Extreme UV single crystal diamond Schottky photodiode in planar and transver con figuration
S.Almaviva a ,Marco Marinelli a ,E.Milani a ,G.Prestopino a ,A.Tucciarone a ,C.Verona a ,⁎,G.Verona-Rinati a ,M.Angelone b ,M.Pillon b
a Dip.di Ing.Meccanica,Universitàdi Roma “Tor Vergata ”,Roma,Italy b
Associazione EURATOM-ENEA sulla Fusione,Frascati,Roma,Italy
a b s t r a c t
a r t i c l e i n f o Article history:
Received 31July 2009
Received in revid form 28October 2009Accepted 5November 2009
Available online 11November 2009Keywords:
Single crystal diamond UV detectors
Schottky photodiode Extreme UV
We report on the study of the performances of two extreme ultraviolet (EUV)photovoltaic single crystal diamond Schottky diodes bad on metal/intrinsic/p-type diamond junction developed at the University of Rome “Tor Vergata ”and having different contact geometries.One detector operates in transver con figuration with a mitransparent metallic contact evaporated on the intrinsic diamond surface,while the cond one operates in planar con figuration with an interdigitated contact structure on the growth surface of the intrinsic diamond layer.Both devices can work in an unbiad mode by using the built-in potential arising from the electrode –diamond interface and show excellent rectifying properties with a recti fication ratio of about 108.
雅思5月考试取消
The devices have been characterized in the EUV spectral region by using He –Ne DC gas discharge radiation source and a toroidal grating vacuum monochromator,with a 5Åwavelength resolution.The extremely good signal-to-noi ratio,the reproducibility of the device respon,the abnce of persistent photoconductivity and undesirable pumping effects suggest the high quality of our CVD diamond for UV applications.The external quantum ef ficiency (EQE)as well as the responsivity have been measured in the spectral range from 20to 120nm and opposite behaviours for the two different geometries propod have been obrved.
©2009Elvier B.V.All rights rerved.
1.Introduction
Diamond appears to be a promising material for UV radiation detection.Its wide band-gap,5.5eV,results in a very low leakage current and its electronic properties as high carrier mobility allow fast time respon [1].Besides,it has a large breakdown electric field (∼10V/μm),a low dielectric constant (i.e.low capacitance),chemical inertness and low intrinsic carrier density,which makes cooling for noi reduction unnecessary [2].Its extreme radiation hardness is well known and another interesting feature,again related to the wide band-gap,is its lective nsitivity to radiation with wavelengths shorter than 225nm (visible-blind detectors)[3].Several attempts have been made to build up UV detectors from natural or synthetic diamonds.A detector often reported in literature is the photoresistor [4,5]having a planar structure and consisting of a photoconductive diamond film with metal electrodes placed on the top surface.It can operate only with external voltage applied and the signal is affected from condary electron emission,which is known to strongly affect the detection properties in the UV and EUV spectral regions.A
英文购物
different geometry reported is a polycrystalline sandwiched photo-diode structure [6,7]with a contact on the diamond growth surface and a backside contact on the silicon substrate.However,the CVD diamond performance is limited in this ca by the polycrystalline structure due to defect states in th
pimpe band-gap introduced by the grain boundaries [8,9],which affect the photoelectric properties and alter the detection characteristics.On the other hand,detector grade natural diamonds are extremely rare and expensive,while high pressure high temperature (HPHT)diamonds have their performance strongly worned by defects and impurities [10].Finally,a more promising approach is to u a planar Schottky photodiode fabricated on homoepitaxially grown p-type diamond epilayer as reported by Koide et al.[11].
A few years ago,at the University of Rome “Tor Vergata ”laboratories,single crystal diamond films grown by chemical vapour deposition (CVD)were ud to obtain a new class of detectors with a layered structure.The performances of a photodetector bad on CVD single crystal diamond in a p-type/intrinsic/metal (PIM)con figuration with a grid-shaped Al contact were reported in a previous publication [12].In order to allow the comparison among the propod PIM structure and the con figuration most reported in interdigitated electrodes,two different con figurations of the PIM device have been tested.
Diamond &Related Materials 19(2010)78–82
⁎Corresponding author.
E-mail address:claudio.verona@uniroma2.it (C.
Verona).
