ! !#
dP d DH aðHdc  Hr Þ
scattering “S” parameters as in a network analyzer. f þ
dHdc dHdc ðHdc  Hr Þ2 þ DH 2 ðHdc  Hr Þ2 þ DH 2
(2)
Results and discussions
Fig. 1(b) shows the magnetic eld dependence of R and X for an The rst term in the above equation describes the absorption
rf current excitation of frequency 2 GHz in the strip coil. As Hdc while the second term represents the dispersion. a denotes the
is swept from Hdc ¼ 1 kOe to Hdc ¼ 0 Oe, R rapidly increases in dispersion-to-absorption ratio. The asymmetricity is prominent
a narrow eld range and exhibits a sharp peak at 712 Oe in conducting samples since the electric and magnetic rf
whereas X shows a sudden jump around the same eld. These components in conducting samples become out of phase with
features reect the absorption and dispersion of the complex each other leading to an admixture of the dispersion into the
susceptibility, i.e., c00 and c0 , respectively, in the vicinity of absorption spectra. a ¼ 0 when the skin effect is negligible as in
electron–spin resonance.31 Conventional EPR spectroscopy with insulating samples while a ¼ 1 for highly conducting samples
lock-in detection technique measures the eld derivative of the where the skin depth is very small compared to the sample size.
rf power absorbed, dP/dH, which is proportional to dc00 /dH. The In this case the absorption and dispersion are of equal strength.
dispersive signal is rarely reported except in a few experi- So, a dP/dH measurement alone cannot isolate the absorption
ments32–34 even though it offers a better understanding of the and dispersion effect whereas, the R and X responses from the
spin dynamics in a material. The R and X responses were tted magnetoimpedance measurements can provide this informa-
to eqn (1) and are shown as solid lines in Fig. 1(b) which tion and enable accurate analysis of the physical parameters.
contains both a symmetric Lorentzian term (rst component) In Fig. 2(a), the eld dependence of R for various frequencies
and a dispersive antisymmetric term (second component). from 1.5 GHz to 2.2 GHz are shown. The peak in R shis

Fig. 2 (a) Resistance (R) of the copper strip coil enclosing the DPPH sample as a function of Hdc for different frequencies (f) of current in the strip
coil. (b) Plot of f vs. Hdc with open circles used to depict the resonance fields (Hr) and the solid line illustrating the linear relationship indicating
EPR.

This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 17311–17316 | 17313
RSC Advances Paper

Fig. 3 (a) The EPR spectroscopic signal (dP/dH) for the DPPH sample measured using the Cryo-FMR spectrometer for excitation frequencies of 2
GHz (squares), 4 GHz (circles), 10 GHz (upward
Ð Ð Ð dP/dH curve at 2 GHz along with the line shape fit
triangle) and 16 GHz (downward triangle). (b)
using eqn (3), the single integrated signal: (dP/dH)dHdc and the double integrated signal: (dP/dH)dHdc of the curve fitted dP/dH curve. Inset:
plot of f vs. Hdc with open circles used to depict the resonance fields (Hr) obtained using the dP/dH data.

towards a higher magnetic eld with increasing frequency of resonance eld (Hr) which corresponds to the zero crossing
current. We performed the line shape analysis for all the point and amplitude of dP/dH increase with increasing
frequencies and extracted the frequency dependent line widths frequency. The inset in Fig. 3(b) shows Hr increasing linearly
(DH) as well as the resonance elds (Hr) using eqn (1). It is with frequency and g/2p ¼ 2.801 GHz kOe1, which is close to
known that the resonance frequency (fr) for EPR is proportional the value observed in the MI measurement. The dP/dH line
g
to the dc magnetic eld and follows the relation fr ¼ Hdc shape was tted to eqn (3):
2p
where g is the gyromagnetic ratio (g ¼ gmB/ħ, where g is the dP 4DHðH  Hr Þ
¼ Aasym h i2
Landé g factor, mB is the Bohr magneton and ħ is the reduced dH
4ðH  Hr Þ2 þ ðDHÞ2
Planck's constant). Therefore, with increasing Hdc the reso-
nance frequency increases linearly. This linear behavior was ðDHÞ2  4ðH  Hr Þ2
 Asym h i2 þ C (3)
observed in the plot of fr vs. Hdc presented in Fig. 2(b) and we 4ðH  Hr Þ2 þ ðDHÞ2
obtain g/2p ¼ 2.799  0.0276 MHz Oe1. This g/2p value
corresponds to a Landé g value of 1.999  0.0197 which is well where, Aasym and Asym are the frequency dependent magnitudes
within the reported value of 2.0036. The small error in the g of the absorptive and dispersive components present in the dP/
value is possibly due to inhomogeneity in dc magnetic eld or dH signal. The intensity of absorption by the sample is directly
determination of the magnetic eld. The line width in this proportional to the relative numbers of unpaired electrons in
frequency range was about 2 Oe which is consistent with the the sample. Therefore, a double integration of the derivative
dilute nature of paramagnetic species (free radicals) in DPPH.36 spectrum of absorbance can be used to estimate the spin
To verify the results obtained through the MI method, we concentration which can be utilized for quantitative EPR
measured the EPR spectra with a broad band ferromagnetic studies. In Fig. 3(b), the line shape t, the single integration and
resonance spectrometer (Cryo-FMR by NanoOsc™ from the double integration of the dP/dH signal is presented.
Quantum Design Inc. USA). This spectrometer makes use of the In Fig. 4(a) R is presented for various angles which Hdc makes
lock-in technique and records the derivative of power absorbed with hrf. When Hdc is perpendicular to hrf the signal is the most
(dP/dH) by the DPPH sample placed on top of a wave guide intense while it disappears when Hdc is parallel to hrf. The R
while Hdc is swept for xed rf excitations of 2 GHz, 4 GHz, 10 response for different masses of DPPH is also presented in
GHz and 12 GHz as shown in Fig. 3(a). We can see that the Fig. 4(b). The signal strength is proportional to the mass of the

