Unconventional electrical

response in chemically-prepared
Gold-Silver Nanostructures

Aishik Panja
A thesis presented for the degree of

Master of Science (Research)

Undergraduate Department

Indian Institute of Science

Bengaluru, India

June 2021
 Unconventional electrical response in
chemically-prepared Gold-Silver
Nanostructures
Aishik Panja

Abstract

This work builds on the recent reports of room-temperature super-

conductivity in gold-silver nanostructures [Anshu Pandey et. al.]. In

addition to the coexistence of zero resistance and diamagnetism, they

show an unusual optical response. In my Bachelor’s project, we had

prepared Au-Ag NSs using glancing angle vapour deposition and had

seen some abnormalities (non-metallic behaviour) in the electrical res-

istance as it was cooled and subsequently warmed up. However, here

we have tried to reproduce the results of Anshu Pandey et. al. and,

therefore, opted for the chemical method as described in their papers.

We also performed absorption spectroscopic measurements with the

NPs in solution just before film preparation. We observe nonmetal-

like behaviour in most of the samples - a V-shaped R vs T being the

common motif. There are also several samples where the resistance had

huge jumps and drops and it even becomes negative at some points.

These will be discussed in some detail in chapter 2.

We also fit the spectroscopy data with theoretical models to get an

estimate of the volume fraction of the constituents, among other para-

meters.

Supervisor: Prof. Ambarish Ghosh

i
 Acknowledgements

I would like to thank Prof. Ambarish Ghosh for giving me the opportunity
to work in his group. I am also indebted to my seniors - graduate students
Ms. Priyanka Suri, Mr. Eklayva Vashisth and project workers Ms. Paramita
Modak and Mr. Ankur Jaiswar for their valuable guidance and support.
I am also grateful to the Kishore Vaigyanik Protsahan Yojana (KVPY) gran-
ted by the Government of India for my undergraduate stipend that helped
keep my finances running.

ii
 Contents

Declaration iv

1 The plasmonics of Au-Ag NSs 1

2 Preparation and Electrical Characterization 7

2.0.1 Sample Preparation . . . . . . . . . . . . . . . . . . . . . . 7

2.0.2 Electrical Measurements . . . . . . . . . . . . . . . . . . . . 9

3 Optical Characterization 38

3.0.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

3.0.2 Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . 41

4 Appendix A 45

5 Appendix B 48

6 Bibliography 53

iii
 Declaration

The work in this thesis is based on research carried out at the Centre for Nanos-
cience and Engineering, Indian Institute of Science, Bengaluru, India. No part of
this thesis has been submitted elsewhere for any other degree or qualification, and
it is the work of the author unless referenced to the contrary in the text.

Copyright c 2021 by Aishik Panja.

“The copyright of this thesis rests with the author. No quotation from it should be
published without the author’s prior written consent and information derived from
it should be acknowledged”.

iv
 Chapter 1
The plasmonics of Au-Ag NSs

Introduction In this chapter, we discuss the physics of localized surface plas-

mons in gold-silver nanostructures. These plasmonic modes naturally arise as a

result of absorption and scattering of light from nano-particles that are well within

the dipole limit (i.e., have size much lower than the wavelength of incident light)

(Mie Theory)[1,2,3] . This is reflected in the absorption spectroscopy of the gold-

silver nanoparticle solution that we prepare using a chemical method (detailed in a

later chapter). The data is then fit to a Mie Theoretic model that also incorporates

Effective Medium Theory - making it possible to determine the volume fractions

of gold and silver in solution.

Localized Surface Plasmons Localized surface plasmons are non-propagating

excitations of the conduction electrons of metallic nanostructures coupled to the

electromagnetic field. When metal NSs are excited with electromagnetic radiation,

typically visible light in the case of gold and silver, the metal electrons are driven to

resonance and this leads to an enhancement in electric field both inside and outside

the nanoparticle. The aforementioned resonance is also known as localized surface

plasmon resonance (LSPR). To solve this problem mathematically, we consider a

metal sphere of dielectric constant (ω) in a medium (dieclectric constant ) in an

external electric field E0 . Next, we solve the Laplace Equation in the system. This

1
 1. The plasmonics of Au-Ag NSs

has been has been workedf in Appendix A of the author’s Bachelor of Science thesis
[1] as well as several textbooks [3,4] on plasmonics and optics.

Figure 1.1: A metallic sphere ((ω)) immersed in a homogeneous medium

(m ) with a constant electric field E0 acting along the z axis. θ is the azi-
muthal angle. Picture taken from Maier, S.A., Plasmonics-Fundamentals
and Applications, Springer (2007), pg. 66, Fig. 5.1

The electric fields are given by -

3m
E~in = E0 (−cosθr̂+sinθθ̂) (1a)
(ω) + 2m   
(ω) − m 2E0 (ω) − m E0
E~out = −E0 cosθ  a3  a3 
   
1 +  r̂ + E0 sinθ 1 −  θ̂(1b)

(ω) + 2m r3 (ω) + 2m r 3

E~out can be re-written as

(ω) − m a3
E~out = E0 (−cosθr̂ + sinθθ̂) − E0 (2cosθr̂ + sinθθ̂). While the first part
(ω) + 2m r3
corresponds to E0 ẑ - the applied field, the seconds part is the induced electric field

that gives the dipole moment and polarizability. Noting[5] that the electric field
(2cosθr̂ + sinθθ̂)
~ =p
due to a dipole of moment p, at ~r and θ is given by E , we have
4πm r3
(ω) − m
p = 4πm a3 E0 (2a)
(ω) + 2m
The polarizability (p/E0 ) is given by :
(ω) − m
α(ω) = 4πm a3 (2b)
(ω) + 2m
The condition for resonance, also called the Frolich Condition, is given by Re((ω)) =

−2m . This is because the medium is a dielectric (Im(m ) = 0), implying Re((ω)+

2m ) = 0 at the Frolich condition.

2
 1. The plasmonics of Au-Ag NSs

Modification of Optical Cross-sections due to LSPRs The absorp-

tion and scattering cross section (CS) (together called extinction CS) due to metal

nanoparticles can be derived from Mie Theory[2] .

The absorption and scattering CSs are given by :

k 4 |α|2
σscat (k) = and σabs (k) = kImag(α)
6π
Here k is the wave vector.

Putting the value of α(ω) from Eqn. 2b, we get

2  
8πk 4 a6 (ω) − m  (ω) − m 
σscat (k) = and σabs (k) = 4πka3 Imag   (3a,b)
3 (ω) + 2m (ω) + 2 

m

Relating the CSs with optical data Consider a particulate medium

(a solution of metal NPs in a solvent, for example) of width h (the width of the

cuvette in case of a solution) that is irradiated with light of intensity Ii (ω) only to

be attenuated to It (ω). Assuming that the light does not suffer multiple scattering

or reflection at the surface of the container, Ii and It are related as -

It
= exp(−N (σext + σsca )h) (4)
Ii
where N is the number density.

