NBS
EPA
National
Bureau of
Standards
Center for
Analytical Chemistry
Washington DC 20234
U.S. Department
of Commerce
United States     Office of Environmental Engineering
Environmental Protection and Technology
Agency       Washington DC  20460
            EPA-600 7 80-123
            June 1980
            Research and Development
            The  Development of
            Potential Thin
            Standards for
            Calibration of X-ray
            Fluorescence
            Spectrometry

            Interagency
            Energy/Environment
            R&D  Program
            Report

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional  grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

      1.  Environmental Health  Effects Research
      2.  Environmental Protection Technology
      3.  Ecological Research
      4.  Environmental Monitoring
      5.  Socioeconomic Environmental Studies
      6.  Scientific and Technical Assessment Reports (STAR)
      7.  Interagency Energy-Environment  Research and Development
      8.  "Special" Reports
      9.  Miscellaneous Reports

This report has been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid  development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments of, and development of, control technologies for energy
systems; and integrated assessments of a wide range  of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
THE DEVELOPMENT OF POTENTIAL THIN STANDARDS FOR
CALIBRATION OF X-RAY FLUORESCENCE SPECTROMETRY
                      by
                  P. A. Pella
        Center for Analytical Chemistry
         National Bureau of Standards
             Washington, DC  20234

-------
                                  DISCLAIMER
     This report has been prepared and reviewed by the Center for Analytical
Chemistry and the Office of Environmental Measurements, National Bureau of
Standards, and reviewed by the U. S. Environmental Protection Agency, and
approved for publication.  Approval does not signify that the contents
necessarily reflect the views and policies of the U. S. Environmental Protec-
tion Agency.  In order to adequately describe materials and experimental
procedures, it was occasionally necessary to identify commercial products by
manufacturer's name or label.  In no instance does such identification imply
endorsement by the National Bureau of Standards or the U, S. Environmental
Protection Agency nor does it imply that the particular products or equipment
is necessarily the best available for that purpose.
                                     ii

-------
                                  FOREWORD
     The role of the National Bureau of Standards (NBS) in the Interagency
Energy/Environment R&D program, coordinated by the Office of Research and
Development, U. S. Environmental Protection Agency, is to provide those
services necessary to assure data quality in measurements being made by a
wide variety of Federal, State, local, and private industry participants in
the entire program.  The work at NBS is under the direction of the Office of
Environmental Measurements and is conducted in the Center for Analytical
Chemistry, the Center for Radiation Research and the Center for Thermody-
namics and Molecular Science.  NBS activities are in the Characterization,
Measurement, and Monitoring Program category and addresses data quality
assurance needs in the areas of air and water measurement methods, standards,
and instrumentation.  NBS outputs in support of this program consist of
the development and description of new or improved methods of measurement,
studies of the feasibility of production of Standard Reference Materials for
the calibration of both field and laboratory instruments, and the development
of data on the physical and chemical properties of materials of environmental
importance in energy production.  This report is one of the Interagency
Energy/Environment Research and Development Series reports prepared to
provide detailed information on an NBS measurement method or standard develop-
ment.
William H. Kirchhoff, Chief
Office of Environmental Measurements
National Bureau of Standards
                                     iii

-------
                                  ABSTRACT
     Thin films containing known concentrations of metals are important for
the calibration of X-ray Fluorescence Spectrometry (XRF), especially for the
analysis of collected airborne particulate matter.  A focused ion-beam
sputtering technique has been investigated as a candidate method for
fabricating thin glass films containing known concentrations of metals on
polycarbonate substrates.  Glass targets were fabricated at NBS for these
studies, and parameters such as ion-acceleration voltage and ion current were
systematically varied to determine any changes in film composition.  It was
found that rather severe changes in instrumental parameters do not affect the
elemental composition of the films appreciably.  Up to eight substrates were
coated at one time and the compositional reproducibility as measured by XRF
for Si, Ca, Zn, and Pb for 13 samples was within five percent relative
standard deviation at mass loadings of glass from 160 - 190 pg/cm2.  Glass
films containing phosphorous and sulfur were also prepared to demonstrate
the feasibility of preparing glass films containing such elements of low
atomic number.   Additional studies consisted of the deposition of finely
ground synthetic glasses on membrane filters, and the characterization of
some selected commercial thin films prepared by thermal evaporation.
                                     iv

