A monolithic silicon based integrated signal generation and detection system for monitoring DNA hybridisation.

Abstract

The aim of this work was to develop an integrated solution to DNA hybridisation monitoring for diagnostics on a monolithic silicon platform. A fabrication process was developed incorporating a gold initiation electrode patterned directly onto a PIN photodiode detector. Patterned interdigitated type electrodes exhibited the smallest reduction in photodiode sensitivity, therefore these were chosen as the ECL initiator design. A novel DNA hybridisation assay was developed based on the displacement of a partially mismatched complementary strand by a perfectly matched labelled complementary strand. Pre-hybridised thiolated oligonucleotide and unlabelled 25% mismatched oligonucleotide were assembled on the gold initiation electrode. On addition of the labelled perfectly complementary oligonucleotide, the mismatched strands were displaced and a signal was generated. The sensitivity of the photodiode to light emitted at 620 nm, the ruthenium emission wavelength, was determined and subsequently, the diode current response to light generated by flow addition of ruthenium solution was found to be measurable to a concentration of 10 fM. Pre-hybridised duplex DNA, consisting of thiolated oligonucleotide and ruthenium labelled complementary oligonucleotide, was assembled on the gold initiation electrode. The difference between the current measured during flow of buffer and the ECL co-reactant TPA was three orders of magnitude, indicating that DNA assembled on the surface comprised sufficient ruthenium to generate a measurable signal. Finally, the displacement of unlabelled partial mismatch oligonucleotide from the sensor surface was monitored on addition of the ruthenium labelled perfectly complementary oligonucleotide in TPA flow and the measured photodiode current response was up to 50 times greater.

Full text

Page 1

Biosensors and Bioelectronics 21 (2005) 565–573
A monolithic silicon based integrated signal generation and
detection system for monitoring DNA hybridisation
Chiara Bertolinoa, Marion MacSweeneyb, Jane Tobinb, Brendan O’Neillb,
Michelle M. Sheehanb,∗, Salvatore Colucciaa, Helen Berneyc
aDepartment of Chemistry I.F.M., University of Torino, via P. Giuria n.7, 10125 Torino, Italy
bBioanalytical Microsystems, NMRC, University College Cork, Ireland
cNanotechnology, Herschel Building Annex, University of Newcastle upon Tyne, NE1 7RU, UK
Received 27 September 2004; received in revised form 3 December 2004; accepted 9 December 2004
Available online 11 January 2005
Abstract
TheaimofthisworkwastodevelopanintegratedsolutiontoDNAhybridisationmonitoringfordiagnosticsonamonolithicsiliconplatform.
A fabrication process was developed incorporating a gold initiation electrode patterned directly onto a PIN photodiode detector. Patterned
interdigitated type electrodes exhibited the smallest reduction in photodiode sensitivity, therefore these were chosen as the ECL initiator
design.
A novel DNA hybridisation assay was developed based on the displacement of a partially mismatched complementary strand by a perfectly
matched labelled complementary strand. Prehybridised thiolated oligonucleotide and unlabelled 25% mismatched oligonucleotide were
assembled on the gold initiation electrode. On addition of the labelled perfectly complementary oligonucleotide, the mismatched strands were
displaced and a signal was generated.
The sensitivity of the photodiode to light emitted at 620nm, the ruthenium emission wavelength, was determined and subsequently, the
diode current response to light generated by flow addition of ruthenium solution was found to be measurable to a concentration of 10fM.
Prehybridised duplex DNA, consisting of thiolated oligonucleotide and ruthenium labelled complementary oligonucleotide, was assembled
on the gold initiation electrode. The difference between the current measured during flow of buffer and the ECL coreactant TPA was three
orders of magnitude, indicating that DNA assembled on the surface comprised sufficient ruthenium to generate a measurable signal. Finally,
the displacement of unlabelled partial mismatch oligonucleotide from the sensor surface was monitored on addition of the ruthenium labelled
perfectly complementary oligonucleotide in TPA flow and the measured photodiode current response was up to 50 times greater.
© 2004 Elsevier B.V. All rights reserved.
Keywords: Electrochemiluminescence; DNA biosensor; PIN photodiode; Point of care diagnostic
1. Introduction
The development of the technique of radioimmunoassays
in the 1960s revolutionised diagnostics, allowing rapid and
quantitative analysis of biomolecules of clinical significance
(Yalow and Berson, 1959, 1960). In the intervening years,
advances in assay configurations, labelling technology, and
the introduction of novel detection platforms has made point
of care diagnostics an attainable goal. The advent of the Hu-
∗Corresponding author. Tel.: +353 21 4904083; fax: +353 21 4270271.
E-mail address: michelle.sheehan@nmrc.ie (M.M. Sheehan).
man Genome Project and discovery of new genes implicated
in disease prognosis and diagnosis has resulted in a rapid
expansion in the field of molecular diagnostics. The current
challenge is to develop reliable, accurate and cost effective
tests based on analysis of genomic sequences based on either
DNA sequencing or hybridisation.
The hybridisation detection format is currently gaining
favour because of its simplicity. With this approach, speci-
ficity is conferred by the sequence of probe DNA. This is
a sequence of DNA that specifically recognises, based on
DNA sequence homology, so-called marker sequences that
are diagnostic or prognostic for certain diseases or disease
0956-5663/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.bios.2004.12.007