Clearly, this is a restatement of Beer-Lambert Law[6,7] . We usually measure the

sample in the cuvette made of quartz which has near 100% transmission.

Also, It /Ii is the transmittance. The absorbance is given as −log10 (It /Ii ). There-

fore the absorbance equals

A = −log10 exp(−N (σext +σsca )h) (5)

Here, we fit the above equation to our absorbance data with several parameters
√
- fAu - the volume fraction of gold, N - the number density, nmed (= m ) - the

refractive index of the surrounding solution.

We use the Maxwell Garnet Method[8,9] to calculate (ω) from fAu and fAg (=

1 − fAu ).

3
 1. The plasmonics of Au-Ag NSs

Maxwell Garnett Method (MGM)[1,9] The statement of MGM is the fol-

lowing : Consider a matrix of (frequency dependent) dielectric constant h on

which spherical inclusions of dielectric constants i of small volume fractions fi are

dispersed. The effective dielectric constant ef f is given by solving the following

equation :

ef f − h i − h
. = Σfi (6)
ef f + 2h i + 2h
Thus for a two component system (gold and silver), we have
ef f − Ag Au − Ag
. = fAu (7a)
ef f + 2Ag Au + 2Ag
 
 2fAu (Au − Ag )
Ag 
1 +

Au + 2Ag 

This leads to : ef f = (7b)

fAu (Au − Ag )
1−
Au + 2Ag
Here, (ω) =ef f

To be noted is that MGM is valid only at extremely low volume fractions (< 10−5 )

such that the inclusions are spatially separated from the matrix and do not interact

electrostatically. On the other hand the Bruggemann’s Method[10] is more universal

as it can account for small volume fractions, places the matrix and the inclusions

on equal footing and incorporates arbitrarily-shaped nanoparticles oriented in pref-

erential directions. However, it has been shown in this author’s Bachelor’s Thesis

that the ef f calculated from the two methods is second order in f , where f is the

volume fraction of the rarer constituent in a two-component mixture. It was also

derived that all systems - irrespective of the shape and orientation of the inclusions

- are equivalent to one having spherical, isotropic inclusions[1,11,12] .

Literature values of (ω) of gold and silver It is very important to

choose the appropriate values of the frequency dependent permittivity of gold and

silver. There are several sources at our perusal - we use the emperical data from

4
 1. The plasmonics of Au-Ag NSs

Ciesielski et. al.[13,14] . We have departed from the Johnson and Christy data[15]

that is conventionally used in these cases as it pertains to the bulk metal whereas

the former describes thin films grown on SiO2 subtstrates. The chart is provided in

the appendix. The  of gold and silver are plotted below as a function of wavelength.

Figure 1.2: The dielectric constant of silver in the 200 nm (0.2µm) - 900nm
range. From Ciesielski et. al.

Figure 1.3: The dielectric constant of gold from Ciesielski et. al.

Why focus on plasmonics ? As touched upon in the introduction, Anshu

Pandey et. al. had observed an unconventional optical response that cannot be

5
 1. The plasmonics of Au-Ag NSs

explained by the theory of surface plasmons in a gold-silver composite of any com-

position (more details in Chapter 3). from plasmonics, the plasmon resonances

of gold and silver are at 2.3 eV and 3.2 eV respectively; therefore, we expect the

composite to have a peak in both absorption and extinction to lie in this range.

However, the following features were observed :

• The absorbance is identically near-zero over the entire spectrum.

• The extinction has no sharp peak anywhere - only a broad hump-like curve

that has its maxima at around 4 eV is seen.

However, in our experiment, we did not notice any such unusual feature. In fact,

our measured absorption matched with theory in the visible regime to a great

extent.

6
 Chapter 2
Preparation and Electrical
Characterization

2.0.1 Sample Preparation

As discussed in the author’s bachelor’s thesis [1] , samples were thermally deposited

using Glancing Angle Deposition (GLAD) in a vacuum chamber.

However, in this project, we have prepared our samples using the wet chemical

methods as described in these papers[16,17,19,20] . The idea is to first prepare silver

nanoparticles and subsequently embed then in gold. We follow this protocol -

1. Chemicals Required

Hexadecyltrimethylammonium bromide (CTAB), Sodium Hydroxide, Ammonium

Bromide, potassium iodide , silver nitrate, sodium borohydride powder, hydrogen-

tetrachloroaurate(III)trihydrate (HAuCl4 ), propan-2-ol (IPA) and Acetone.

2. Procedure

Aqueous solutions of 0.1M CTAB (10 mL), 0.1M NaOH(0.1 ml), 1 M ammonium

bromide(1 mL), 0.1 M potassium iodide (40 µL) and 1 mM silver nitrate(5 mL)

7
 2.0.1. Sample Preparation

were mixed. The mixture was continuously stirred at 1000 rpm till slight cloud-

iness appears followed by addition of 2 mL of 0.1 M aqueous solution of sodium

borohydride to produce the silver nanoparticles.

Then 800 µL of 1mM aqueous solution of HAuCl4 was added to the silver nan-

oparticles followed by 3 ml of IPA and subsequently dissolving in acetone in 1:1

ratio (amount of solvents varied at each run). There is an instantaneous colour

change from golden brown/ maroon to greyish black. It is then centrifuged for 3-4

min at 6500 revolutions per minute (RPM). After centrifugation, the supernatant

was discarded, and the precipitate was dried in vacuum.

Note that the amount of each reagent may be varied at each run; in later runs, we

scale it by 2.

3. Film Casting Procedure

The dry precipitate was dissolved in around 400 µL of chloroform. In parallel we

make a saturated solution of Potassium Hydroxide in IPA. The drop casting drill

is the following - Drop cast a small quantity (2µl to 10µl, depending on the desired

size of the spot) of the chloroform solution → add the KOH/IPA solution to it

and wait for 1 minute → rinse thoroughly with Methanol. This process is to be

repeated till one gets a smooth film with electrical continuity.

The drop cast is carried out inside a makeshift plastic container with sliding walls

that is covered in yellow sheet to reduce exposure to ambient light and its top

surface covered with 3 layers of a filtering cloth. Additionally, nitrogen gas is

blown into the chamber so as to create an inert environment. Finally, some of

the later samples are covered with the epoxy AralditeT M to protect the film from

the elements while they are removed from the box to be prepared for electrical

measurements.

Note that we keep on making changes to the reaction conditions such as quantity

of the constituents, the speed of centrifugation and the drop-casting environment.

Substrates

8
 2.0.2. Electrical Measurements

We use quartz substrates that has electrodes/pads made of chromium and gold for

the electrical connections. The pads are fabricated at the Annexe of the National

Nanofabrication Centre, CeNSE using vapour deposition.

