Influence of microstructure and composition on hydrogenated siliconthin film properties for uncooled microbolometer applicationsD. B. Saint John, H.-B. Shin, M.-Y. Lee, S. K. Ajmera, A. J. Syllaios et al. Citation: J. Appl. Phys. 110, 033714 (2011); doi: 10.1063/1.3610422 View online: http://dx.doi.org/10.1063/1.3610422 View Table of Contents: http://jap.aip.org/resource/1/JAPIAU/v110/i3 Published by the American Institute of Physics. Related ArticlesTemperature dependent photoluminescence of lateral polarity junctions of metal organic chemical vapordeposition grown GaN J. Appl. Phys. 110, 093503 (2011) Investigation of Ge1-xSnx/Ge with high Sn composition grown at low-temperature AIP Advances 1, 042118 (2011) Mechanisms of suppressing secondary nucleation for low-power and low-temperature microwave plasma self-bias-enhanced growth of diamond films in argon diluted methane AIP Advances 1, 042117 (2011) Surface transition induced island formation on thin strained InGaN layers on GaN (0001) in metal-organic vapourphase epitaxy J. Appl. Phys. 110, 073527 (2011) Direct imaging of InSb (110)-(1×1) surface grown by molecular beam epitaxy J. Appl. Phys. 110, 073526 (2011) Additional information on J. Appl. Phys.Journal Homepage: http://jap.aip.org/ Journal Information: http://jap.aip.org/about/about_the_journal Top downloads: http://jap.aip.org/features/most_downloaded Information for Authors: http://jap.aip.org/authors

Downloaded 02 Nov 2011 to 131.183.220.113. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

Influence of microstructure and composition on hydrogenated silicon thinfilm properties for uncooled microbolometer applications

D. B. Saint John,1 H.-B. Shin,2 M.-Y. Lee,2 S. K. Ajmera,3 A. J. Syllaios,3 E. C. Dickey,1

T. N. Jackson,2 and N. J. Podraza2,4,a)

1Department of Materials Science and Engineering, The Pennsylvania State University, University Park,Pennsylvania 16802, USA2Center for Thin Film Devices and Materials Research Institute, Department of Electrical Engineering,Pennsylvania State University, University Park, Pennsylvania 16802, USA3L-3 Communications Electro-Optical Systems, 3414 Herrmann Dr., Garland, Texas 75041, USA4Department of Physics and Astronomy, University of Toledo, Toledo, Ohio 43606 USA

(Received 15 September 2010; accepted 12 June 2011; published online 5 August 2011)

Doped n- and p-type hydrogenated silicon (Si:H) thin films prepared by plasma enhanced chemical

vapor deposition have been investigated for uncooled microbolometer applications. The material

microstructure has been studied by in situ real time spectroscopic ellipsometry collected during thin

film deposition or ex situ spectroscopic ellipsometry measurements on a static sample with a

multiple sample analysis technique. The key electrical properties of interest, including film

resistivity (q), temperature coefficient of resistance (TCR), and 1/f noise, have been measured as a

function of deposition conditions for p-type amorphous hydrogenated silicon (a-Si:H) films and

microcrystalline content for n-type amorphous (a), microcrystalline (lc), and mixed-phase

amorphousþmicrocrystalline (aþlc) Si:H films. The TCR and 1/f noise values were compared for

p- and n-type a-Si:H samples in the resistivity range of 100 < q < 3000 X cm and show that for a

given resistivity, amorphous p-type films exhibit a lower 1/f noise, which might be expected due to a

larger density of majority carriers. VC 2011 American Institute of Physics. [doi:10.1063/1.3610422]

INTRODUCTION

For several years, hydrogenated silicon (Si:H) has been

used commercially as the active sensing layer in resistive

uncooled microbolometer imaging arrays in the infrared

spectral range.1–4 Although Si:H has been a subject of wide

interest for thin film transistors and photovoltaic devices,5–8

the impact of the growth evolution and microstructure on

microbolometer device performance is not completely

understood. For uncooled microbolometer applications, the

key material properties that enable effective device perform-

ance include electrical resistivity, which results in imaging

layer resistances compatible with read-out circuitry, a large

temperature coefficient of resistance (TCR), and low electri-

cal noise. All of these must be achievable in a material hav-

ing process compatibility with standard IC/MEMS

fabrication.2,9 While these requirements have been achieved

with hydrogenated amorphous silicon (a-Si:H), further opti-

mization of the material system to enable future generations

of high performance uncooled microbolometers requires a

deeper understanding of the relationship between the resis-

tivity, TCR, and 1/f noise as a function of doping and micro-

structure. Prepared by plasma enhanced chemical vapor

deposition (PECVD), a-Si:H is a particularly flexible mate-

rial system since deposition parameters such as substrate

temperature, dopant carrier gas type, doping gas ratios,

plasma power density, and hydrogen dilution of the silane

precursor can be broadly varied to impact the microstructure,

and hence, the electrical properties of the film. In addition,

understanding the specific impact of the film microstructure

on the transport properties will allow for tailoring an

improved class of films since electronic band states can be

designed through the optimization of deposition conditions,

in addition to the potential of alloying with other elements

such as carbon and germanium, which have been used suc-

cessfully in Si:H based photovoltaic devices.