0925-9635/$–e front matter ©2009Elvier B.V.All rights rerved.doi:
10.1016/j.diamond.2009.11.007
Contents lists available at ScienceDirect
Diamond &Related Materials
j o u r na l h o m e p a g e :w w w.e l s ev i e r.c o m /l o c a t e /d i a m o n d
Afirst detector operates in a planar configuration with interdig-itatedfingers on the diamond surface.One t offingers is made of aluminium and the cond one is made of p-type diamond.The cond one has been designed in transver configuration,but at variance with Ref.[12],instead of the grid-shaped contact a mitransparent electrode was deposited on the diamond surface in order to improve the electricfield homogeneity.
2.Experimental
The PIM diamond detector consists of a multilayered structure obtained by a three step deposition process.A conductive boron-doped diamond homoepitaxial layer with approximately5Ωcm resistivity,ud as a backing contact,is deposited by Microwave Plasma Enhanced CVD(MWPECVD)on a commercial low-cost synthetic HPHT type Ib single crystal diamond substrate,4×4×0.5mm3in size.After that an intrinsic diamond layer is homo-epitaxially grown on the doped one and is ud as the detecting layer. Due to the small penetration depth of UV radiation in the10–200nm range[13],the detecting region of diamond has a thickness of approximately2µm.The intrinsic layer is deposited by using a parate reactor in order to avoid any boron contamination.The intrinsic diamond surface is oxidized,after the growth,by isothermal annealing at500°C for1h in air,in order to remove the hydrogenized surface conductive layer.Finally,a mitransparent Al electrode (3
mm in diameter)with a thickness of about10nm is deposited on the diamond surface by thermal evaporation,while annealed silver paint is utilized in order to provide an ohmic contact with the B-doped layer.
For the cond diamond detector with interdigitatedfinger electrodes(IDT-PIM in the following),two steps of a standard photo-lithographic technique are ud for the fabrication process.
First,an intrinsic diamond layer is homoepitaxially grown by MWPECVD on a commercial HPHT single crystal diamond substrate. As previously mentioned,annealing in air is employed in order to remove the surface conductive layer of the as-grown diamondfilm. After the annealing process,p-type diamond interdigitatedfingers are lectively grown on the top of the intrinsic diamond layer by using a patterned Cr plasma-resistant coplanar mask.After removal of the chromium mask by wet etching,the interdigitated Al electrode is fabricated using a cond mask which is aligned to the pattern previously obtained.The Alfingers are patterned by a standard lift-off photo-lithographic technique and by thermal evaporation on the CVD intrinsic diamond surface.The width and the gap between twofingers are both20µm.A SEM image of the IDT-PIM device and a scheme of PIM device are shown in Fig.1.
举办的英文
Both devices have been tested over the extreme UV spectral region from20to120nm,using He–Ne DC gas discharge as radiation source and a toroidal grating vacuum monochromator(Jobin Yvon model LHT-30)with a5Åwavelength resolution.The dimension of the optical aperture was0.25×6mm2;a manual shutter was ud to switch on and off the UV radiation.The photorespon measurements have been performed in a vacuum chamber,at a pressure of 0.03mbar.By using a three dimension mechanical(X–Y–Z)stage powered by stepper motors,it is possible to locate the photodetector in front of the beam light and to compare its respon with that of a calibrated NIST silicon photodiode[14]placed in the same position, which measures the absolute photonflux.A hole,2mm in diameter,is ud to collimate the radiation on the nsitive area of the detectors and to obtain the same illuminated area on the silicon photodiode. The photocurrent is measured by an electrometer(Keithley6517A), using the internal voltage source.The detectors are rever biad with a negative voltage on the boron-doped contact while the Al contact is grounded.It must be noticed,however,that the particular detector structure ud is also able to work in an unbiad mode,by using the internal voltage drop at the electrode–diamond junction.
Becau of the different geometries adopted for the two devices, they are measured differently.The PIM detector is encapsulated in a copper/vetronite shielded housing with a2mm pinhole.In such hou
sing,the Al contact is grounded and the photocurrent is measured from the p-type diamond electrode so that the signal is not affected by the possible prence of condary electron emission current from the illuminated contact.