2.0.2 Electrical Measurements

Introduction Electrical measurements are carried out in a refrigerator that

consists of a cylindrical chamber having a cuboidal protrusion at its bottom where

optical windows are located. Inside this chamber, there is a smaller liquid nitrogen

chamber - one end of that is attached to the lid of the main chamber and the other

end attached to a copper stub that fits into the aforementioned protrusion. A prin-

ted circuit board (PCB) is attached to this stub that has several electrodes which

are connected to the substrate pads via copper wires. These wires are tethered

to the substrate with indium wires. There is a Pt-100 thermometer and a heater

attached to the stub as well.

We have a resistance box made of aluminium that has several resistors of resist-

ances between 10 Ohm and 106 Ohm connected between Bayonet Neill–Concelman

(BNC) terminals. We send current and voltage to the PCB, and therefore our

sample, using coaxial cables via a Military connector. The other end of these

cables is connected through coaxial-to-BNC adapters to another aluminium box

(call it the terminal box) that has BNC connectors on the other side.

As we want to measure the four-probe and two-probe resistances simultaneously,

We have two Stanford Research SR 830 DSP lock-in amplifiers. One of them (call

this Lock-in 1) supplies a 1 V reference signal of frequency 227.7 Hertz across the

10,000 ohm resistor (also called load resistance) on the resistance box and the res-

ultant current of 10−4 A(= 1V

104 Ω
) is sent to the I+ terminal of our sample located on

the terminal box. The I- terminal is grounded using a 50 ohm BNC terminator. For

the two-probe measurement, we connect the I+ and I- terminals through T-shaped

adapters and BNC cables to Lock-in 1. Similarly, for the four-probe resistance, we

9
 2.0.2. Electrical Measurements

connect the V+ and V- terminals to the other lock-in (Lock-in 2). We connect the

REF IN terminal of Lock-in 2 to the TTL OUT terminal of Lock-in 1 to complete

the circuit.

The values of voltage displayed on the Channel 1 of the lock-ins are divided by the

input current (104 A) to get the resistance is ohms.

The cables from the heater and thermometer also pass through the military con-

nector to a LakeShore temperature controller.

The lock-ins and the temperature controller are operated on LabVIEW via a GPIB

interface. The lock-ins are interconnected using an IEEE 488 GPIB which also con-

nects Lock-in 2 with the the temperature controller. The Lock-in 1 is connected to

the computer through a GPIB-USB-HS Instrument Control Device. At the start

of the experiment, the phase of both the lock-ins as displayed in Channel 2 are set

to zero. The phase of the reference signal is also set to zero.

We attach the substrate to the stub with strips of carbon tape. The entire stub,

except the sample substrate, is covered with multiple layers of super-insulation.

After loading the sample, the refrigerator is pumped using a Pfeiffer HiCube 80

pump for a few hours at the maximum frequency of 1500 Hz. Subsequently, we set

the temperature in the temperature controller to 90 K and pour liquid nitrogen

from the top. After the desired temperature is reached, we turn on the heater to

either high or low depending on the required speed of heat-up.

10
 2.0.2. Electrical Measurements

a)

b)

Figure 2.1: a) The stub with carbon tapes stuck. Also seen is the PCB and
the copper wires attached to it.
b) The upper part of the cryogenic chamber resting on the lower part. It
contains i) the LN2 chamber (shown wrapped in super-insulation), ii) the
tubes on the top of the lid for electrical connections and an inlet for liquid
nitrogen (white funnel) and iii) the stub attached to the LN2 chamber.

Results Sample 1
This was our very first attempt at making the sample according to the above

protocol. Earlier, we had tried another protocol (Protocol 1[16,17,19,20] which had

yielded our desired Ag-Au solution (as confirmed by spectroscopy) but the drop-

cast film did not have electrical continuity.

We follow the protocol (Protocol 3 of Supplementary Information)given above ver-

11
 2.0.2. Electrical Measurements

batim, except for the amount of HAuCl4 (2.5 ml instead of 800 µl). The drop-cast

was performed on a hot plate that allows the liquids (IPA, methanol and chloro-

form) However, the sample was loaded into the vacuum refrigerator 13 days after it

was prepared. During this time, it was often kept undesiccated and understandably

got oxidised.

Sample 2

Sample 2 was the first sample that was put for electrical measurement right after

it was prepared. During preparation, the solution was heated at 600 C while being

stirred. The amount of the constituents was as mentioned in Protocol 3 except for

HAuCl4 - we added 2 ml of HAuCl4 instead of the prescribed 800 µl. We had also

added 7 ml of IPA (and no acetone) to the ≈ 18 ml of the final solution immediately

after it was prepared and centrifuged it for 5 minutes at 5500 rpm.

The substrate has six parallel pads that are labelled 1 through 6 sequentially. The

drop-cast was carried out on a hot plate that was heated to 60-70 0 C to accelerate

the drying process. The dropcast film covers the four pads numbered 2 through

5. In the 4-probe measurement, we send in current through terminals 2 and 5 and

measure voltage between terminals 3 and 4.

12
 2.0.2. Electrical Measurements

a b

c d
Figure 2.2: (a) and (b) are respectively the 2 probe and 4 probe resistance values
as a function of temperature over the entire cycle encompassing the cool-down and
heat-up processes(c) and (d) are the 2-probe and 4-probe values plotted together
in the cool-down and heat-up processes, respectively .

It is evident that none of these processes point to a typical metal that follows a

linear R vs T profile with a positive slope. Moreover, the heat-up and cool-down

profiles are totally different and they do not even converge at room temperature,

implying a permanent change in the film structure over a thermal cycle. In the

cool-down process, the 4-probe and 2-probe resistances follow a similar profile,

indicating that the film is homogeneous at macroscopic scale before deformation

sets in.

Sample 3 Sample 3 was prepared exactly according to Protocol 3 and cent-

rifuged at 5000 rpm for 5 minutes. However, we used bigger centrifuge tubes (10

13
 2.0.2. Electrical Measurements

ml) to centrifuge the ≈ 45 ml of final solution as opposed to 2 ml tubes. As a result,

a major portion of the pellets ended up getting stuck to the walls of the tubes and

could not be dissolved in chloroform. Despite this, we had drop-cast the solution

and were able to get electrical continuity between the electrodes numbered 3 and

4, although the 4-probe was non-existent (Overload).

a)
b)

Figure 2.3: The a) film after it was unloaded from the refrigerator and the b)
2 probe resistance (R34 ) over one thermal cycle (as a function of temperature).

The 2-probe resistance decreases with decreasing temperature till ≈ 225 K, after

which it starts [Link] is quite similar to the Sample 2 profile. During heat-

up, the resistance continuously increases until it reaches 250 K when it flattens out

and eventually starts decreasing. There is a sharp drop in the 300 K - 320 K range

and when cooled back to the room temperature (297 K), it ends up at a much

lower resistance than where it started. Thus, at room temperature, we have three

distinct values of resistances - one at the beginning (≈ 49.85Ω), another during

heat-up (≈ 51.2Ω) and the final resistance (≈ 47.7Ω). In hindsight, these values

may not seem too far off - these are indeed only 3% off each other. However, these

points lie at the extremities of the R vs T graph and are therefore significant.