Si:H films prepared by PECVD have been shown to ex-

hibit changes in microstructure as functions of processing

conditions.10–13 For example, films prepared at low hydrogen

dilution generally remain amorphous throughout their

growth, while films prepared at higher dilutions may initially

grow as amorphous until a critical thickness, at which micro-

crystallites first nucleate and subsequently increase in their

coverage of the surface area until they coalesce into a single-

phase microcrystalline layer (lc-Si:H). The critical thickness

for microcrystallite nucleation is dependent upon a variety of

growth conditions and significantly varies upon the specific

deposition parameters chosen.

In this work, Si:H thin films have been deposited by

PECVD, resulting in q values ranging from 0.1 to 3000 X cm,

and corresponding TCRs ranging from –0.5 to –4.5%/K. A

combination of spectroscopic ellipsometry (SE) and electrical

measurements (q, TCR, 1/f noise) have been used to study the

effects of the deposition conditions on the film microstructure

and the resulting impact on the electrical properties. A series

of p-type a-Si:H prepared at differing hydrogen dilution levels

have been used to demonstrate that changes in the growth con-

ditions impact the electrical properties of amorphous material.

Amorphous (a), microcrystalline (lc), and mixed-phasea)Author to whom correspondence should be addressed. Electronic mail:

Nikolas.Podraza@utoledo.edu.

0021-8979/2011/110(3)/033714/7/$30.00 VC 2011 American Institute of Physics110, 033714-1

JOURNAL OF APPLIED PHYSICS 110, 033714 (2011)

Downloaded 02 Nov 2011 to 131.183.220.113. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

(aþ lc)-Si:H n-type films have been evaluated in order to

quantify differences in the electrical properties resulting from

microstructural variations, namely, the microcrystallite frac-

tion. Spectroscopic ellipsometry has been applied in situ in

real time during film growth or ex situ post-deposition in order

to characterize the microstructure and optical properties, in

the form of the complex dielectric function spectra

(e¼ e1þ ie2). Differences between e of the amorphous and

microcrystalline material allow for the extraction of the

microcrystallite fraction in mixed-phase films, giving us

insight into the microstructural composition. Correlations con-

necting the microstructure with the electrical transport proper-

ties are established, as are the roles of the process parameters

in controlling the key material and electrical properties.

EXPERIMENTAL DETAILS

The p- and n-type Si:H films studied here were prepared

using RF (13.56 MHz) PECVD. The p-type a-Si:H films

were deposited onto crystalline silicon (c-Si) wafers coated

with �500 A of low pressure chemical vapor deposition

(LPCVD) silicon nitride (SiNx), while the n-type Si:H films

were deposited onto c-Si wafers with 1 lm of thermally

grown oxide (SiO2) coated with �330 A PECVD SiNx. The

n-type Si:H films were capped with an additional PECVD

SiNx layer prior to atmospheric exposure. The deposition pa-

rameters included a low substrate temperature compatible

with IC fabrication, Ts¼ 200 �C for n-type Si:H and 350 �Cfor p-type a-Si:H; RF plasma power density of 0.53 W/cm2

for n-type Si:H and 0.07 W/cm2 for p-type a-Si:H, and low

total pressures �1.0 Torr, using ultra high purity (> 99.99%)