-------
                                  CONTENTS
Foreword	     iii

Abstract	      iv

Figures	      vi

Tables	     vii

     1.   Introduction 	       1

     2.   Experimental Procedures  	       4

               Focused-Ion Beam Sputtering 	       4

               Study of Polycarbonate Substrates 	       4

               Synthesis of Glass Targets  	       7

               Initial Experiments 	       9

               Glass Film Composition Studies A	      17

               Glass Film Composition Studies B	      21

               Effect of Sputtering Parameters on Glass Film
                  Composition	      22

               Radial Composition of Glass Films 	      27

               Supplementary Studies 	      33

               Conclusion	      36

References	      38

-------
                                    FIGURES
Number                                                                    Page

  1   Cascade collision resulting in sputtering	  .     5

  2   Schematic diagram of sputtering apparatus	     6

  3   X-ray fluorescence spectrum of glass target A  	    10

  4   XRF spectrum of sputtered glass film (target A)	    11

  5   NBS holder for 37 mm polycarbonate substrates.  A = outer ring;
      B = inner ring; C = polycarbonate substrate  	    12

  6   Model MIM/TLA 15 (front view)	    15

  7   View into chamber of MIM/TLA 15.  A = rotating substrates;  B =
      glass target	    16

  8   SEM photo of target B film:  Mass loading = 50 ug/cm2	    25

  9   SEM photo of target B film:  Mass loading = 160 pg/cm2	    26

 10   EDXRF spectrum of film from target C	    29

 11   EDXRF spectrum of film from target D	    30

 12   Al strip holder	    31
                                     VI

-------
                                    TABLES


Number                                                                    Page

  1   Nominal Composition of Glass Targets (A-D) Fabricated at NBS ....  8

  2   Element X-ray Intensities from Sputtered SiC>2 Films Compared
        with a Blank Polycarbonate Substrate (MIM/TLA 5.5) 	 13

  3   X-ray Intensity Data for Various Elements Vs. Sputtering Time
        (Target A)	14

  4   X-ray Conditions Used for EDXRF Analysis 	 18

  5   Analysis of Glass Films From Sputtering Target A (Instrument MIM/TLA
        5.5) Weight Percent Found	19

  6   Analysis of Glass Films From Sputtering Target B (Instrument MIM/TLA
        5.5) Weight Percent Found	20

  7   Comparison of the Average Composition of Sputtered Films Using Two
        Different Instruments (Target B) 	 22

  8   Summary of Sputtering Conditions Employed	23

  9   Composition of Glass Films Determined by EDXRF (Target B)	 24

 10   Composition of Glass Films at Higher Mass Loadings (Target B).  .  . .28

 11   Composition of Glass Films Deposited on Substrates Held by
        Stationary Strip Holder	32

 12   Comparison of Mass Loadings of Various Metal Films on Polycarbonate
        Substrates by Gravimetry 	 34

 13   Ratio of Intensity to Mass Loading (I/M)  for Zn,  Pb, Fe, and
        Ti Films (s = standard deviation of a single measurement)	35
                                     vii