14
 2.0.2. Electrical Measurements

Sample 4 We prepared Sample 4 exactly according to the protocol given at

the beginning of the chapter. During preparation, the solution was heated at 600 C.

This time, the drop-cast was performed inside a box that had a hairdryer attached

to it to circulate hot air inside the chamber.

a) b)

Figure 2.4: The a) Sample 4 after it was unloaded from the refrigerator and
the b) The chamber with the hairdryer attached at the top.

a) b)

Figure 2.5: The a)The 2-probe (R25 ) and b)4-probe resistance (R2534 ) of
Sample 4 as a function of temperature over 3 thermal cycles.

The two-probe resistance follows linear (metallic) curves that also show hysteresis.

We confirm this by carrying out 3 thermal cycles. However, the four-probe resist-

15
 2.0.2. Electrical Measurements

ance was very similar in nature to the one seen in Sample 2 in that the cool-down

and heat-up processes are too far off with the resistance at room temperature being

markedly different in the two processes. In what is an extreme case of hysteresis,

the resistance at any given temperature increases in every thermal cycle.

Sample 5 We prepared Sample 5 exactly according to the protocol, only this

time the solution was not heated while being stirred. The substrate was plasma

treated using a Harrick PlasmaBasic Plasma Cleaner and drop-cast carried inside

the aforementioned chamber. As a result, this film is significantly more consistent

than the previous ones.

Figure 2.6: The Sample 5 thin film. Photograph taken after electrical meas-
urements.

16
 2.0.2. Electrical Measurements

a) b)

Figure 2.7: The a)The 2-probe (R25 ) and b)4-probe resistance (R2534 ) of
Sample 5 as a function of temperature over 3 thermal cycles.

The 2-probe resistance profile, though not strictly linear, can be approximated to

that of a metal. The 4-probe has extreme hysteresis - in the cool-down process,

the resistance initially decreases with temperature and starts to increase later. The

4-probe profile follows a completely different profile.

A common feature in Samples 4 and 5 that we did not see in the previous samples

is that the 2-probe and 4-probe profiles are totally different in nature. In other

words, the sample is (almost) metal-like in the 2-probe length scale (≈ 1 cm), but

behaves unconventionally at the smaller 4-probe length scale.

Sample 6 We change the geometry of the pads so as to make the drop-cast

spot smaller in size, thereby minimizing the effects brought by a variable length

scale across the sample. The substrate had eight concentric pads that, at the cen-

ter, had a separation less than 1 mm.

While preparing Sample 6, we had added 1 ml of HAuCl4 as opposed to the pre-

scribed 0.8 ml, while other components were added as per the protocol. We did

17
 2.0.2. Electrical Measurements

not heat the solution while it was stirred.

a) b)

Figure 2.8: The a) The drop cast thin film in Sample 6, photograph taken
after electrical measurements b) The final solution of Sample 6, identifiable
by its golden maroon color.

a) b)

Figure 2.9: The a) The 2-probe and b) 4-probe resistance of Sample 6. The
current was sent through the outer pads and voltage measured over the two
middle pads (marked with indium wires in the previous figure).

Here, both the 2-probe and 4-probe profiles are linear except for temperatures

above 300 K in the heat-up process where the resistance drops sharply. We had

seen this in Sample 3 as well.

18
 2.0.2. Electrical Measurements

Sample 7 sample 7 was prepared exactly according to the aforementioned

protocol and at room temperature. For drop-cast, we used the third type of sub-

strate that has six pads. During the drop-cast, we vented the chamber with Nitro-

gen gas in order to create a more inert environment.

a) b)

Figure 2.10: The a) The drop cast thin film in Sample 7, photograph taken
after electrical measurements, the carbon tape and indium wires still stuck
to the substrate b) The final solution of Sample 7.

Clearly, the film is not smooth at all despite the substrate being plasma cleaned.

a) b)

Figure 2.11: The a) The 2-probe and b) 4-probe resistance of Sample 7. For
the 4-probe resistance, we send current through pads numbered 2 and 6 and
measured voltage between 3 and 4.

The two probe was linear apart from a few blips. The four-probe had two linear

19
 2.0.2. Electrical Measurements

regions with different slopes that exhibited hypothesis. To check how the differ-

ence in resistance between the cool-down and heat-up processes (due to hysteresis)

depends on temperature, we took the sample through cd3 to 229K (yellow curve)

and immediately heated it up (hu3, cyan curve). We infer that hysteresis is an

ever-present feature of this sample - in other words, if we stop the cool-down at

any given point and start heating up, the two curves start diverging.

Sample 8 Sample 8 was prepared according to the protocol given at the

beginning of the chapter. The drop-cast techniques were same as those in Sample

7. However, probably due to over-etching caused by IPA and methanol, the film is

not very smooth and the gold pads tend to peel off. Therefore, we only manage to

get a 2-probe resistance while 4-probe is non-existent i.e., overload.

a)
b)

Figure 2.12: The a) The drop cast thin film in Sample 8, photograph taken
after electrical measurements, the pads looking very worn out b) The final
solution of Sample 8.

20
 2.0.2. Electrical Measurements

Figure 2.13: The R vs T of Sample 8 in a 2-probe configuration (the two

middle pads are used).

The 2-probe profile is mostly linear.

21
 2.0.2. Electrical Measurements

Sample 9 Sample 9 was also prepared exactly according to the protocol.

a) b)

Figure 2.14: The a) The drop cast thin film in Sample 9, photograph taken
after electrical measurements, b) The final solution of Sample 9.

b)
a)

Figure 2.15: The a) The 2-probe and b) 4-probe resistance of Sample 9. For
the 4-probe resistance, we send current through pads numbered 2 and 6 and
measured voltage between 3 and 5.

Apart from the R vs T profile being uncharacteristic of a metal, we observe hys-

teresis in the sample as we did in the previous samples. As in Sample 7, we had

stopped the cool-down at an intermediate temperature (266K, in this case) and

heated the sample up, just to see the deviation (CD2 and HU2 in the figures). We

22
 2.0.2. Electrical Measurements

also observe sharp drops and jumps at several temperatures, including at temper-

atures greater than 300 K.

Sample 10 We scaled up the amount of every constituent by 2 keeping the

concentration intact. We had added 2 ml of HAuCl4 and only IPA (no acetone) in

10:8 (solution:IPA) ratio before centrifuging it. After drop-cast, the film is covered

with a cured PDMS layer.

a)
b)

Figure 2.16: The a) The drop cast thin film in Sample 10, photograph taken
after electrical measurements, the pads looking very worn out b) The final
solution of Sample 10.

a) b)

Figure 2.17: The a) The 2-probe and b) 4-probe resistance of Sample 10. For
the 4-probe resistance, we send current through pads numbered 1 and 5 and
measured voltage between 2 and 3.