source gases. The p-type dopant gas is trimethylboron

[B(CH3)3] with a doping ratio D¼ [B(CH3)3]/[SiH4]¼ 0.10,

and the n-type dopant gas is phosphine [PH3] with a doping

ratio D¼ [PH3]/[SiH4]¼ 0.01. The primary variable was the

ratio of hydrogen to silane source gases, R¼ [H2]/[SiH4],

which ranged from 0 to 500. The p-type a-Si:H deposition

process was investigated in situ in real time using a rotating-

compensator multichannel ellipsometer at an angle of inci-

dence, Hi¼ 70o, and a spectral range from 1.5 to 5.0 eV,14

while the n-type Si:H films were measured ex situ using a

similar instrument at two angles of incidence, Hi¼ 65o and

75o, and a spectral range from 1.24 to 5.00 eV.15

The p-type a-Si:H films, prepared at H2-dilutions rang-

ing from 50 � R � 500, remain in the amorphous growth re-

gime throughout all thicknesses studied: �500–800 A. The

structural model for these films consists of a semi-infinite c-

Si substrate/�500 A LPCVD SiNx/bulk a-Si:H film/surface

roughness/ambient, where the free structural parameters con-

sist of the bulk film thickness and surface roughness thick-

ness. For such films, the time evolution of the bulk and

surface roughness layer thicknesses, as well as the bulk layer

dielectric functions, e, were extracted from the in situ real

time spectroscopic ellipsometry data using a global Rr mini-

mization procedure, where r is the unweighted error

function.16

The n-type Si:H films have been deposited in order to

produce single-phase amorphous and microcrystalline mate-

rial, along with a range of mixed-phase amorphous

þmicrocrystalline samples with varying microcrystallite

content. For this series, e and the microstructural parameters

are fit using a least squares regression analysis and an

unweighted error function, r, using ellipsometric spectra col-

lected on static samples ex situ after deposition. These films

are modeled using a semi-infinite c-Si substrate/1 lm SiO2/

�330 A SiNx/Si:H film/Si:HþSiNx interface/SiNx capping

layer/surface roughness/air ambient stack. The free parame-

ters consist of the bottom SiNx thickness, the bulk layer Si:H

thickness, the Si:Hþ SiNx interface thickness, the thickness

of the SiNx capping layer, the surface roughness thickness,

and a parameterization of the Si:H dielectric function.

The dielectric functions of the bottom SiNx layer were

obtained directly from the ellipsometric spectra collected on

a wafer without the Si:H layer deposited where the free pa-

rameters consist of the SiNx bulk and surface roughness

thicknesses, as well as the parameterization of e for SiNx

using a Tauc-Lorentz17 oscillator and a Sellmeier oscilla-

tor.18 The e for the top SiNx layer are extracted using the

same type of parameterization, but from a combination of

ellipsometric measurements collected for two samples where

an a-Si:H layer has been deposited: one without the SiNx

capping layer to quantify the underlying structure and the

other with the SiNx capping layer so that the SiNx optical

properties can be clearly determined. The Si:H parameteriza-

tion for e consists of either a single Tauc-Lorentz oscillator

for a-Si:H or two Tauc-Lorentz oscillators sharing a common

bandgap for microcrystalline lc-Si:H. For the mixed-phase,

(aþlc)-Si:H, either one or two oscillators may be required

to adequately fit the spectra depending on the microcrystal-

lite fraction. For all parameterizations, a constant additive

term to e1, represented by e1, is also included. The optical

properties of the surface roughness layer are represented by

a Bruggeman effective medium approximation19 consisting

of a 0.50 top layer film/0.50 void material fraction mixture.

Similarly, the Si:Hþ SiNx interface is represented by a

Bruggeman effective medium approximation consisting of a

0.50 Si:H/0.50 SiNx mixture. All models use reference

dielectric function spectra for the c-Si substrate and ther-

mally grown SiO2.20

The relative microcrystallite fraction, flc, of the n-type

(aþlc)-Si:H material in the top �250 A of the Si:H layer

was then extracted by fitting the high energy ellipsometric

spectra from 3.5 to 5.0 eV using a multiple sample analysis

procedure. The n-type Si:H samples exhibiting clearly

defined amorphous dielectric function characteristics have

been simultaneously analyzed using a common a-Si:H

dielectric function parameterization, while those n-type Si:H

samples exhibiting the strongest microcrystalline characteris-

tics have been simultaneously analyzed using a common lc-

Si:H dielectric function parameterization. After the common

a-Si:H and lc-Si:H dielectric functions were determined, the

high energy (3.5–5.0 eV) ellipsometric spectra for all sam-

ples were fit to a structural model where the Si:H parameter-

ized dielectric function is replaced by a Bruggeman effective

medium approximation consisting of a-Si:H and lc-Si:H vol-

ume fractions. From this analysis, the relative microcrystal-

lite fraction of each n-type (aþlc)-Si:H film is obtained,

although this only describes the content of the top �250 A of

033714-2 Saint John et al. J. Appl. Phys. 110, 033714 (2011)

Downloaded 02 Nov 2011 to 131.183.220.113. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

each film, since absorption in the 3.5–5.0 eV range limits the

penetration depth of the probing light.

Rectangular test structures of varying size and aspect ra-

tio were fabricated by depositing and patterning low-resist-

ance metal contacts using a lift-off technique and defining

the test structure geometry through photolithography and

dry-etching. Conductivity and 1/f noise measurements have

been performed with the sample in thermal equilibrium in a

temperature-controlled dark chamber. Thermal hysteresis

behavior was not observed during the temperature cycling.

Current–voltage (I-V) characteristics are taken in the –5

toþ 5 V range to confirm linearity and to confirm the ohmic

character of the contacts. Conductivity is measured at a con-

stant current of 20 nA and no drift was observed over the

time of the measurement. The TCR is obtained from the tem-

perature-dependent I-V measurements over a range of 15 to

80 �C and the numerically derived, normalized temperature

derivative of conductivity, 1/r (dr/dT), to obtain an activa-

tion energy, which is then converted into the reported TCR.1

Noise is measured over a frequency range of 1 Hz to 1 KHz

in the dark, in a temperature-controlled chamber, using a low

noise current source and current amplifier (Keithley 428).