Fig. 4 (a) Resistance (R) of the copper strip coil enclosing the DPPH when hrf makes 90 , 45 and 0 with Hdc. (b) Variation of R with Hdc at 2 GHz
for 23 mg, 19 mg and 13 mg of DPPH. Inset: single integration of the R curve for different masses of DPPH.

17314 | RSC Adv., 2020, 10, 17311–17316 This journal is © The Royal Society of Chemistry 2020
Paper RSC Advances

DPPH with 23 mg of DPPH exhibiting the largest response. 5 K. L. Brower and S. M. Myers, Appl. Phys. Lett., 1990, 57, 162.
Since R is proportional to c00 , a single integration of the R 6 V. V. Khramstov, A. A. Bobkov, M. Tseytlin and
response can be used to estimate the number of spins in the B. Driesschaert, Anal. Chem., 2017, 89, 4758.
sample. The single integration of R is presented in the inset of 7 S. Van Doorslaer and D. M. Murphy, EPR Spectroscopy:
Fig. 4(b) which provides the EPR intensity of absorption by the Applications in Chemistry and Biology, ed. M. Drescher and
DPPH samples and increases with increase in number of spins. G. Jeschke, Springer-verlag, Berlin, 2012, pp. 1–29.
For 13 mg, 19 mg and 23 mg of DPPH, we obtain 1.9853  1019, 8 M. Che and E. Giamello, Catalyst Characterization, ed. B.
2.9017  1019 and 3.5125  1019 spins, respectively. Imelik and J. C. Vedrine, Springer, Berlin, 1994, p. 131.
Although the presented strip coil method does not need 9 E. Reijerse and A. Savitsky, Electron Paramagnetic Resonance
microfabrication, use of this method is constrained by the Instrumentation in eMagRes, ed. R. K. Harris and R. L.
natural frequency of the strip coil and maximum frequency of Wasylishen, 2017, DOI: 10.1002/
the signal sourced (f ¼ 3 GHz) by the impedance analyzer. In 9780470034590.emrstm1511.
this study, natural resonance of the strip coil was around 2.6 10 G. R. Eaton, EPR Imaging and In-vivo EPR, CRC Press,
GHz. However, one can prepare a cylindrical coil or a cuboidal London, 2018.
coil depending on the shape and physical state (thin lm/bulk 11 C. Clauss, M. Dressel and M. Scheffler, J. Phys.: Conf. Ser.,
sample/powder) of the sample and it can be extended to study 2015, 592, 012146.
liquid samples as well. We had successfully tested for ferro- 12 Y. Wiemann, J. Simmendinger, C. Clauss, L. Bogani,
uids (not shown here). These limitations should be taken into D. Bothner, D. Koelle, R. Kleiner, M. Dressel and
consideration. M. Scheffler, Appl. Phys. Lett., 2015, 106, 193505.
13 K. Jing, Z. Lan, Z. Shi, S. Mu, X. Qin, X. Rong and J. Du, Rev.
Conclusions Sci. Instrum., 2019, 90, 125109.
14 H. Malissa, D. I. Schuster, A. M. Tryshkin, A. A. Houck and
In summary, we have presented a simple technique to probe S. A. Lyon, Rev. Sci. Instrum., 2013, 84, 025115.
EPR in the standard DPPH sample by passing high frequency 15 B. Johansson, S. Haraldson, L. Pettersson and O. Beckman,
currents in a copper strip coil which surrounds the sample Rev. Sci. Instrum., 1974, 45, 1445.
while measuring the magnetic eld dependence of electrical 16 E. D. Dahlberg and S. A. Dodds, Rev. Sci. Instrum., 1981, 52,
impedance. We analyzed the line shapes using symmetric and 472.
asymmetric Lorentzian functions and the g-value was extracted. 17 A. C. Torrezan, T. P. M. Alegre and G. Medeiros-Ribeiro, Rev.