In the ca of the IDT-PIM,condary electron emission cannot be shielded becau the signal is collected from the contacts on the irradiated surface.Therefore the measured photocurrent of IDT-PIM detector can contain both photoconductive current and photoemis-sion current arising from the Alfingers and from the p-type diamond expod to the UV irradiation.
blastfurnace
3.Results and discussion
Fig.2shows the I–V characteristics of the two diamond Schottky diodes measured from−3V to10V at room temperature in a vacuum chamber at a background pressure of10−4mbar.The dark current is lower than10−13A for rever bias voltages for both devices.A very good rectification ration of about108was obrved for both devices at ±3V.The photocurrent vs.applied voltage is also reported in the samefigure when the devices are expod to UV radiation from a He–Ne lamp and at30.4nm(He line)and73nm(Ne line).Both devices operate in the rever bias mode becau when operating in the forward bias mode,the photocurrent is masked by the dark current [15].It is evident t
hat the devices show a photocurrent respon even at zero voltage bias.The photocurrent is almost constant with increasing positive voltage,while the dark current increas by about two orders of magnitude.Remarkably,thus,the best signal-to-dark current ratio is obtained at zero bias voltage,so that in the following,the devices have been operated with no external bias voltage
applied.
Fig.1.a)SEM image of the IDT-PIM device:(1)p-type diamond,(2)intrinsic diamond and(3)Al contact.b)Structure of the PIM device.
79
S.Almaviva et al./Diamond&Related Materials19(2010)78–82
The detector time respon,upon exposure to UV radiation,has been measured by opening and closing a manual shutter during the acquisition.The temporal respon of the tested devices is reported in Fig.3(a)under UV illumination of the He –Ne DC gas discharge radiation source.The respon is reproducible and undesired effects such as persistent photocurrent and priming or memory effects,which are often obrved in diamond UV detectors [16–18],are not prent.Fig.3(b)shows ri and fall times of the signal of about 60ms,which correspond to the acquisition rate of the ud electronic chain.
pgf
The normalized emission spectra of a DC discharge He –Ne lamp measured in an unbiad mode by the detectors are reported in Fig.4.All spectral lines are clearly resolved and obrved with a goo
d signal-to-noi ratio,demonstrating the high photodetection capabilities of the CVD single crystal diamond grown in the extreme UV spectral region.In particular,the weak intensity lines of the He –Ne spectrum in the wavelength range 20–30nm [19]are easily resolved by PIM detector (e Fig.5).
The absolute spectral respon of the detectors is measured by comparison with a calibrated photodiode expod to the same source on the same optical area of about 1mm 2.
The spectral responsivity,expresd in amperes per watt (A/W),is de fined as the photocurrent per unit incident optical power and can be evaluated from the relationship R d =R Si I d /I Si where R Si is the responsivity of the calibrated silicon photodiode at a given wave-length,I Si and I d are the photocurrents measured by the silicon photodiode and the diamond detector,respectively.
The absolute spectra responsivity curves of the two devices are shown in Fig.6.
The responsivity of the PIM device decreas monotonically as the wavelength increas up to about 80nm,but at 120nm an increment is obrved.At 98nm the signal is below the noi level so that only an upper limit can be provided (e error bar in Fig.6).However,the prence of a minimum in the responsivity at around 100nm can be clearly deduced from Fig.6.
The responsivity of the IDT-PIM detector is much lower than that of the PIM detector at short wavelengths (below 50nm),with a maximum at about 73nm.The incread nsitivity of the IDT-PIM device at intermediate wavelengths could be ascribed to the contribution of photoemission current as already reported in the literature [20].For both devices the absolute responsivity measured at around 50nm is comparable to results recently reported in literature for diamond bad EUV detectors [21].
Fig.6also reports the responsivity of an unshielded detector,tested in a previous publication [12],that operates using a third contact in a transversal con figuration using an Al grid-shaped contact instead of a mitransparent homogeneous Al contact.The new PIM propod in this paper a much higher responsivity than the previously tested device at λb 60nm.The continuous electrical contact generates an electric field near the detector surface more uniform and parallel than that generated by the grid-shaped contact improving the responsivity at low wavelengths.On the other
hand,
Fig.2.I –V characteristic in dark and in light of IDT-PIM (a)and PIM (b)
detectors.
Fig.3.a)Time respons under illumination of He –Ne DC gas discharge radiation source.b)Magni fication of fall time of both devices.
80S.Almaviva et al./Diamond &Related Materials 19(2010)78–82
the shadowthe respon of the grid-shaped contact detector is rather flat and a higher nsitivity with respect to the PIM device is obrved above 60nm.However,it should be pointed out that the grid-shaped device was not encapsulated and the improved nsitivity obrved above 60nm could be due to condary photoemission current contribution.