The noise power spectral density is extracted with a

HP3562A dynamic signal analyzer. The noise power spectral

density for the samples reported here had 1/f dependence at

low frequency with a transition to a constant (white noise)

characteristic at higher frequency. The white noise was

found to be equal to the calculated Johnson noise for each

test structure resistance implying both low contact resistance

and low test set noise. The noise values reported here are the

1/f noise power spectral densities at 1 Hz (well within the 1/f

noise regime for all samples), taken from spectra like those

shown in Fig. 1. Figure 1 shows two sets of noise spectra;

one from relatively high-noise a-Si:H and one from rela-

tively low-noise a-Si:H.

RESULTS

The results of the in situ real time SE to study the

growth evolution of the a-Si:H films are presented in Fig. 2

which shows the surface roughness thickness as a function of

the bulk layer thickness for the R¼ 500 film. The R¼ 500

film and all films prepared at lower R exhibit an amorphous

roughening transition [a! a],10,21–25 which is denoted by an

increase in surface roughness after the initial coalescence of

clusters on the substrate and subsequent smoothening, typi-

cally at bulk layer thicknesses � 100 A for films prepared

under these conditions. The films remain amorphous both

before and after the [a! a] roughening transition. No amor-

phous to mixed-phase (amorphousþmicrocrystalline) transi-

tion [a ! (aþ lc)],5,10,13 characteristic of the initial

appearance of microcrystallites, is observed even for the

highest H2-dilution within the pressure regime used (�1

Torr). An example of a film undergoing the [a ! (aþlc)]

transition at �250 A with a sharp increase in the surface

roughness thickness beyond that bulk layer thickness from

Ref. 10 is also shown in Fig. 2 for comparison. All films in

this p-type series are determined to nucleate as amorphous

material on the �500 A SiNx-coated c-Si substrates

and remain amorphous throughout the deposition since no

[a ! (aþ lc)] transition is observed. As shown in Fig. 3, efor the R¼ 50 and R¼ 500 films, also indicates that these

films remain amorphous since only a single resonance fea-

ture is observed.26,27 The spectra in e, however, do vary with

increasing H2-dilution. As is evident in Fig. 3, the amplitude

of both the real and imaginary parts increase with H2-dilu-

tion. To quantify this change, e from this series are fit to a

Bruggeman effective medium approximation consisting of eobtained for the representative densest p-type a-Si:H film

exhibiting the highest amplitude; in this case, the film pre-

pared at R¼ 300, and a relative fraction of voids, fv, where

evoid¼ 1.28 The relative void fraction as a function of H2-

dilution is shown in Fig. 4 and indicates that materials pre-

pared at low dilution, R< 200, may exhibit a lower density

while those prepared at higher dilution, R� 200, appear

denser and do not show a significant change in relative void

FIG. 1. The normalized power spectral density in the low frequency region

is shown for two a-Si:H samples, exemplifying relatively high and low

noise, respectively. A dashed line illustrates the theoretical 1/f line shape of

a sample exhibiting intermediate noise.

FIG. 2. Surface roughness layer thickness plotted vs bulk layer thickness

obtained from in situ real time spectroscopic ellipsometry (SE) analysis for

the plasma enhanced chemical vapor deposition (PECVD) p-type,

D¼ [B(CH3)3]/[SiH4]¼ 0.10, hydrogenated amorphous silicon (a-Si:H) de-

posited at H2-dilution, R¼ [H2]/[SiH4]¼ 500, plotted for comparison with

an undoped Si:H film which undergoes a transition from amorphous to

microcrystalline growth (see Ref. 10).

033714-3 Saint John et al. J. Appl. Phys. 110, 033714 (2011)

Downloaded 02 Nov 2011 to 131.183.220.113. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

fraction up to R¼ 500. The resistivity, TCR, and 1/f noise at

1 Hz for p-type a-Si:H films prepared at H2-dilutions from

R¼ 50 to 200 are shown in Fig. 5. For this series of a-Si:H

films, the resistivity, TCR, and 1/f noise increase with H2-

dilution.

The series of n-type Si:H films were studied using exsitu SE and a multiple sample analysis technique to ascertain

the microstructure of the samples and track the influence of

these effects on the electrical properties. Common a-Si:H

and lc-Si:H spectra in e, shown in Fig. 6, are determined for

this series of films from the simultaneous fitting of several

sets of near phase-pure a-Si:H and lc-Si:H samples. The

a-Si:H and lc-Si:H spectra in e extracted by this approach,

respectively, show a characteristic single resonance feature

for a-Si:H26,27 and the presence of two resonance features

near �3.5 and 4.1 eV, representative of the dampened criti-

cal point features of c-Si for lc-Si:H.29,30 These reference

dielectric function spectra and a Bruggeman effective me-

dium approximation are used to determine the relative

microcrystallite fraction in the top �250 A of each n-type

Si:H layer. The resistivity, TCR, and 1/f noise at 1 Hz vary

FIG. 3. Complex dielectric function spectra (e¼ e1þ ie2) obtained from insitu real time SE analysis for the PECVD p-type, D¼ [B(CH3)3]/

[SiH4]¼ 0.10, a-Si:H deposited at the extremes of H2-dilution, R¼ [H2]/

[SiH4]¼ 50 and 500.