-------
                                    SECTION 1

                                  INTRODUCTION


     X-ray fluorescence spectrometry has proven to be a versatile and rapid
method for multielement analysis in many analytical applications.  Of particu-
lar interest are substances which can be analyzed in the form of a deposit on
a filter, mesh or membrane, or as a thin pressed pellet.  Collected airborne
particulate matter, particulates in waste water and geochemical samples are
just a few examples of this type.  In some respects such "thin samples" are
ideal for x-ray fluorescence analysis because the detection limits are
suitably low and interelement effects such as x-ray absorption and/or enhance-
ment are negligible.  The absence of any sizable interelement effects makes
possible a linear instrument response to element mass per unit area.  Some
problems associated with the "thin specimen model" are x-ray self-absorption
in particles notably for the light elements such as Al, Si, P, and S commonly
referred to as particle-size effects, and self-absorption in the substrate.
Both of these effects require the application of correction factors.  There-
fore, the basic premise is that accurate multielement analysis of thin
samples can be performed based on the thin specimen model, together with
empirical or theoretical correction factors.

     For a thin specimen the count rate I. of characteristic x-rays from an

element i is linearly related to the element mass per unit area, m. (g/cm2):

                                   I± - S. m.                              (1)


where S. is the sensitivity or calibration factor for element i, and is in

units of counts/s per pg/cm2.  S, is usually determined experimentally with

thin standard samples.  A thin specimen is arbitrarily defined as one having
a total mass per unit area given by*

                                        0.1

                                    "" T                                (2>

where the effective, average mass absorption coefficient y in cm2/g is

                          U = PI cosec 0j + y2 cosec 62                    (3)

*assumes x-ray photon excitation

-------
where vi is the sample mass  absorption coefficient  for  the primary radiation
energy, M2 is  tnat f°r tne characteristic  x-ray  energy, and 6j and 63 are the
glancing angles of incidence and  emergence,  respectively.  The deviation from
linear response due to sample self-absorption  is approximately equal to 0.5 um.
This so-called critical  thickness is  a good  guide to determining sample "thin-
ness" and can be estimated from equation 2.

     There are several types of "thin" samples presently available from com-
mercial sources (1,2) which  are used  to calibrate x-ray spectrometers.  These
generally consist of:

     (1)  Thin films of  single metallic elements deposited by thermal evapora-
tion on mylar  or thin aluminum substrates.   The  mass loading is usually deter-
mined by the manufacturer using gravimetry.

     (2)  Dried solution deposits on  membrane  filters.  These are prepared by
a multi-drop technique where ten  or more elements are present in concentrations
ranging from 1 to 50 mg/cm2  as stated by the manufacturer.

     (3)  Synthetic particulate deposits consisting of previously characterized
rock samples ground and  dispersed on  membrane  filters and fixed to the filter
with a 30-40 mg/cm2 layer of paraffin wax.

     In addition to commercial sources, a number of workers have reported on
techniques for preparing thin standard specimens (3-11).  Pella et al. (3)
reported on a technique  for  depositing a reground NBS-Standard Reference
Material (NBS-SRM 1571 Orchard Leaves)  on membrane filters and coated with
vapor-deposited Parylene films.   Baum et al. (6) described a capillary matrix
and lyophilization procedure for  preparing solution deposited standard samples.
Dzubay et al.  (9) have developed  calibration samples consisting of polymer
films each containing a  single calibration element.  The film is cast from a
homogeneous solution containing known amounts  of an organometallic compound
and a polymer.  Olson et al.  (10)  used an air  filter calibration facility for
preparing standard samples.   Recently,  Semmler et al. (11) describes a tech-
nique where micron sized particles of known  compounds suspended in a carrier
solution are deposited onto  the surface of a Nuclepore filter.   Particle size
is controlled by a separate  sedimentation step following grinding in a boron
carbide mortar and pestle.   Binding of the deposit to the filter is accom-
plished with collodion film.