Page 2

566
C. Bertolino et al. / Biosensors and Bioelectronics 21 (2005) 565–573
states. The translation of the duplex formation event into a
measurable output signal can be performed in a number of
ways, directly by capacitance or impedance (Souteyrand et
al.,1997;Berneyetal.,2000;Usluetal.,2004),electrochem-
ically (Millan et al., 1994; Wang et al., 1999), by frequency
response (Su et al., 1994; Caruso et al., 1997; Mannelli et al.,
2003) or using optical labelling (Piunno et al., 1995; Abel
et al., 1996; Hsueh et al., 1998). Depending on the applica-
tion, different types of information may be required, from
the presence or absence of a sequence in mutation analysis
to quantitative measurements in functional genomics (Case-
Green et al., 1998).
Currently, the preferred approach for DNA array hybridi-
sationinterrogationisbasedonfluorescentdetection(Blohm
andGuiseppi-Elie,2001;Epsteinetal.,2002;Mantripragada
et al., 2004). However, ECL generation has an advantage in
that there is no requirement for expensive excitation sources,
as the reaction can be initiated by an applied potential. ECL
involves the generation of an electroactive species at an elec-
trode surface that undergoes electron transfer reactions to
reach excited states that emit light (Richter, 2004). The first
appearance of ECL in the literature can be found as far back
as 1927, where light emission from Gringnard compounds at
appliedpotentialswasreported(Duffordetal.,1927).ECLof
luminol in alkaline solutions was reported in 1928 (Harvey,
1929) and in the early 1960s the detailed characteristics of
ECL of ruthenium complexes began (Hercules and Lytle,
1966). It was at this time that ruthenium emerged as a use-
ful tool for many analytical applications, and proved to be a
highly sensitive and selective detection method (F¨ ahnrich et
al., 2001). Ruthenium is still the most commonly used ECL
compound, due primarily to its ability to produce an elec-
trochemiluminescent reaction in aqueous buffered solutions,
in the presence of dissolved oxygen and other impurities,
at room temperature, at easily attainable potentials and with
very high efficiency (Knight, 1999).
During the last decade, the use of Tris(2,2?-bipyridyl)
ruthenium (Ru(bpy)32+) labels for practical applications in
immunoassays and nucleic acid assays has become a reality.
ECL has proven to be a versatile analytical technique with
high sensitivity and selectivity and is used commercially for
immunoassay applications (Choi and Bard, 2004; Bard and
Whitesides,1993a,b,1994).Inthebiosensorfieldofresearch,
ECL has been used to detect drug concentrations (Michel et
al., 1999a), to quantify DNA (Hsueh et al., 1998), the detec-
tion of proteins (Michel et al., 1999b), enzymes (L’Hostis et
al.,2000)andantibiotics(TomitaandBulh˜ oes,2001).Ruthe-
niumECLhasalsobeenusedforthedetectionofmicroorgan-
isms and specifically for those that can be used for biological
warfare (Yu et al., 2000). Co-reactant ECL is being used in
a range of analytical applications. In this case, photons are