FIG. 4. Relative void fraction, fvoid, for the D¼ [B(CH3)3]/[SiH4]¼ 0.10 p-

type a-Si:H films prepared as a function of H2-dilution, R¼ [H2]/[SiH4].

FIG. 5. (a) 1/f noise at 1 Hz; (b) film resistivity, q; and (c) temperature coef-

ficient of resistance, a, for p-type, D¼ [B(CH3)3]/[SiH4]¼ 0.10, a-Si:H de-

posited on c-Si wafers coated with �500 A of LPCVD SiNx as functions of

H2-dilution, R¼ [H2]/[SiH4], ranging from 50�R� 200.

FIG. 6. Complex dielectric function spectra (e¼ e1þ ie2) for the PECVD n-

type, D¼ [PH3]/[SiH4]¼ 0.01, a-Si:H, and microcrystalline hydrogenated

silicon (lc-Si:H) determined from multiple sample ex situ SE analysis.

Reference complex dielectric function spectra for c-Si are also shown (see

Ref. 20).

033714-4 Saint John et al. J. Appl. Phys. 110, 033714 (2011)

Downloaded 02 Nov 2011 to 131.183.220.113. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

with the relative microcrystallite fraction as shown in Fig. 7.

The relative microcrystallite fraction extracted from this

technique has a strong correlation with electrical perform-

ance, with monotonic decreases in TCR, resistivity, and 1/f

noise with increasing microcrystallite fraction.

DISCUSSION

Figures 1 and 2 demonstrate that the p-type a-Si:H films

produced as a function of variable H2-dilution are amor-

phous, regardless of the dilution level up to R¼ 500, for

these other fixed deposition conditions. Disordered amor-

phous material is not unexpected with the inclusion of boron

and carbon from the trimethylboron, though microcrystalline

material may be obtained using higher processing pressures,

plasma power densities, or lower doping gas ratios than those

used in this study.31,32 However, e indicates that films pre-

pared at a low H2-dilution exhibit a lower density than those

prepared at a higher dilution, as shown in Fig. 4. Variations

in e and the microstructure of a-Si:H materials as a function

of H2-dilution are commonly observed. In the case of Si:H-

based thin film photovoltaic devices, the amorphous material

produced at the highest H2-dilution possible, prior to the

nucleation of microcrystallites, is generally considered to ex-

hibit improved ordering. This behavior is reflected in e and

the surface roughness evolution, as evidenced through

improvements in device performance, compared to less

highly diluted materials.10–12 Resistivity, TCR, and 1/f noise

are observed to increase with increasing dilution as shown in

Fig. 5. In general, high TCR is beneficial and higher 1/f noise

is detrimental to microbolometer device performance. We

believe the increase in resistivity may be attributed to one of

several probable mechanisms controlling the final incorpora-

tion of boron into the films, though no direct measurement of

dopant incorporation was performed in this work. Under

deposition conditions involving high H2-dilution, the effec-

tive breakdown of B(CH3)3 in the plasma might be pre-

vented, limiting the dissociated boron available for

incorporation into the film. Alternatively, excessive hydro-

gen at the growth surface may impact the effective sticking

coefficient of boron hydride radicals by reducing the number

of available dangling bond sites for chemisorption of the

dopant precursor. Borane radicals have been linked to

increased deposition rates, by extracting terminal hydrogen

from the surface and creating more dangling bond sites for

SiH3 radicals.33 This sort of behavior, whereby borane tends

to remove hydrogen and create a dangling bond rather than

sticking to the surface itself, is expected to limit the final

borane incorporation in the film. As such, to simultaneously

control resistivity and obtain material prepared at high H2-

dilution, it is necessary to better control the dissociation of

B(CH3)3 in the plasma and its adsorption to the surface to

increase boron incorporation. This control might be achieved

by several methods, including simultaneously varying the

doping gas ratio, D, and the H2-dilution, R, or increasing the

processing pressure or plasma power density for a given dop-

ing ratio, such that the majority of B(CH3)3 is broken down

and effectively incorporated regardless of the H2-dilution

levels. Thus, the increase in resistivity of the p-type films

with H2-dilution may be linked to the relative amount of

dopant incorporated into the films, with the TCR and 1/f

noise also affected. Confirmation of changes in the dopant

concentration was beyond the scope of this study, but merits

further attention.