     There are several important  criteria which  need to be met  if thin specimen
samples are to be acceptable for  calibration purposes.   First,  the uniformity
of the sample as well as its homogeneity must be within an acceptable well-
defined tolerance limit.  If the  sample contains particulate matter, both the
size and distribution should be characterized.   Secondly, the mass loading of
the sample must be known to  an acceptable degree of accuracy.   And finally,
the sample should possess sufficient  durability  and integrity.   The goal of
the NBS-EPA XRF program  is to perform the research and development required to
produce thin standard samples especially useful  for calibration of x-ray
fluorescence analysis of collected airborne  particulate matter.  For this pur-
pose we have investigated candidate methods  for  preparing such samples and the
results are summarized in this report.  These  consist of (a)  the deposition of

-------
thin glass films on polycarbonate substrates by focused-ion beam sputtering
(12-15), (b) the deposition of NBS-SRM 1648 urban particulate on membrane fil-
ters, and (c) the deposition of finely ground synthetic glasses on membrane
filters.  In addition, we have also characterized some selected thin films from
a commercial supplier prepared by thermal evaporation to assess the state-of-
the-art for some types of thin films.  The sputtering experiments described in
this report were performed initially with the model MIM/TLA 5.5 instrument,
and subsequent work with the larger MIM/TLA 15 system.

-------
                                    SECTION 2

                             EXPERIMENTAL PROCEDURES
1.  FOCUSED-ION BEAM SPUTTERING

     Ion sputtering is essentially a technique which involves the transfer of
molecules from a target to a substrate by the action of an ion beam incident
upon the target as shown in Figure 1.  The ion gun  is of the field accelerated
design.  Permanent magnets cycloidally accelerate electrons thermally emitted
from a double-filament thoriated tungsten electrode (see Figure 2).  These
electrons produce ionization of an inert gas, usually argon, thus creating a
high intensity plasma.  The argon ions are extracted from the plasma and accel-
erated at the target using optically aligned molybdenum grids.  The resulting
15 cm diameter ion beam is neutralized using an immersion flood neutralizer.
The neutralized ion beam is further collimated to about 12 cm with a baffle
plate coated with high purity aluminum oxide to reduce contamination of the
sputtered film due to the container walls.  The ion beam is accelerated at high
potential from 700-1100 volts at a flux of 0.5 mA/cm2 toward the target at a
45° angle.  Since the substrate is not in the plasma environment, substrate
damage does not occur and is a distinct advantage of this particular sputter-
ing technique.  Also, operating parameters such as  deposition rate, angle of
deposition, and temperature which can effect thin film properties can be varied
easily and independently.  The sputtered material is deposited at a relatively
low energy, typically 100 eV, so that no impact damage to the substrates result.
To ensure deposits of high purity the system is operated at low pressure (10~5
mm Hg range).  The main advantages to be gained by using such a technique
appear to be as follows:

     (1)  Production of standard thin films from synthetic materials such as
glasses with known elemental composition which can be varied as desired.

     (2)  High adherences of the films to the substrate which obviates the
need for adhesive materials to maintain the standard material on the substrate.

     (3)  Easy control of the mass loading on the substrate by variation of the
sputtering time.


2.  STUDY OF POLYCARBONATE SUBSTRATES

     Membrane filter materials such as Millipore and Nuclepore are finding
increasing use as substrates for preparing samples for x-ray fluorescence anal-
ysis.   This is because these materials have relatively low levels  of trace

-------
INCIDENT  ION
                                        SPUTTERED ATOMS
                                              PUTTERING  SOURCE SURFACE
               Figure 1.   Cascade collision resulting in sputtering.

-------
                  Source Filament
                /
          Anode
         \
Glass Target \
              \
               \
                        . 1st Stage Accelerator
                       - 2nd
                       Flow  Neutralizer
                               Collimator
                            Substrate Holder
     Figure 2.  Schematic diagram of sputtering apparatus.