23
 2.0.2. Electrical Measurements

Sample 11 We prepare sample 11 according to the protocol and amount

of each component scaled by a factor of 2. The deposition was performed inside

a chamber that was vented with gaseous nitrogen and the film was covered with

Araldite.

a) b)

Figure 2.18: The a) The drop cast thin film in Sample 11, photograph taken
after electrical measurements, b) The final solution of Sample 11.

a) b)

Figure 2.19: The a) The 2-probe and b) 4-probe resistance of Sample 11.
During the heat-up process, the voltage leads must have got disconnected -
hence the 4-probe was overload.

Sample 12 Sample 12 was also prepared according to the aforementioned

protocol. The film was covered with araldite and dried before loading.

24
 2.0.2. Electrical Measurements

a) b)

Figure 2.20: The a) The drop cast thin film in Sample 12, photograph taken
after electrical measurements, b) The final solution of Sample 12.

a) b)

Figure 2.21: The a) The 2-probe and b) 4-probe resistance of Sample 12 as a

function of temperature.

Sample 13 - Au control film Sample 13 is the control gold sample that

is prepared using standard NP preparation techniques - a gold salt (HAuCl4 ) is

added to a capping agent (Polyvinylpyrrolidone or PVP). It is diluted with IPA,

cewntrifuged, precipitate dissolved in chloroform and deposited on the substrate.

25
 2.0.2. Electrical Measurements

a) b)

Figure 2.22: The a) The 2-probe and b) 4-probe resistance of Sample 13 as a

function of temperature.

By and large, this is linear and therefore metallic, as expected.

Sample 14 Sample 14 was prepared according to protocol. The reagent

amounts were scaled by 2 (we added 2 ml of HAuCl4 ) and the film covered by

Araldite after preparation.

a)
b)

Figure 2.23: The a) The drop cast thin film in Sample 14, photograph taken
after electrical measurements, b) The final solution of Sample 14.

26
 2.0.2. Electrical Measurements

a) b)

Figure 2.24: The a) The 2-probe and b) 4-probe resistance of Sample 14.

The R vs T curves for Sample are mostly linear in both the 2-probe and 4-probe.

Moreover, the heat-up and cool-down paths are very close together - no sign of

hysteresis that was seen in some previous samples.

Sample 15 Sample 15 was prepared

b)
a)

Figure 2.25: The a) The drop cast thin film in Sample 15, photograph taken
after electrical measurements, b) The final solution of Sample 15.

27
 2.0.2. Electrical Measurements

a)

b)

Figure 2.26: The a) The 2-probe and b) 4-probe resistance of Sample 15.
Parts of the graph are zoomed in the inset.

In Sample 15, there is a huge spike at temperatures below 150 K that can be

explained as follows - the two phases of the lock-in amplifiers showing the current

and voltage are set to zero. However, at the onset of these sharp spikes, the phase

28
 2.0.2. Electrical Measurements

of the voltage lock-in (that measures the 4-probe resistance) departs from zero

and mostly hovered around 80 degrees. As a result the input current (generated

from the reference voltage sent across the 10kΩ resistor) and the output voltage are

always out of phase leading to the resistance taking these unusual values. However,

as seen in the inset, the sample behaves ’normally’ in the 300K-150K - resistance

going down initially and later increasing.

We are unsure why this happens. Physically this means the input current and

output voltage are out of phase.

Sample 16 Sample 16 was prepared according to protocol. The reagent

amounts were scaled by 2 (we added 1.6 ml of HAuCl4 ) and the film covered by

Araldite after preparation.

a) b)

Figure 2.27: The a) The drop cast thin film in Sample 16, photograph taken
after electrical measurements, b) The final solution of Sample 16.

29
 2.0.2. Electrical Measurements

a ) R (2 probe)

b ) R (4 probe)

Figure 2.28: The a) The 2-probe and b) 4-probe resistance of Sample 16.

Here, the phase difference between the two lock-ins departs from zero at certain

points in the cycle and it is only at those points that the resistance shoots. The

phase difference varies between 70 and 90 degrees.

30
 2.0.2. Electrical Measurements

a) b)

Figure 2.29: The a) The drop cast thin film in Sample 17, photograph taken
after electrical measurements, b) The final solution of Sample 17.

a ) R (2 probe)

b ) R (4 probe)

Figure 2.30: The a) The 2-probe and b) 4-probe resistance of Sample 17.

31
 2.0.2. Electrical Measurements

Sample 17 This time, the phase difference between the signal and reference
exceeds 90 degrees. As a result, the resistance becomes negative (as the time-

averaged output from a lock-in amplifier is proportional to cosθ, where θ is the

phase difference [ref - [Link] We see

this behaviour in Samples 18 and 20 as well.

Sample 18 Sample 18 was prepared according to the protocol and the re-
agents were scaled by a factor of 2. The drop-cast film was covered with araldite.

a)
b)

Figure 2.31: The a) The drop cast thin film in Sample 18, photograph taken
after electrical measurements, b) The final solution of Sample 18.

32
 2.0.2. Electrical Measurements

a ) R (2 probe)

b ) R (4 probe)

Figure 2.32: The a) The 2-probe and b) 4-probe resistance of Sample 18.

As seen in samples 15, 16 and 17, there is a phase mismatch in Sample 18. The

phase difference between the input and signal oscillates between 0 and 180 degrees.

As a result the recorded ’resistance’ is all over the place, even taking negative values

when the phase exceeds 900 .

Sample 19 Sample 19

33
 2.0.2. Electrical Measurements

a)
b)

Figure 2.33: The a) The drop cast thin film in Sample 19, photograph taken
after electrical measurements, b) The final solution of Sample 19.

a ) R (2 probe)

b ) R (4 probe)

Figure 2.34: The a) The 2-probe and b) 4-probe resistance of Sample 19.

34
 2.0.2. Electrical Measurements

Sample 19 does not show the unusual and abrupt phase difference between the ref-

erence and the signal that is accompanied by the resistance taking unconventional

values. This behaves more like Samples 2, 3, 4 etc. where the common theme is an

U or a V shaped R vs T curve.

Sample 20 Sample 20 was prepared according to the aforementioned pro-

tocol. The drop-cast film that was covered with araldite was not consistent enough

to have a stable 4-probe resistance.

a) b)

Figure 2.35: The a) The drop cast thin film in Sample 20, photograph taken
after electrical measurements, b) The final solution of Sample 20.

35
 2.0.2. Electrical Measurements

a ) R (2 probe)

b ) R (2 probe)

Figure 2.36: The a) The 2-probe resistance of Sample 20. Fig b) is the
zoomed-in section of Fig a) in the 250-300 K range that captures the R vs T
behaviour before the sharp drop. During CD, The resistance decreases and
subsequently increases (a staggered V-shaped curve) before the drop. In the
HU process, the phase difference between the signal and reference stays above
90 degrees hence the sharp vertical lines.