generated at a single potential step in the presence of an oxi-
dised or reduced intermediate species that can react with the
ECL luminophore to produce excited states.
There are two commercially available systems for detec-
tion of ECL, one from IGEN Inc. (Gaithersburg, MD, USA)
and the other from Roche Diagnostics (Basel, Switzerland).
Both are based on Igen’s ORIGEN®Technology. However,
advances in microelectronic fabrication technology and the
trend towards miniaturisation and point of care testing has
meanttheemergenceofintegratedsensorapproachestoECL
generation and detection. One of the first demonstrations of
DNA quantification using a microfabricated system was re-
ported by Hsueh et al. (1998). The assembled microcell was
comprised of three silicon layers, a Pyrex cover and a com-
mercial PIN diode. The base silicon had platinum deposited
on the surface, which acted as the anode, the intermediate
silicon layer was etched to form fluidic channels and had a
platinum cathode. Magnetic beads loaded with ruthenium la-
belled single stranded oliognucleotides were captured on the
anodeandthegeneratedECLwasmeasuredintheTPAcore-
actant. This relatively bulky assembly showed the possibili-
ties of detection of ruthenium label, but hybridisation reac-
tionswerenotperformed.Thesystemhadalimitofdetection
of 10−8M Ru(bpy)32+. The total integration of the photodi-
ode detector element and initiation electrode was performed
by a group at the Insititute of Microtechnology at Neuchˆ atel
(Michel et al., 1999a,b; L’Hostis et al., 2000; Fiaccabrino
et al., 1998). A silicon pn photodiode detector element was
integratedwithaplatinum,goldorcarboninterdigitatedelec-
trode.Initially,thedetectionlimitsusingRu(bpy)32+andTPA
as a model were determined. The device was then used to de-
tect the presence of codeine. A limit of detection of 100?M
was obtained and the measurement was performed in a dif-
ferential mode.
The aim of this research was to fabricate and characterise
an integrated system for initiation and detection of ECL and
its application to a DNA assay. Ruthenium labelled target
DNA provides the ECL luminophore. Hybridisation of this
target to immobilised probe on the initiation electrode was
monitored by the PIN photodiode on addition of TPA core-
actant. When TPA is in excess the emitted light intensity is
directly proportional to the amount of oxidised Ru(bpy)32+
(Hsueh et al., 1998).
2. Materials and methods
AllreagentsweresuppliedbySigma–AldrichIrelandLtd.,
Ireland, unless otherwise stated.
2.1. Microfabrication
2.1.1. Gold chip fabrication
A 1?m thickness of silicon dioxide was thermally grown
on the surface of a silicon wafer. Evaporation of 35nm
chrome, followed by 200nm gold onto the wafers was per-
formed. The wafer was diced into 4mm×4mm chips. The
chips were cleaned prior to performing the DNA immobili-
sation by incubation for 4min in an 80◦C solution of 5:1:1
water,30%hydrogenperoxideand30%ammonia.Thechips
were then rinsed with distilled water and dried.