                     6

-------
elements and their mass is quite low (e.g. on the order of 1 mg/cm2).  Nuclepore
is a polycarbonate material which has quite distinct advantages over Millipore
in that it contains lower concentrations of trace elements such as Cl and Fe
and is more durable under x-ray photon bombardment.  We have checked the purity
and measured the effects of humidity on polycarbonate material.  The need to
assess the effect of humidity is important in order to use gravimetry for
measuring the mass loading of the films.  Examination of the polycarbonate
by energy-dispersive x-ray spectrometry (EDXRF) revealed detectable amounts of
Ca and Fe.  However, these concentrations are believed to be at levels suffi-
ciently low not to cause any blank problems.

     A number of Nuclepore polycarbonate filters were weighed to .001 mg in a
constant humidity environment (55 ± 1% relative humidity) and at room tempera-
ture 24 ± 1% °C.  Variations in the weight of each filter were on the order of
± 0.005 mg which was a little less than 0.1 percent of the total mass of the
filter.  The first study of humidity effect was done by exposing a number of
filters to 100 percent relative humidity environment for two days and then
removing these and immediately reweighing them in 50 percent relative humidity
environment.  The weights were recorded periodically over a time span of about
four hours.  After about ten minutes the filter weights were in agreement with-
in 0.1 percent of the original weights of the filters.

     A second study was performed in a similar manner by placing preweighed
filters in an 11 percent relative humidity environment for two days and re-
weighing them after ten minutes in the 50 percent relative humidity environ-
ment.  Again the filter weights were within 0.1 percent agreement with these
preweighed values.  These studies showed that these filters quickly reached
equilibrium with the environment in which they were weighed and indicated that
changes in humidity do not cause any serious problems.
3.  SYNTHESIS OF GLASS TARGETS

     Synthetic glasses were selected for producing thin films because these
materials can, in principle, be fabricated containing various elements of
interest at desired concentration levels.  Several glass disks were prepared
by D. Blackburn and D. Kaufmann of the NBS Inorganic Glass Section.  In the
initial studies performed with the MIM/TLA 5.5 instrument, glass disks 7.6 cm
in diameter were fabricated.  For work with the model MIM/TLA 15 glass targets
17.8 cm in diameter were cast.  The nominal composition of the glass targets
is summarized in TABLE 1.

-------
      TABLE 1.   NOMINAL COMPOSITION OF GLASS TARGETS (A-D) FABRICATED AT NBS
      Compound
Si02
A1203
Fe30^
CaO
MgO
KoO
36.0 36.0 45.0
13.5
13.5
18.0 18.0
9.0
10.0
—
—
—
—
—
__
      ZnO                  —            20.0

      PbO                  —            26.0
      Na2SOi4               —             —           20.0

      Na20                 —             —           35.0

      A1(P03)3             —             —             —           70.0

      NaP03                —             —             --           30.0
The constituents for  target A were  fused  at  1511  °C  for  three hours in a Pt
crucible with continuous stirring.   The melt was  cast  into a 7.6 cm diameter
mold on a preheated hot plate and then annealed at 685 °C for ten minutes and
then allowed to cool  to room temperature.  Target B  constituents were fused at
1400 °C, and poured into molds  (7.6 and 17.8 cm)  at  1200 °C, and annealed over-
night at 650 °C.  Fusion and annealing temperatures  for  targets C and D (17.8
cm) were 1225 °C and  420 °C, and 1300 °C  and 525  °C, respectively.  For casting
targets in the 17.8 cm mold about 700-900 g  of glass was melted.  The glass
target surface was first coarsely ground  on  a diamond disk (220 grit) then
polished with 22.5 y  Al203, then 9.5 p, and  finally  0.5 y sapphire.  The 17.8
cm disks were all ground to a thickness of 0.3 cm.  The homogeneity of the
target A surface was  studied to at  least  a depth  of  2 ym by scanning electron
microscopy fitted with a Si (Li) detector.   The electron beam was scanned
across at least 25 different points on the target, and the characteristic
x-rays of Mg, Al, Si, K, Ca, and Fe were measured.  The results indicated that
the target was homogeneous within the Poisson counting statistics error of 1-2%.
Targets B, C, and D were examined by EDXRF and were found to be homogeneous
with respect to elemental content within  3%.