As seen in samples 15-18, the phase difference between the output and reference

signal deviates from zero and takes values greater than 90 degrees (between 170

and 180 degrees in this case)

Samples 15, 16, 17, 18 and 20 are unusual in that the signal and reference fall

out of phase at several points in the experiment; as a result the resistance takes

36
 2.0.2. Electrical Measurements

seemingly unphysical values. We want to check if these results are reproducible

Summary Most samples depart from the conventional metallic nature where
the R vs T is a linear curve with a positive slope. The notable features in the

samples are the following :

• During cooldown, in several samples, the resistance first decreases with tem-

perature and later starts to increase. In many cases the upward (increasing)

portion is larger than the decreasing part.

• For most samples, the R vs T curves for the heat-up and cooldown process

are far apart - an extreme case of hysteresis.

• In some samples, there are sharp drops in resistance during heat-up in the 300

K - 320 K range Even if we consider this a result of microscopic deformation

(at the Nanoparticle level) as a result of continuous thermal cycles, the fact

that this occurs so close to room temperature is very surprising.

• In Samples 15-20 (with the exception of Sample 19), the phase of the input

current (generated from the reference voltage of 1 V) deviates from that

of the measured voltage (referred to as signal) across the voltage terminals.

When the phase difference exceeds 90 degrees, the time-averaged output

becomes negative. These phase changes are abrupt- the sample behaves

’normally’ at other times. On many occasions, during heat-up,the phase

difference reverts to its original near-zero value.

37
 Chapter 3
Optical Characterization

3.0.1 Introduction

As elaborated in Chapter 1, Anshu Pandey et. al. had seen an unconventional op-

tical response in the samples that had superconducting transitions[18] . The absorb-

ance had been near-zero throughout the spectrum, whereas the extinction spectrum

was very broad and had a peak at around 4 eV.

Both these features are unusual. It was shown in this author’s Bachelors’ Thesis
[1] (pages 60-64) that this corresponds to unnatural values in the Drude Lorentz fit
[3,21,22] .

We go over those methods and results in brief.

The first task is to extract the  from the optical data. We use the equations (Equa-

tions 3a,b) relating the extinction and absorption coefficients with the (ω), med

and the size and number density of the particles. As the absorbance is close to zero

, (ω) is a real number. Solving for (ω) from the expression of σext and plotting

it, we get two branches.

38
 3.0.1. Introduction

Figure 3.1: The effective dielectric constant of the Au-Ag NSs calculated
using the parameters a = 7.5 nm and N = 1015 .

We want to fit (ω) with the following equation :

ωp2
 = ∞ + (8)
ω02 − ω 2
This is a modified form of the Drude-Lorentz Equation that takes interband trans-

itions into account. As these IB transitions occur at a much higher frequency than

the UV-Vis-NIR regime, they are all clubbed together in the ∞ [Link] relax-

ation term γ = 0 as (ω) is a real number. There is also the presence of a ω0

term that is the natural frequency of oscillation of the bound particles. Usually,

this term takes into account the lattice vibrations in non-metals with high-energy

electronic transitions in the background. For most metals ω0 = 0.

39
 3.0.1. Introduction

Figure 3.2: A plot of the effective dielectric constant against energy for the
Au-Ag NSs. The fit data is also plotted. We take the upward branch as that
is what fits with the Drude-Lorentz Model. Code presented in Appendix B.

∞ = −3.85, ωp = 3.81, γ = 0 and ω0 = 7.85

As a comparison, the values of these parameters for a few materials are[1,6,15,21] :

Parameters (in eV) Gold Silver α-SiC MgO

∞ 9 5 6.7 3.01

ωp 9.1 9.2 0.18 0.12,0.017

γ 0.072 0.021 5.9 × 10−4 9.4 × 10−4 , 0.013

ω0 0 0 0.098 0.05,0.08

MgO is modelled a double oscillator which explains the two sets of frequencies in

the table.

Clearly, the values are unphysical, especially the fact that ∞ is negative which

implies imaginary IB frequencies. The value of ω0 is also astronomically high com-

pared to even a non-metal like α-SiC.

40
 3.0.2. Measurements

3.0.2 Measurements

We measure the absorbance of the final solution as the drop-cast film is balckish in

colour and the absorbance would be very high anyway; thereby not providing any

additional information.

We use a simple, purpose-built setup for our measurements. We have a light source

(Ocean Optics HL 2000 or Ocean Optics D 2000) that is connected via an optical

fibre cable to a cuvette holder. Another fibre cable connects the cuvette holder to

a spectrophotometer (Ocean Optics USB 4000 or an Ocean Optics Flame), which

is then connected to the computer. We use the OceanOptics OceanView software

to collect data. The experimental procedure is as follows :

We first do a baseline correction with DI water as it is the solvent in our samples.

We set an appropriate integration time so that the spectrometer does not get sat-

urated. To reduce the noise, we increase the number of scans per measurement.

We then set the baseline with the lamp switched off, thereby eliminating the ef-

fects of ambient light. If we find the data particularly noisy, we do a Fast Fourier

Transform (FFT) smoothing on OriginPro to get a higher-quality data. We save

our data in .txt files

Next, we run a code that takes the data files and fits (sum of least squares fit)

the data to our plasmonic model described in Chapter 1. In most of the fitting

codes, we have considered arrays of size > 100 for each fitting parameter. The

fAu array goes from 0 to 1, the nmed from 1 to 2. The number density range is

chosen rather arbitrarily, using the trial-and-error method, and is usually in the

[1015 , 1018 ] range. The particle size, is usually varied from 10 nm to 20 nm but is

not included in the fitting loop. This is mostly because the particles have a wide

range of size in the solution (also confirmed from TEM images) and it is convenient

to stick with a value and change it before running the code if the situation arises.

Unless otherwise mentioned, in all our codes, the value of d is 15 nm.

The plots of some of our samples are given below:

41
 3.0.2. Measurements

a b

c d
Figure 3.3: (a), (b) , (c) abd (d) are respectively the Abasorbance vs Wavelength
and the least sum of squares fit curves for Samples 3,4,6 and 8. The data was
taken using the Halogen-tungsten light (Ocean Optics Hl 2000) and the Ocean
Optics USB 4000 spectrometer. Both these devices work in the 350-900 nm range.
We have removed the noise using FFT Smoothing on OriginPro. The fit parameters
are given in the text boxes. To be noted is that the particle size is 15 mm in all
these cases unless otherwise mentioned.

Clearly, Samples 3,4 and 6 have a decent agreement with the theoretical data,In all

these samples, the fraction of gold (fAu ) roughly lies between 10 and 15 percent. It

would be seen in the later samples that fAu is in the 10%-20% range. TEM images

of some other samples also confirm that Ag is the major constituent. The ratios

obtained from TEM does not exactly match with the fit parameters but the trend

is well exhibited.

42
 3.0.2. Measurements

In Sample 8, the tail does not go to zero, something that cannot be explained by

this theoretical model. We see this behaviour in several other samples as well.