The electrical detection of EPR signal using an impedance Sci. Instrum., 2009, 80, 075111.
analyzer which is traditionally used to characterize dielectric 18 A. Ghirri, C. Bonizzoni, M. Righi, F. Fedele, G. Timco,
samples could aid the broad scientic community to probe EPR R. Winpenny and M. Affronte, Appl. Magn. Reson., 2015, 46,
and understand the spin dynamics in the low frequency regime 749.
of the microwave spectrum. 19 S. K. Misra, Multifrequency electron paramagnetic resonance:
theory and applications, John Wiley & Sons, 2011.
Conflicts of interest 20 G. R. Eaton and S. R. Eaton, EPR: Instrumental Methods, ed. L.
J. Berliner and C. J. Bender, Springer, New York, 2004, pp.
The authors declare no conict of interest. 59–114.
21 A. Kitagawa, J. Sens., 2011, 2011, 813636.
Acknowledgements 22 H. Ueda, Z. Kuri and S. Shida, J. Chem. Phys., 1962, 36, 1676.
23 L. V. Panina, K. Mohri, K. Bushida and M. Noda, J. Appl.
R. M. acknowledges the Ministry of Education, Singapore (Grant Phys., 1994, 76, 6198.
number: R144-000-381-112). 24 A. E. P. De Araujo, F. L. A. Machado, F. M. De Aguiar and
S. M. Rezende, J. Magn. Magn. Mater., 2001, 226, 724.
References 25 M. A. Corrêa, F. Bohn, A. D. C. Viegas, M. A. Carara,
L. F. Schelp and R. L. Sommer, J. Magn. Magn. Mater.,
1 E. K. Zavoisky, Zhur. Eksperiment. i Theoret. Fiz., vol. 15, 1945, 2008, 320, 25.
pp. 344–350For historical perspective, see G. R. Eaton, 26 J. M. Gonzalez, A. Garcia-Arribas, S. V. Shcherbinin,
S. S. Eaton and K. M. Salikhov, Foundations of modern EPR, V. N. Lepalovkij, J. M. Collantes and G. V. Kurlyandskaya,
World Scientic, Singapore, 1998, pp. 45–50. Measurement, 2018, 126, 215.
2 M. M. Roessler and E. Salvadori, Chem. Soc. Rev., 2018, 47, 27 G. V. Kurlyandskaya, S. V. Shcherbinin, S. O. Volchkov,
2534. S. M. Bhagat, E. Calle, R. Perez and M. Vasquez, J. Magn.
3 K. L. Bower, Rev. Sci. Instrum., 1977, 48, 135; P. M. Lenhan Magn. Mater., 2019, 459, 154.
and J. F. Conley Jr, J. Vac. Sci. Technol., B: Microelectron. 28 N. A. Buznikov, A. P. Safronov, I. Orue, E. V. Golubeva,
Nanometer Struct., 1998, 16, 2134; F. Jelezko and V. N. Lepalovskij, A. V. Svalov, A. A. Chlenova and
J. Wrachtrup, Phys. Status Solidi A, 2006, 203, 3207. G. V. Kurlyandskaya, Biosens. Bioelectron., 2018, 117, 366.
4 K. A. Muller, W. Berlinger and F. Waldner, Phys. Rev. Lett., 29 U. Chaudhuri and R. Mahendiran, Appl. Phys. Lett., 2019,
1968, 21, 814. 115, 092405.

This journal is © The Royal Society of Chemistry 2020 RSC Adv., 2020, 10, 17311–17316 | 17315
RSC Advances Paper

30 U. Chaudhuri and R. Mahendiran, J. Magn. Magn. Mater., 34 J. S. Hyde, W. Froncisz and A. Kusumi, Rev. Sci. Instrum.,
2018, 488, 165377. 1982, 53, 1934.
31 A. G. Marshall and D. C. Roe, Anal. Chem., 1978, 50, 756. 35 V. A. Ivanshin, J. Deisenhofer, H.-A. Krug Von Nidda,
32 P. Fajer and D. Marsh, J. Magn. Reson., 1983, 55, 205. A. Loidl, A. A. Mukhin, A. M. Balbashov and M. V. Eremin,
33 C. Mailer, H. Thomann, B. H. Robinson and L. R. Dalton, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61, 6213.
Rev. Sci. Instrum., 1980, 51, 1714. 36 W. R. Hagen, J. Phys. Chem. A, 2019, 123, 6986.

17316 | RSC Adv., 2020, 10, 17311–17316 This journal is © The Royal Society of Chemistry 2020