-------
4.  INITIAL EXPERIMENTS

     Initial experiments were designed to obtain information on sputter rate
and the cleanliness of the ion-beam sputtering apparatus which was used.
Since this machine (model MIM/TLA 5.5) was used primarily for demonstration
purposes, and was in almost continuous operation for producing films of several
elements, the sputter gun was disassembled and was cleaned thoroughly before
sputtering our targets.  A disk of high-purity SiC>2 was sputtered for five
hours on one polycarbonate substrate and the resulting Si02 film was examined
by x-ray fluorescence.  The main impurities found were Fe, Mn, Cr and Ni,
which were attributed to the sputtering of the stainless steel ion current
monitor by the primary beam.  A small amount of Al contamination arose from
sputtering of the walls of the ion gun.  Four deposits of the synthesized glass
target A were sputtered for various times to obtain information on the sputter
rate and to determine the length of time necessary to obtain a film of adequate
thickness to provide good x-ray counting statistics for the XRF measurements.
In Figures 3 and 4 are xrf spectra of the glass target A and film, respec-
tively.  From these experiments it was determined that under a particular set
of machine operating parameters, films from 100 to 400 ug/cm2 could be depos-
ited in 4 to 16 hours.  The sputtering time could be shortened by using higher
beam currents but it was found that considerable heating of the polycarbonate
substrates occurred with resultant damage.  Henceforth, machine operating
conditions were deliberately chosen so that the substrate temperature did not
exceed 80 °C.  Temperature studies of the polycarbonate substrate filters
(37 mm diameter) showed that cycling temperatures from 25 to 80 °C did not
change the weight of the filters when remeasured at room temperature.  It was
also found that the glass-coated polycarbonate substrates tended to wrinkle
when removed from the substrate holder.  This made it extremely difficult to
remount in a flat reproducible configuration for xrf measurements.  For this
reason an aluminum ring holder was constructed to maintain the substrates flat
as shown in Figure 5.  The assembly could then be transferred directly to our
xrf instrument for measurement.  Both Teflon and aluminum insert rings were
made for use in order to keep the substrate in a stretched state in its alumi-
num ring holder.  The Teflon insert ring was especially useful when the glass-
coated substrate needed to be demounted for weighing to measure the mass of
the glass deposit.

     To reduce the Fe and Cr contamination from the sputtering system, the
stainless steel ion current monitor was vnrapped with Al foil and all surfaces
in the system where possible sources of contamination could arise were lined
with Al foil.  Because the substrate holder in the system was made to accom-
modate only one sample at a time, another holder was constructed at NBS to
accommodate three samples at one time which rotated when the glass films were
being deposited.  A high purity SiC>2 target was then sputtered for 720 minutes
to provide three samples.  The films were removed and examined for impurities
by energy-dispersive xrf using a Ni secondary emitter for excitation.  The
resultant intensities were compared with a blank substrate tabulated in TABLE
2.  As seen from the table, appreciable amounts of Al, Ar, and Fe appear in the
sputtered Si02 films.  The Ar contamination arises from the argon entrapped in
the glass during the sputtering operation.

-------
Figure 3.
X-ray fluorescence spectrum of glass target A.
                10

-------
Figure 4.  XRF spectrum of  sputtered  glass film  (target A)
                            11

-------
Figure 5.  NBS holder for 3/ mm polycarbonate substrates.  A = outer ring; B = inner ring; C = polycarbo-
           nate substrate.