These are the data for some other samples.

a b

c d
Figure 3.4: (a), (b) , (c) abd (d) are respectively the Absorbance vs Wavelength
for Samples 12,14,15 and 19. The light source was a Halogen-tungsten lamp and
the spectrometer was Ocean Optics Flame.A particle size of 15 mm in all these
cases unless otherwise mentioned.

Note that ‘lsqs’ refers to the least sum of squares.

In other words, lsqs = min(Σi (gi − fi )2 ), where fi ’s are the absorbance values as i

goes over all the data [Link] ’s are the data points in the fitting curve. For every

set of the fitting parameters, we have a unique set of gi ’s. ‘lsqsr’, which is defined
lsqs
as lsqsr = gives the accuracy of fit.
Σi fi2

43
 3.0.2. Measurements

Summary

• We analyse 8 samples here and all but two have their lsqsr values less than

0.1. Those are all good fits.

• The fraction of gold varies between 12 and 26 percent. TEM images of 3

similarly-prepared samples also corroborate this trend, although one needs

to carry out further TEM imaging to be sure,

• The peak is in the 400-500 nm range. There is no correlation between the

amount of gold salt added in the reaction mixture and the peak position. We

attribute this to the fact that the constitution of the nanoparticles may not

correlate with the reagent amounts.

44
 Chapter 4
Appendix A

Derivation of Surface Plasmon Resonance NB : Appendix A has been

taken verbatim from the author’s Bachelors’ thesis [1] .

Consider a metallic sphere of radius a and permittivity (ω) in a dielectric medium

(m ). A constant electric field E0 k̂ is switched on. From the Laplace Equation

(∇2 φ = 0) the potential due to this field φ(r, θ) takes the following forms inside

and outside the sphere[18] :

P l +C −(l+1) ]P
Φout = l [Bl r lr l (cosθ) (A1a)

Al rl Pl (cosθ)
P
Φin = l (A1b)

where Pl are Legendre Polynomials of order l(≥ 0) and Al , Bl and Cl are constants.

To be noted is that the ‘−(l + 1)’ is term is absent in Φin as it leads to a singularity

at r = 0.

One needs to determine Al , Bl and Cl using the boundary conditions.

At large r, E = E0 , implying φ(r → ∞)=-E0 rcosθ. This leads to, from equation

(A1a)
 and the fact that P1 (cosθ) = cosθ,

Bl = 0 for l 6= 1


(A2)

B1 = −E0



45
 4. Appendix A

Now, from the equality of tangential components of the electric field, i.e.
   
δφin δφout
δθ r=a = δθ , one has the following :
r=a
 
P∞ dPl P∞  Cl  dPl
l l
l=0 Al a = l=0 Bl a + l+1  which leads to
 
dθ a dθ



 Cl
 Al = for l 6= 1


a2l+1
C1




 A1 = − E0 (A3)

a3
   
δφin δφout
From the equality of radial components, i.e. (ω) δr r=a = m δr one
r=a
has :  
P∞ P∞  Cl (l + 1)
(ω) l−1 P l−1 −
l=0 Al la = m l=0 Bl la
l
  Pl .
al+2 

Using
 equation (A2), this leads to :
m Cl (l + 1)





 A l = − for l 6= 1
(ω) la2l+1


 

 m  2C1 
A = − E0 + (A4)

1
 
3 




 (ω)  a

Equating the corresponding parts of equations (A3) and (A4), one eliminates Al

and

gets the following :

Cl = 0 for l6= 1





 (ω) − m
C1 = a3 E0 (A5)


(ω) + 2m
Using (A5) one finds the values of Al :


Al = 0 for l 6= 1





 (ω) − m
A1 = −E0 (A6)


(ω) + 2m

Combining Equations (A6) and (A5) with Equation (A1), one has :
(ω) − m E0 cosθ
Φout = −E0 rcosθ + a3 (A7a)
(ω) + 2m r2
3m
Φin = − E0 rcosθ (A7b)
(ω) + 2m

46
 4. Appendix A

Taking the gradient of Φ, one gets the electric field as follows :

3m
E~in = E0 (−cosθr̂+sinθθ̂) (A8a)
(ω) + 2m   
(ω) − m 2E0 (ω) − m E0
E~out = −E0 cosθ  a3  a3 
   
1 +  r̂+E0 sinθ 1 −  θ̂(A8b)

(ω) + 2m r3 (ω) + 2m r3

3m
Clearly, E~in = E0 (Equation (2)) and E~out can be also be written as
(ω) + 2m
(ω) − m a3
E~out = E0 (−cosθr̂ + sinθθ̂) − E0 (2cosθr̂ + sinθθ̂)
(ω) + 2m r3

The first part of the above equation is the external field and the second part is due

to the dipole induced on the metallic particle. The dipole moment is given by
(ω) − m
p = 4πm a3 E0 (A9a)
(ω) + 2m
and the polarizability by
(ω) − m
α(ω) = 4πm a3 (A9b)
(ω) + 2m

When extinction spectroscopic measurements are performed on the sample, one ob-

serves a resonance at that frequency ω where Re((ω)) = −2m (A10)

This is also known as the Frolich’s Condition. The scattering and absorption cross-

sections which together constitute the extinction cross-section are respectively given
k 4 |α|2
by σscat (k) = and σabs (k) = kImag(α) (A11a,b)
6π
where k is the wave vector.

47
 Chapter 5
Appendix B

MATLAB Code for data fitting

nm = 10ˆ(-9);

r = 15*nm/2;

samplef = fopen(’sample [Link]’);

s = textscan(samplef,’%f %f %f’,’Delimiter’,’ ’);

fclose(samplef);

lambdanms = s1; lambdas = lambdanms*nm;

counts = length(lambdas)

abssnew = [];

%silver

agf = fopen(’[Link]’);

p = textscan(agf,’%f %f %f’,’Delimiter’,’ ’);

fclose(agf);

ns = p2; ks = p3;

48
 5. Appendix B

%gold

auf = fopen(’[Link]’);

q = textscan(auf,’%f %f %f’,’Delimiter’,’ ’);

fclose(auf);

ng = q2; kg = q3;

count = length(p1)

lambdanm = p1*1000 ;

lambda = p1/106 ;

eig = 2 ∗ ng. ∗ kg;

erg = ng.2 − kg.2 ;

eg = erg + 1j ∗ eig;

ers = ns.2 − ks.2 ;

eis = 2 ∗ ns. ∗ ks;

es = ers + 1j ∗ eis;

nm = 10( − 9);

%%%%%%%%%%%%%

sp = 1;

if s1(1) < lambdanm(1)

sp = 1;

else

for i = 2:count

if lambdanm(i-1)<s1(1) lambdanm(i)>s1(1)

sp = i;

end

end

end

lambdanmnew = [] ; egnew=[] ; esnew=[];

49
 5. Appendix B

for i = sp:count

lambdanmnew(i-sp+1) = lambdanm(i);

egnew(i-sp+1) = eg(i);esnew(i-sp+1) = es(i);

end

countnew = length(lambdanmnew)