Even without knowledge of boron concentration in these

films, some conclusions might be inferred through optical

measurement methods. Variations in e show that material

prepared with low H2-dilution is less dense than that pre-

pared at higher H2-dilution. Consequently, for the less dense

material to be more conductive, one can reason that addi-

tional dopants must be incorporated at low H2-dilution with

regard to those dopant atoms which participate in conduc-

tion. As the n-type Si:H data illustrates, there is some corre-

lation between TCR and resistivity, whereby greater

resistivity leads to greater TCR. Because of this effect, it is

difficult to directly ascribe the behavior of TCR to variations

in H2-dilution, rather than the variations in density and resis-

tivity that result from dilution. From a practical standpoint

however, the correlation between H2-dilution and the result-

ing resistivity and TCR is useful to those trying to tailor the

properties of similar films.

The series of n-type Si:H material have been studied by

ex situ SE, and a multiple sample analysis procedure has

been used to determine common amorphous and microcrys-

talline spectra in e, as shown in Fig. 6. The reference e for a-

Si:H consisted of a single Tauc-Lorentz oscillator, which is

characteristic of a-Si:H materials, while the reference e for

lc-Si:H consisted of two Tauc-Lorentz oscillators sharing a

common gap. The lc-Si:H e spectra exhibits broadened and

dampened amplitude features with resonance energies near

�3.5 and 4.1 eV, which are consistent with the critical point

positions of c-Si and the dampening and broadening of these

features due to the smaller grain size and increased disorder

FIG. 7. (a) 1/f noise at 1 Hz; (b) film resistivity, q; and (c) temperature coef-

ficient of resistance, a, for n-type, D¼ [PH3]/[SiH4]¼ 0.01, hydrogenated

silicon (Si:H) deposited onto c-Si wafers with 1 lm of thermally grown ox-

ide (SiO2) coated with �330 A PECVD SiNx as a function of the relative

microcrystallite volume fraction, flc, in the top �250 A of the Si:H layer.

033714-5 Saint John et al. J. Appl. Phys. 110, 033714 (2011)

Downloaded 02 Nov 2011 to 131.183.220.113. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

in the material.29,30 These reference e spectra are combined

with a Bruggeman effective medium approximation so that

the relative microcrystallite fraction in the top �250 A is

extracted for the remainder of the mixed-phase (aþ lc)-Si:H

films in the series. The resistivity, TCR, and 1/f noise have

been evaluated as functions of the relative microcrystallite

fraction in this series of materials, as shown in Fig. 7. As

expected, the resistivity decreases with the increasing degree

of microcrystallinity, which is accompanied by a decrease in

the TCR and 1/f noise. The low TCR values obtained for the

predominately microcrystalline materials are consistent with

those of doped lc-Si in the respective resistivity range.32

High TCR values are observed for the a-Si:H samples,

although a relative microcrystallite fraction at the top sur-

face, flc� 0.07, can significantly reduce the TCR. The con-

trolled incorporation of smaller amounts of microcrystallites

in the bulk of the Si:H layer has yet to be methodically

evaluated.

A comparison of the TCR values as a function of resis-

tivity for both the n- and p-type Si:H materials is shown in

Fig. 8. It is clear that there are increases in TCR with increas-

ing resistivity for both n- and p-type doped Si:H materials,

and that the highest TCR values occur for materials in the

amorphous regime. Therefore, a comparison of the TCR and

1/f noise of all a-Si:H samples is shown in Fig. 9. For the

range of resistivity from �1000–3000 kX cm, the p-type a-

Si:H samples exhibit lower TCR values, but also lower 1/f

noise. This behavior indicates that 1/f noise and TCR are not

solely correlated to film resistivity, but that variations in the

material structure, such as that induced by the choice of car-

rier type, may be responsible for the control of these

properties.

When trying to understand the differences in 1/f noise

between n- and p-type samples of similar resistivity, a simple

interpretation can be made, provided that similar noise

mechanisms operate in a-Si:H with either dopant. It is gener-

ally assumed that the mobility of holes is lower than that of

electrons34 and thus, for films of similar resistivity, p-type a-

Si:H must have more charge carriers participating in conduc-

tion than n-type a-Si:H. The empirical noise equation given

by Hooge,35 suggests that 1/f noise is inversely proportional

to the number of carriers participating in the conduction.

This equation gives a simple interpretation for films of simi-

lar resistivity: p-type materials are expected to require

greater carrier densities to compensate for their lower carrier

mobilities, thus this difference in the number of conduction

participants yields lower 1/f noise. This reasoning assumes

that the Hooge parameter is independent of carrier type, con-

firmation of which demands accurate mobility measurements

for determination of the carrier densities and the Hooge pa-

rameter itself.