Page 3

C. Bertolino et al. / Biosensors and Bioelectronics 21 (2005) 565–573
567
2.1.2. Photodiode fabrication and packaging
Positive–Intrinsic–Negative (PIN) photodiodes consist of
a large neutrally doped intrinsic region sandwiched be-
tween p-doped (boron) and n-doped (phosphorus) semicon-
ducting regions. The starting material for the PIN photodi-
odes was n-type, float zone (Fz) silicon [111], resistivity
700–1300?/cm.Depositionofalayerof330˚AthickLPCVD
silicondioxideontothesurfaceofthesiliconwasperformed,
followed by phosphorous implantation (6e11dose at an im-
plantenergyof80keV)toproducealowdopedn-typeregion,
which helps to reduce current leakage in the final device. Sil-
icon dioxide of 3900˚A thickness was deposited, the p+ re-
gion was patterned and the oxide was etched to 500˚A oxide
in patterned areas. Boron was implanted at a dose of 5e15
and an energy of 35keV through the 500˚A oxide. This ox-
ide was removed by wet etch and the boron was driven in
by an activation anneal at 1000◦C. The oxide thickness in
the p+ region was 2100˚A. The oxide on the wafer back was
removed and phosphorus was back implanted (5e15dose at
an implant energy of 35keV) for back contact. Contact etch
and metallisation (10,000˚A aluminium) was performed. Fi-
nally, the gold initiation electrodes were integrated onto the
diodes. The gold electrode fabrication involves a lift-off pro-
cess, which differs from the lithography used to pattern the
photodiodes. The mask set for the gold electrode lithogra-
phy consists of two masks, one for the electrode design and
the other for the bond pads of the electrodes. A photoresist
layer was patterned on the wafers, followed by gold evapo-
ration. The gold for the electrodes was evaporated onto the
wafer to a depth of 200 and 2000˚A. This was followed by a
lift-off process. A simplified cross section of the completed
PIN photodiodes with integrated gold electrodes is shown in
Fig. 1.
The photodiode die was attached to a 24-pin dual in line
(DIL)packageforelectricalcharacterisationusingAblebond
961-2 (Ablestik, Rancho Dom´ ınguez, CA, USA). The pack-
ages were placed in a Heraeus vacuum oven at 125◦C for
2h to set the adhesive. Two wire bond connections were
required to prepare the device for the electrical characteri-
sation. Firstly, the bond pad of the diode was connected to
the package, subsequently, a bond from the surface of the
package to one of the package bond pads was made. The
surface bond on the packaging device provides a ground
state for the electrical characterisation. For characterisa-
Fig. 1. Cross section of the PIN photodiode device.