The External Quantum Ef ficiency (EQE)spectrum,estimated by EQE=1240·R d /λ[nm],is reported in Fig.7for the PIM devices.To understand the different behaviours of EQE for the two devices,a physical analysis of the detection process should be performed.Becau of the particular geometry of the device,this is very dif ficult for the IDT-PIM ca.Moreover,as mentioned above,the photocur-rent measured by IDT-PIM detector includes the photoemission current,which also depends on the wavelength [20].On the contrary,the shielded PIM device is not affected by condary electron contribution,and the more homogeneous electric field con figuration of the PIM device allows a simpl
e analysis of the detection process.The physics of the device is clearly bad on the existence of a Schottky barrier generated at Al –diamond interface,as demonstrated by the fact that the detector is able to operate even with no external bias applied.
The EQE depends on the absorptance A (λ)(number of photon absorbed)of the diamond active layer and taking into account the PIM structure (e Fig.1(b)),we can simulate the experimental data by the following equation:
EQE ðλÞ=g ⋅ηðλÞ⋅A ðλÞ=g ⋅ηðλÞ⋅ð1−R ðλÞÞ⋅e
高中英语改错
−αAl ðλÞd Al
ð1Þ
where g is a photoconductive gain,η(λ)internal quantum ef ficiency,R (λ)the re flectivity of aluminium/diamond structure [22],αAl (λ)the Al absorption coef ficient,d Al (∼10nm)the thickness of Al contact.The number of electron –hole pairs created per absorbed photon η(λ)is unity for λN 95.4nm and it has a value of 95.4/λfor λb 95.4nm [21,23],since for higher photon energies,condary ionization is energetically possible.
The curve expresd by Eq.(1)is reported in Fig.7from 40nm to 120nm for g =1.Clearly it does not reproduce the experimental data.Changing the value of g (e the g =0.1curve in Fig.7),would only produce a vertical shift of the curve therefore not improving much the fit.One reason for this lack of agreement could be that the R(λ)values are accurate only in the ideal condition of a perfectly smooth surface.However,such a strong discrepancy suggests the occurrence of a different process.The penetration depth of diamond shows a deep minimum at about 100nm who trend is qualitatively similar to the responsivity curve obrved in Fig.6.This suggests the existence of a dead layer located at the diamond surface,probably related to the recombination of photo-generated carriers clo to the metal –diamond interface [21].Introducing such a dead diamond layer in the simulation,we can indeed reproduce the experimental behaviour.Under this assumption,Eq.(1)must be multiplied by the additional term exp(−αdiam d diam ),where αdiam (λ)is the diamond absorption coef ficient and d diam is the thickness of the dead diamond
layer.
Fig.4.He –Ne emission spectrum measured by the two
devices.Fig.5.He –Ne spectrum measured by PIM detector in the range 20–30
nm.
reno怎么读Fig.6.Responsivity of both
devices.
Fig.7.External quantum ef ficiency (EQE)of the PIM device between 20and 120nm.The dotted and solid lines correspond to Eq.(1)for g =1and g =0.1respectively.The dashed curve corresponds to Eq.(1)with g =1and adding a 10nm thick dead diamond layer.
81
S.Almaviva et al./Diamond &Related Materials 19(2010)78–82
The efficiency curve reported in Fig.7,calculated by using d diam= 10nm,now shows a good agreement with the experimental data.
4.Conclusions
Two detectors were fabricated at the University of Rome“Tor Vergata”with a structure that acts as a metal/intrinsic/p-doped diamond photovoltaic Schottky diode.The two detectors operate in different configurations:one in transver geometry and the other one in planar configuration.
We have measured the electrical characteristics and tested the performances under continuous vacuum UV photon irradiation of the two devices.A general result of our experiments is that diamond detectors exhibit a low dark current and a very good signal-to-noi ratio.The respons are reprodu
cible and undesired effects such as persistent photocurrent,priming or memory effects are negligible for both devices.The respon time could not be measured,being much lower than the acquisition rate of the ud electronic chain(∼60ms). The results indicate the high quality of our CVD diamond grown for UV applications.
The responsivity and the EQE of the two devices show an opposite behaviour as a function of the radiation wavelengths due to the different operative configurations.In particular the PIM detector is more efficient at lower wavelengths and prents a drop of nsitivity at approximately100nm.The IDT-PIM is less efficient at low wavelengths and has a maximum efficiency at about73nm. Moreover,the measured EQE of the PIM detector was compared with the theoretical curve taking into account the detector structure. The simulation results em to indicate that the prence of a thin dead diamond layer at the diamond–metal interface should be taken into account.Work is in progress to study the influence of the metallic contact material upon the detection performance of the devices.References
[1]J.E.Field,Properties of Diamond,Academic Press,London,1979.