The TCR of a film is correlated with some activation

energy,1 although the meaning of this energy has multiple

interpretations. A simplistic model would reason that this

energy would be the separation between an effective mobil-

ity edge and the Fermi level.34,36 If the charge carriers

around the mobility edge are considered as two populations,

those moving through the higher energy extended states with

mobilities near that of crystalline silicon and those trapping

and de-trapping in the band tails with very low mobility, the

interaction between these two populations may offer an

approach to understanding the role of localized states and the

influence of dopant type on the TCR. The density of states at

the conduction band tail, which is known to have a steeper

slope, may result in a more thermally-responsive carrier pop-

ulation, accounting for the higher TCR measured in the n-

type material. If these band tail states assist in the transport

by providing additional carriers that can be subsequently

promoted to extended states, then a steeper slope is expected

to cause an increase in the temperature dependence of the

effective carrier mobility. Without more detailed knowledge

of the band edges and their interactions with extended states,

it is difficult to differentiate between the influences of local-

ized tail state densities, defects, midgap states, and more

practical problems, such as the role of interfaces.

SUMMARY

Doped n- and p-type Si:H thin films for uncooled micro-

bolometer applications have been studied using SE and elec-

trical property measurements. In situ real time SE has been

used to study microstructural variations in a series of a-Si:H

FIG. 8. The activation energy and temperature coefficient of resistance plot-

ted for all n- and p-type Si:H films as a function of the film resistivity, q.

FIG. 9. (a) Temperature coefficient of resistance and (b) 1/f noise at

1 Hz plotted for all n- and p-type a-Si:H films as a function of the film

resistivity, q.

033714-6 Saint John et al. J. Appl. Phys. 110, 033714 (2011)

Downloaded 02 Nov 2011 to 131.183.220.113. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions

p-type materials, while ex situ SE has been used to determine

common amorphous and microcrystalline complex dielectric

functions for n-type a-Si:H and lc-Si:H and the relative

microcrystallite content in the top surface of (aþ lc)-Si:H

films. For the p-type a-Si:H series, variations in H2-dilution

result in increases in the resistivity, TCR, and 1/f noise with

increasing R. The n-type Si:H series is used to assess the role

that variations in the degree of microcrystallinity play on the

electrical properties. With increasing microcrystallinity,

expected decreases in the resistivity, TCR, and 1/f noise are

observed, with a sharp decrease in TCR for flc> 0.07. Com-

parisons of TCR and 1/f noise for n- and p-type a-Si:H mate-

rials over the same resistivity range show that the p-type

material exhibits lower TCR values but also lower 1/f noise.

The 1/f noise behavior may be rationalized as an effect of

the increased carrier density required by p-type a-Si:H to

have similar resistivity to n-type material; an artifact of hav-

ing matching resistivities with carrier types of different mo-

bility. An explanation of the differences in TCR behavior

requires further study of the interactions between the con-

duction and valence band tail carrier populations with more

mobile extended state populations, since dynamic transport

through localized and extended states cannot be de-con-

volved from the effective activation energy observed in sim-

ple TCR measurements.

ACKNOWLEDGMENTS

This research was sponsored by the U.S. Army Research

Office and U.S. Army Research Laboratory and was accom-

plished under Cooperative Agreement Number W911NF-0-

2-0026. A portion of the research reported in this paper has

been supported by L-3 Internal Funds.

1A. J. Syllaios, T. R. Schimert, R. W. Gooch, W. L. McCardel, J. H. Tregil-

gas, and B. A. Ritchey, Mater. Res. Soc. Symp. Proc. 609, A14.4.1 (2000).2For example: J. F. Brady, T. R. Schimert, D. D. Ratcliff, R. W. Gooch, B.

Ritchey, P. McCardel, K. Rachels, S. J. Ropson, M. Wand, M. Weinstein,

and J. Wynn, Proc. SPIE 3698, 161 (1999).3J. L. Tissot, F. Rothan, C. Vedel, M. Vilain, and J. J. Yon, Proc. SPIE

3436, 605 (1998).4For example: T. Schimert, J. Brady, T. Fagan, M. Taylor, W. McCardel,

R. Gooch, S. Ajmera, C. Hanson, and A. J. Syllaios, Proc. SPIE 6940,

694023 (2008).5J. Koh, Y. Lee, H. Fujiwara, C. R. Wronski, and R. W. Collins, Appl.

Phys. Lett. 73, 1526 (1998).6D. V. Tsu, B. S. Chao, S. R. Ovshinsky, S. Guha, and J. Yang, Appl. Phys.

Lett. 71, 1317 (1997).7J. Yang, K. Lord, S. Guha, and S. R. Ovshinsky, Mater. Res. Soc. Symp.

Proc. 609, A15.4 (2000).