tion in an aqueous environment, the photodiode and inte-
grated gold initiation electrode were packaged on dipstick
printed circuit boards (PCBs). The dipstick PCBs (ILFA,
Germany) were manufactured from a FR4 substrate with di-
mensions of 45mm×11mm and a thickness of 1.5mm. The
wire bonds were encapsulated with Hysol®FP4450 (Henkel
Loctite Adhesives Ltd., Hertfordshire, UK) IC Encapsulant.
The packaged diodes were then placed in the Heraeus vac-
uum oven for 2h at 125◦C for the glob top encapsulant to
set.
The ECL limits of detection and DNA hybridisation ex-
periments were performed using a Teflon Flow Cell fabri-
cated in-house to the dimensions of the PC board packaging
of the sensor. The different solutions were manually injected
onto the device using a disposable plastic syringe. The volt-
age was applied and the current signal was collected by a
Hewlett Packard Parameter Analyser 4145A and the device
was connected to the analyser through a Personality Board
16058A Test Fixture (Hewlett-Packard Company, Palo Alto,
CA, USA). The data was transferred to a PC by a VeePRO
6.0 Agilent program.
2.2. Self-assembly of DNA on gold chips
2.2.1. Oligonucleotides
A set of complementary oligonucleotides were chosen
for this work; Wild type complementary “WTC” 5?-Thiol-
TGAATTGGCTCAGCTGGCT-3?: Wild type probe “WTP”
5?-AGCCAGCTGAGCCAATTCA-3?:
type probe “WTP Mismatch” 5?-GGCCAACTGAACC-
AGTTCG-3?, the bold text indicates nucleotide base
changes creating the mismatched oligonucleotide. WTC
was modified at the amino terminal end with thiol groups
(Thiol-WTC). WTP was modified at the amino terminal
end with amine groups (Amine-WTP). In addition, WTP
were individually labelled at the amino terminal end by,
digoxigenin (DIG-WTP) or ruthenium (Ru-WTP) (Section
2.4). Oligonucleotides were supplied, modified or labelled
by Proligo (France), unless otherwise stated.
Mismatchedwild
2.2.2. Prehybridised duplex DNA
Prehybridised duplex DNA was prepared by incubat-
ing equal volumes and concentrations of thiol-WTC and
complementary WTP in 2× SSC [0.03M Na3C6H5O7
(sodium citrate), 0.3M NaCl, pH 7.0 buffer] at room
temperature for 30min. A 100?l aliquot of this DNA
solution was applied to gold chips and incubated for
90min at room temperature. Chips were then washed in
2× SSC.
2.2.3. Thiol modified DNA and hybridisation
Thiol-WTC were prepared at different concentrations in
0.5M potassium phosphate buffer, pH 7. A 100?l aliquot
of this DNA solution was applied to gold chips and incu-
batedfor90minatroomtemperature.Thechipswerewashed
with deionised water. A 200?l aliquot of blocking solution