[2]J.Prins,in:A.Paoletti,A.Tucciarone(Eds.),Applications of Diamond Films in
Electronics in“The Physics of Diamond”,IOS Press,Amsterdam,1997.
[3]J.-F.Hochedez,J.Alvarez,F.D.Auret,P.Bergonzo,M.-C.Castex,A.Deneuville,J.M.
Defi,B.Fleck,P.Gibart,S.A.Goodman,O.Hainaut,J.-P.Kleider,P.Lemaire,J.
Manca,E.Monroy,E.Munoz,P.Muret,M.Nesladek,F.Omnes,E.Pace,J.L.Pau,V.
Ralchenko,J.Roggen,U.Schuhle,C.Van Hoof,Diamond Relat.Mater.11(2002) 427.
[4]A.Balducci,M.Marinelli,E.Milani,M.E.Morgada,A.Tucciarone,G.Verona-Rinati,
M.Angelone,M.Pillon,Appl.Phys.Lett.86(2005)193509.
[5]T.Teraji,S.Yoshizaki,H.Wada,M.Hamada,T.Ito,Diamond Relat.Mater.13(2004)
858.
[6]V.I.Polyakov,A.I.Rukovishnikov,N.M.Rossukanyi,A.I.Krikunov,V.G.Ralchenko,
A.A.Smolin,V.I.Konov,V.P.Varnin,I.G.Teremetskaya,Diamond Relat.Mater.7
(1998)821.
[7]L.Thaiyotin, E.Ratanaudompisut,T.Phetchakul,S.Cheirsirikul,S.Supadech,
Diamond Relat Mater11(2002)442.
[8]R.D.McKeag,R.B.Jackman,Diamond Relat.Mater.7(1998)513.
[9]L.Barberini,S.Cadeddu,M.Caria,Nucl.Instrum.Methods460(2001)127.
不在场[10]E.Pace,A.Vinattieri,A.Pini,F.Bogani,M.Santoro,G.Messina,S.Santangelo,Y.
Sato,Phys.Status Solidi,A Appl.Res.181(2000)91.
[11]Y.Koide,Liao Meiyong,J.Alvarez,Diamond Relat.Mater.Des.15(2006)1962.
[12]S.Almaviva,Marco Marinelli,E.Milani,G.Prestopino,A.Tucciarone,C.Verona,G.
Verona-Rinati,M.Angelone,M.Pillon,Diamond Relat.Mater.Des.18(2009)101.
[13]D.Palik,Handbook of Optical Constants of Solids II,Academic Press,New York,
1991.
[14]website:
[15]M.Imura,M.Liao,J.Alvarez,Y.Koide,Diamond Relat.Mater.18(2009)296.
[16]C.E.Nebel,A.Waltenspiel,M.Stutzmann,M.Paul,L.Schäfer,Diamond Relat.
Mater.9(2000)404.
[17]A.De Sio,E.Pace,Nucl.Instr.Methods A552(2005)203.
[18]M.Liao,Y.Koide,J.Alvarez,M.Imura,J.P.,Phys.Rev.B78(2008)045112.
[19]website:v/PhysRefData/ASD/lines_form.html.
[20]T.Saito,K.Hayashi,H.Ishihara,I.Saito,Metrologia43(2006)S51.
[21]A.BenMoussa,A.Theisn,F.Scholze,J.F.Hochedez,U.Schuhle,W.Schmutz,K.
Haenen,Y.Stockman,A.Soltani,D.McMullin,R.E.Vest,U.Kroth,C.Laubis,M.
Richter,V.Mortet,S.Gissot,V.Delouille,M.Dominique,S.Koller,J.P.Halain,Z.
Remes,R.Petern,M.D'Olieslaeger,J.M.Defi,Nucl.Instr.Methods A568(2006) 398.
[22]D.M.Kolb,J.D.E.McIntyre,Surf.Sci.28(1971)321.
[23]C.Canali,E.Gatti,S.F.Kozlov,P.F.Manfredi,C.Manfredotti,F.Nava,A.Quirini,
Nucl.Instr.Methods A160(1979)73.
82S.Almaviva et al./Diamond&Related Materials19(2010)78–82