8O. Vetterl, F. Finger, R. Carius, P. Hapke, L. Houben, O. Kluth, A. Lam-

bertz, A. Muck, B. Rech, and H. Wagner, Sol. Energy Mater. Sol. Cells

62, 97 (2000).9P. W. Kruse, Uncooled Thermal Imaging (SPIE, Bellingham, Washington,

2001).10R. W. Collins, A. S. Ferlauto, G. M. Ferreira, C. Chen, J. Koh, R. J. Koval,

Y. Lee, J. M. Pearce, and C. R. Wronski, Sol. Energy Mater. Sol. Cells 78,

143 (2003).11J. M. Pearce, N. Podraza, R. W. Collins, M. M. Al-Jassim, K. M. Jones, J.

Deng, and C. R. Wronski, J. Appl. Phys. 101, 114301 (2007).12J. A. Stoke, N. J. Podraza, J. Li, X. Cao, X. Deng, and R. W. Collins, J.

Non-Cryst. Solids 354, 2435 (2008).13N. J. Podraza, J. Li, C. R. Wronski, E. C. Dickey, and R. W. Collins, J.

Vac. Sci. Technol. A 27, 1255 (2009).14J. Lee, P. I. Rovira, I. An, and R. W. Collins, Rev. Sci. Instrum. 69, 1800

(1998).15B. Johs, J. A. Woollam, C. M. Herzinger, J. N. Hilfiker, R. Synowicki, and

C. Bungay, Proc. Soc. Photo-Opt. Instrum. Eng. 72, 29 (1999).16Y. Cong, I. An, K. Vedam, and R. W. Collins, Appl. Opt. 30, 2692 (1991).17G. E. Jellison, Jr., and F. A. Modine, Appl. Phys. Lett. 69, 371 (1996);

ibid. 69, 2137 (1996).18For example: R. W. Collins and A. S. Ferlauto, in Handbook of Ellipsome-

try, edited by H. G. Tompkins and E. A. Irene (William Andrew, Norwich,

NY, 2005), pp. 125–129.19H. Fujiwara, J. Koh, P. I. Rovira, and R. W. Collins, Phys. Rev. B 61,

10832 (2000).20C. M. Herzinger, B. Johs, W. A. McGahan, J. A. Woollam, and W. P.

Paulson, J. Appl. Phys. 83, 3323 (1998).21N. J. Podraza, C. R. Wronski, and R. W. Collins, J. Non-Cryst. Solids 352,

950 (2006).22N. J. Podraza, C. R. Wronski, and R. W. Collins, J. Non-Cryst. Solids 352,

1263 (2006).23N. J. Podraza, C. R. Wronski, M. W. Horn, and R. W. Collins, Mater. Res.

Soc. Symp. Proc. 910, A. 03.2.1 (2006).24N. J. Podraza, C. R. Wronski, M. W. Horn, and R. W. Collins, Proceedings

of the 4th World Conference on Photovoltaic Energy Conversion, 1657

(Honolulu, Hawaii, 2006).25Y. A. Kryukov, N. J. Podraza, R. Collins, and J. G. Amar, Phys. Rev. B

80, 085403 (2009).26J. K. Rath, R. E. I. Schropp, and P. R. Cabarrocas, Phys. Status Solidi C 5,

1346 (2008).27A. S. Ferlauto, G. M. Ferreira, J. M. Pearce, C. R. Wronski, R. W. Collins,

X. Deng, and G. Ganguly, J. Appl. Phys. 92, 2424 (2002).28Y. Lu, S. Kim, M. Gunes, Y. Lee, C. R. Wronski, and R. W. Collins,

Mater. Res. Soc. Symp. Proc. 336, 595 (1994).29A. S. Ferlauto, J. Koh, P. I. Rovira, C. R. Wronski, R. W. Collins, and G.

Ganguly, J. Non-Cryst. Solids 266–269, 269 (2000).30D. E. Aspnes, A. A. Studna, and E. Kinsbron, Phys. Rev. B 29, 768

(1984).31S. Filonovich, H. Aguas, I. Bernacka-Wojcik, C. Gaspar, M. Vilarigues, L.

Silva, E. Fortunato, and R. Martins, Vacuum 83, 1253 (2009).32P. I. Rovira, A. S. Ferlauto, R. J. Koval, C. R. Wronski, R. W. Collins, and

G. Ganguly, Mater. Res. Soc. Symp. Proc. 609, A19.6 (2000).33J. Perrin, Y. Takeda, N. Hirano, Y. Takeuchi, and A. Matsuda, Surf. Sci.

210, 114 (1989).34For example: R. A. Street, Hydrogenated Amorphous Silicon (Cambridge

University Press, New York, 1991), pp. 124–142.35F. N. Hooge, Phys. Lett. A 29, 139 (1969).36T. Abtew, M. Zhang, Y. Pan, and D. A. Drabold, J. Non-Cryst Solids 354,

2909 (2008).

033714-7 Saint John et al. J. Appl. Phys. 110, 033714 (2011)

Downloaded 02 Nov 2011 to 131.183.220.113. Redistribution subject to AIP license or copyright; see http://jap.aip.org/about/rights_and_permissions