Page 4

568
C. Bertolino et al. / Biosensors and Bioelectronics 21 (2005) 565–573
(5mg/ml BSA, 25?l 10% SDS in 1ml of 0.1M NaCl) was
added to each chip and incubated for 60min at 4◦C, then
washed with deionised water. WTP prepared in 2× SSC was
incubated on the gold chips for 90min at room temperature.
The chips were washed with 2× SSC.
2.3. Characterisation of assembled DNA using a
modified ELISA
A 100?l aliquot of 1:100 dilution of horseradish perox-
idase labelled rabbit anti-digoxigenin antibody (Dako, Gal-
way, Ireland) was incubated on the chips for 60min at room
temperature.Thechipswerewashedwith2×SSCand200?l
of 3,3?,5,5?-tetramethylbenzidine (TMB) was applied to the
chips for 15min. The reaction was stopped by addition of
100?l of 0.1M HCl. The absorbance at 450nm was read
using a Wallac 1420 VICTOR 2 multilabel, multitask plate
reader (Perkin-Elmer, Boston, MA, USA). The test sam-
plescomprisedself-assembledthiol-WTCongoldhybridised
to a complementary strand labelled with digoxigenin, DIG-
WTP. The negative control or blank sample comprised self-
assembled Thiol-WTC on gold hybridised to the unlabelled
WTP.
2.4. Ruthenium labelling of DNA
A quantity of 0.31mg of bis(2,2?-bipyridine)-4?-me-
thyl-4-carboxybipyridine-ruthenium
bis(hexafluorophosphate) NHS ester was dissolved in
17.2?l of dimethylsulfoxide (DMSO) anhydrous. An 84?l
aliquot of Amine-WTP was added to 93?l of sodium
borate buffer, pH 8.5. A 7?l aliquot of the NHS ester
solution was then added. This was incubated overnight
at room temperature, shaking gently. Purification by
HPLC on RP18 column, in triethylammonium acetate
buffer (TEAA) pH 6.5 and acetonitrile, was performed.
The pure fraction was lyophilised and then dissolved in
500?l of 2M NaCl. The absorbance at 450nm (ruthenium
complex) and 260nm (oligonucleotide) was read. From
the absorption results, the concentration of the Ru-WTP
oligonucleotide was calculated. The solution was stored
at −20◦C.
N-succinimidylester-
2.5. Ruthenium solution for photodiode characterisation
A 1.95ml volume of TPA was mixed with 0.05ml of
Tween®20 and 1.36g of potassium phosphate in 85ml
deionised water and stirred for 5min. The pH was adjusted
to the required pH using 0.1M NaOH. 7.48mg of Tris(2,2?-
bipyridyl)dichlororuthenium(II) hexahydrate was added to
the solution and the final volume was brought to 100ml. The
final concentration was 0.1?M of ruthenium complex in a
0.05% TPA solution.
3. Results and discussion
3.1. Optimisation of ruthenium ECL reaction
RutheniumECLwasinvestigatedforuseasanopticalsig-
nallingcomplexforDNAhybridisationdetection.Theruthe-
nium ECL reaction emits light at a wavelength of 620nm.
The amount of luminescence is dependent on several factors
including;thevoltageatwhichtheECLisinitiated,theruthe-
nium concentration in the donor solution, flow rate, and the
pH of the buffer solution.
The optimal conditions for the use of [Ru(bpy)3]2+for
ECLreactionshavetobedetermineddependingonthesystem
considerations. The coreactant species solution used and the
target analyte will often dictate the parameters within which
the ECL reaction can occur. There are different published
parameters for optimum generation of chemiluminescence.
Tomita and Bulh˜ oes (2001) recommend pH 9 but they use a
reactive intermediate in the production of Cefadroxil, as an
electrondonorsolution.Micheletal.(1999a)usepH6.8with
TPA as the coreactant and initiate the reaction at 1.3V.
Forthiswork,weinvestigatedtheinfluenceofpHandini-
tiationvoltageonthechemiluminescenceproducedwithTPA
as the coreactant. The ruthenium concentration was 100?M
and the initiation voltage chosen was 1V. Low amounts of
luminescence is observed at a pH of 7.2 and 7.5. Increasing
the pH of the buffer solution to 8.0 resulted in a significant
increase in the amount of luminescence observed. A further
pHincreasetopH8.5resultedinareductionintheamountof
luminescence produced (Fig. 2). The optimum pH for gener-
ating the maximum amount of photons was pH 8.0 (Fig. 2).
The voltage at which the ruthenium ECL is initiated
also influences the amount of luminescence that is observed
(F¨ ahnrich et al., 2001) and depends on the nature of the elec-
trode material. Several publications have stated that the op-
timum voltage for ruthenium ECL is 1.1V (Michel et al.,
1999b;TomitaandBulh˜ oes,2001).Duringtherutheniumop-
timisation experiments a comparison of luminescence gen-
erated at 0.8 and 1.0V was made in a solution of 0.1?M
[Ru(bpy)3]2+and 0.1M TPA. Increasing the initiation volt-
age from 0.8 to 1V resulted in a significant increase in the
amountofluminescenceobserved.However,thecurrentgen-
erated at 0.8V initiation voltage was easily detectable and,
though the voltage is not the optimum, it is the limit before
Fig. 2. Comparison of different amounts of ECL emitted at different buffer
pH, initiation voltage 1V, 100?M [Ru(bpy)3]2+and 0.1M TPA. The exper-
iments were repeated in triplicate and the standard deviation is shown in the
error bars.