S = 0;

for i = 1:countnew

for j = 2:counts

if lambdanms(j-1) <= lambdanmnew(i) lambdanms(j)>=lambdanmnew(i)

abssnew(i) = (s2(j-1)+s2(j))/2;

S = S+abssnew(i)2 ;

end

end

end

length(abssnew)

plot(lambdanmnew,abssnew);hold on

f = linspace(0,1,20); nmed = linspace(1,1.7,10);

N = linspace(1*10( 16), 20 ∗ 10( 16), 100); h = 0.01;

absor = zeros(length(f),length(nmed),length(N),length(abssnew));

diff = zeros(length(f),length(nmed),length(N),length(abssnew));

sumsqdiff=zeros(length(f),length(nmed),length(N));

for j=1:length(f)

50
 5. Appendix B

for k = 1:length(nmed)

for m = 1:length(N)

for i = 1:length(abssnew)

absor(j,k,m,i) = -log10(exp(-N(m)*h*(2*pi/lambda(i))*imag(4*pi*(nmed(k)2 )∗r3 ∗

(esnew(i)∗((1+2∗f (j))∗egnew(i)+(2−2∗f (j))∗esnew(i))./((1−f (j))∗egnew(i)+

(f (j) + 2) ∗ esnew(i)) −

nmed(k)2 )/(esnew(i) ∗ ((1 + 2 ∗ f (j)) ∗ egnew(i) + (2 − 2 ∗ f (j)) ∗ esnew(i))./((1 −

f (j)) ∗ egnew(i) + (f (j) + 2) ∗ esnew(i)) + 2 ∗ nmed(k)2 )))); %soln

end

end

end

end

for j=1:length(f)

for k = 1:length(nmed)

for m = 1:length(N)

for i = 1:countnew

sumsqdiff(j,k,m) = sumsqdiff(j,k,m) + (absor(j,k,m,i)-abssnew(i))2 ;

end

end

end

end

sumsqdiffmin = sumsqdiff(1,1,1);jmin = 1;kmin = 1;mmin = 1;

for j=1:length(f)

for k = 1:length(nmed)

for m = 1:length(N)

if sumsqdiff(j,k,m) < sumsqdiffmin

51
 5. Appendix B

sumsqdiffmin = sumsqdiff(j,k,m);sumsqdiffminratio = sumsqdiffmin/S;

jmin = j;fmin = f(j);kmin = k;nmedmin = nmed(k);mmin = m;Nmin = N(m);

end

end

end

end

fmin

nmedmin

Nmin

sumsqdiffmin

sumsqdiffminratio

Absorfinal = [];

for j=1:length(f)

for k = 1:length(nmed)

for m = 1:length(N)

for i = 1:countnew

Absorfinal(i) = absor(jmin,kmin,mmin,i);

end

end

end

end

plot(lambdanmnew,Absorfinal);

52
 Chapter 6
Bibliography

1. Panja, A., 2020, Optical and Electrical Characterization of Gold-Silver Nano-

structures, BS thesis, IISc, Bengaluru

2. Mie, G., R. Ann. Phys. 1908, (german)

3. Maier, S.A., Plasmonics-Fundamentals and Applications, Springer (2007)

4. Novotny, L., Hecht, B., "Principles of Nano-Optics", Cambridge University Press

(2006). 5. Jackson, J. D., Classical Electrodynamics, Second Edition, Wiley, New

York, 1975. (Third Edition, 1999)

6. Beer (1852). "Bestimmung der Absorption des rothen Lichts in farbigen Flüssigkeiten"

[Determination of the absorption of red light in colored liquids]. Annalen der Physik

und Chemie (in German). 162 (5): 78–88.

7. Lambert, J.H. (1760). Photometria sive de mensura et gradibus luminis, colorum

et umbrae [Photometry, or, On the measure and gradations of light intensity, col-

ors, and shade] (in Latin). Augsburg, (Germany): Eberhardt Klett.

8, Maxwell Garnett, J. C., "Colours in metal glasses and in metal films," Transac-

tions of the Royal Society, Vol. CCIII, 385-420, London, 1904.

9. Sihvola, A., "Two Main Avenues Leading To the Maxwell Garnett Mixing Rule",

Journal of Electromagnetic Waves and Applications, 15:6, 715-725(2001)

10. Bruggeman, D. A. G. (1935). "Berechnung verschiedener physikalischer Kon-

53
 6. Bibliography

stanten von heterogenen Substanzen. I. Dielektrizitätskonstanten und Leitfähigkeiten

der Mischkörper aus isotropen Substanzen". Annalen der Physik. 416 (7): 636–664.

11. Kreibig, U., Vollmer, M., Optical Properties of Metal Clusters, Springer (1995)

12. Serge, B., "Anisotropic effective medium theories", Journal de Physique I 4(2),

DOI: 10.1051/jp1:1994139

13. Ciesielski, A., Skowronski, L, Trzinski, M., Szoplik, T., Controlling the optical

parameters of self-assembled silver films with wetting layers and annealing, Appl.

Surf. Sci. 421B, 349-356 (2017)

14. Ciesielski, A., Skowronski, L, Trzinski, M., Szoplik, T., Evidence of germanium

segregation in gold thin films, Surf. Sci. 674, 73-78 (2018)

15. P. B. Johnson and R. W. Christy, Optical Constants of the Noble Metals, Phys.

Rev. B 6, 4370 (1972)

16. Thapa, D.K., et. al., Coexistence of Diamagnetism and Vanishingly Small

Electrical Resistance at Ambient Temperature and Pressure in Nanostructures,

arXiv:1807.08572 (2019)

17. Mahapatra P S, Saha S K, Mahadevu R, Islam S, Mondal P, Kumbhakar

S, Sai T P, Patil S, Chandni U and Pandey A 2019 Observation of excess res-

istance anomaly at resistive transitions in Ag/Au nanostructures, arXiv preprint

arXiv:1912.05428

18. Islam S, Mahadevu R, Saha S K, Mahapatra P S, Bhattacharyya B, Thapa D

K, Sai T P, Patil S, Pandey A and Ghosh A 2019 Currentvoltage characteristics in

Ag/Au nanostructures at resistive transitions, arXiv preprint arXiv:1906.02291

19. Hooda M, Kumar P, Balakrishnan V and Yadav C, Observations of zero

electrical resistance of Ag-Au thin films near room temperature, arXiv preprint

arXiv:1906.00708

20. Dalai M K, Singh B B, Sethy S K, Sahoo S P and Bedanta S., Superconduct-

ivity in Ag implanted Au thin film, arXiv preprint arXiv:1906.02091

21. Ashcroft, N.W., Mermin, N.D., Solid State Physics, Harcourt College Publish-

ing, 1976

54
 6. Bibliography

22. Shalaev, V., Cai, W., "Optical Metamaterials- Fundamentals and Applica-

tions", Springer (2010)

55