Page 5

C. Bertolino et al. / Biosensors and Bioelectronics 21 (2005) 565–573
569
the DNA-thiol linkage on the gold is damaged (results not
shown).
3.2. DNA immobilisation on gold
To perform a DNA analysis with our ECL device, it
was necessary to firstly study the immobilisation of probe
DNA onto the gold electrode and subsequent hybridisation.
A series of experiments were performed on bare gold chips,
4mm×4mm. Duplex DNA, consisting of Thiol-WTC and
DIG-WTP which had been allow to prehybridise off chip a
priori and at different concentrations were subsequently as-
sembledonthegoldchips.Thisself-assemblywasusedtode-
termine maximum packing efficiency without constraints of
probedensityonhybridisation.Theresultswerecomparedto
in situ hybridisation of DIG-WTP to assembled Thiol-WTC
(Fig.3).ThemeasurementofthepresenceofassembledDNA
was performed using a modified ELISA (Section 2.3).
FortheprehybridisedduplexDNA,themaximumamount
ofDNAmoleculesassembledonthesurfaceoccursathighest
concentrations. However, for the in situ hybridisation at this
concentration, it is likely that the packing of the assembled
single stranded probe prevents optimum hybridisation due
to the steric and charge repulsion effects on the incoming
complementary strand. The maximum hybridisation for the
single stranded probe assembly occurred at a concentration
of 0.1?M probe.
The next stage was to demonstrate the concept of DNA
displacement as a method of generating a sensor biorecogni-
tion layer. Duplex DNA, consisting of prehybridised 0.1?M
Thiol-WTC and 0.1?M WTPMismatch was assembled on
thegoldchips.OnadditionofperfectlycomplementaryDNA,
WTP with or without the DIG label, the WTPMismatch was
displaced and hybridisation occurred. The effect of incuba-
tion time on displacement efficiency was determined using
a modified ELISA. Displacement occurred after 5min and
longer incubation times did not result in an increase in the
amount of displacement or hybridisation (Fig. 4).
In the ideal final sensor configuration, the mismatch se-
quence would be labelled with ruthenium and the reduction
in signal would be monitored on displacement with perfect
Fig.3. Acomparisonofself-assemblyofprehybridiseduplexDNAtoinsitu
hybridisation at different thiolated probe concentrations. The experiments
were repeated in triplicate and the standard deviation is shown in the error
bars.
Fig. 4. Effect of complementary DNA incubation time on displacement of
mismatchedprehybridisedDNA.Theexperimentswererepeatedintriplicate
and the standard deviation is shown in the error bars.
complementary sequence. The advantage of creating a pre-
hybridised biorecognition layer as the assay starting point, is
that the DNA packing density is at a maximum, but the du-
plexformfacilitatessubsequentspaceforhybridisationwhen
the mismatch sequence is displaced. There is also a reduced
contribution of non-specific interactions.
3.3. Characterisation of the PIN photodiodes
3.3.1. Comparison of size and effect of patterned gold
initiator electrode
In order to select the optimum photodiode size and initi-
ation electrode type, the responsivity of the photodiodes to
varying wavelengths in the visible spectrum was monitored.
The responsivity of the photodiodes at each wavelength was
calculated using the following formula:
R = (IPIN/A2
whereRisresponsivity(W/cm2),IPINisthecorrectedphoto-
diodecurrent(A),Adetectortheareaofthephotodiodedetector,
and Cal the calibration value (W), the standard reading for
a particular wavelength given by the reference detector. The
corrected PIN photodiode current was obtained by subtract-
ingthedarkcurrentofthePINphotodiodefromthemeasured
raw photodiode current.
Threedifferentelectrodedesigns,planar,digitatedandin-
terdigitated, were evaluated. The planar electrodes were de-
signedtoprovidethewholesurfaceareaofthephotodiodeas
thebiologicalsupportlayerandalsoalargeinitiationareafor
the ruthenium ECL reaction. The interdigitated and digitated
goldgeometriesweredesignedtoprovidealargesurfacearea
for DNA immobilisation and ECL initiation and they allow
direct exposure of the PIN photodiode surface to the pho-
tons of light produced by the ruthenium. These electrodes
comprised 200nm thick gold fingers, the interdigitated gold
electrode covered 18.18% of the overall area and digitated
gold covered 24.41%.
A comparison between the responsivity of a bare photo-
diode and photodiodes which had been patterned with the
three different gold electrode designs was made (Fig. 5A).
The photodiodes patterned with the interdigitated electrodes
had the highest responsivity results, exhibiting a 21% reduc-
tioninaverageresponsivityincomparisontothecontrolbare
detector)/Cal(1)