-------
                                     TABLE 2

        ELEMENT X-RAY INTENSITIES FROM SPUTTERED Si02 FILMS COMPARED WITH

                  A BLANK POLYCARBONATE SUBSTRATE  (MIM/TLA 5.5)
        Film
  Gross Counts (Average) in 1000 Seconds

Al         Si        Ar        K       Ca
Fe
Si02 #1
Si02 #2
Si02 #3
Blank
16,134
17,494
14,188
2,098
186,412
179,072
178,212
1,391
28,131
27,936
26,944
1,474
3,847
3,978
3,853
1,212
939
1,019
941
1,451
14,724
14,903
14,645
1,834
XRF Conditions:  45 KV, 30 MA, Ni secondary emitter, in vacuum
To get an estimate of the radial uniformity of the sputtered glass film seven
1 X 1 cm vitreous carbon squares were distributed between and around two poly-
carbonate substrates on the substrate holder.  The glass target A was then
mounted and sputtered for 20 hours onto each of the polycarbonate substrates
and the seven carbon squares.  This gave a film approximately 400 yg/cm2.  The
films on five of the carbon substrates were examined by SEM using a Si (Li)
x-ray detector.  The K^ x-ray intensities of Al, Si, K, Ca, and Fe were meas-
ured on each of the carbon squares and were found not to vary by more than 3%
of the measured x-ray intensity.

     To obtain information on the mass loading of the glass films with time of
sputtering, three sets of three substrates each were glass coated for 180, 360,
and 540 minutes, respectively, and the x-ray intensities were measured by
EDXRF.  A summary of this data is presented in TABLE 3.  As seen from the
table, the mass loading of the film does increase approximately in proportion
to the sputtering time.

     The results of these preliminary studies indicated that glass targets
appeared to be suitable for producing thin standards by this technique.  The
uniformity of these films were acceptable.  Several problems, however, did
arise as a result of these studies.  First, elemental contamination of the
films from the sputtering environment was significant.  Secondly, target heat-
ing became appreciable during ion bombardment and also caused localized heating
of the substrates with some apparent damage.  Also, the instrument used could
not accommodate more than three substrates at one time which was clearly
unsatisfactory.  These problems were overcome with the advent of the larger
Model MIM/TLA 15 instrument which became available to us in 1978 (see
Figures 6 and 7).  The main features of this larger instrument are sum-
marized as follows:
                                      13

-------
TABLE 3.  X-RAY INTENSITY DATA FOR VARIOUS ELEMENTS VS. SPUTTERING TIME
                               (TARGET A)

Sputtering
_. . Sample
Time, mm. *
180 1
2
3
Average =
Rel. Std. Dev.
360 1
2
3
Average =
Rel. Std. Dev.
540 1
2
3
Average =
Rel. Std. Dev.
Al
17,169
16,440
17,567
17,059
3.4%
25,381
25,855
26,089
25,775
1.4%
36,568
36,211
37,659
36,813
2.0%
Si
23,791
23,000
25,341
24,044
5.0%
34,762
33,073
34,760
34,198
2.8%
42,304
40,907
43,825
42,345
3.4%
Ar
42,511
39,939
44,436
42,295
5.3%
71,986
72,956
74,745
73,229
1.9%
119,766
119,502
127,401
122,223
3.7%
K
80,734
78,267
84,688
81,230
4.0%
128,505
124,147
130,598
127,750
2.6%
156,215
151,758
163,889
158,287
3.9%
Ca
293,873
283,927
308,935
295,578
4.3%
485,841
481,355
495,574
487,590
1.5%
748,193
733,963
784,836
755,664
3.5%
Fe
643,062
620,588
687,640
650,430
5.2%
1,031,281
989,017
1,044,297
1,021,532
2.8%
1,328,058
1,271,554
1,385,397
1,328,336
4.3%

-------
Figure 6.  Model MIM/TLA15 (front view)
                    15

-------
Figure 7.   View into chamber of MIM/TLA 15.
           target.                     16
A = rotating substrates; B = glass

-------
     (1)  The substrate stage  (15.24 cm  in diameter) was water cooled, rotatable
at 1 RPM, and allowed up to eight  (37 mm diameter)  substrates to be coated
s imultaneously.

     (2)  The angle of beam incidence was variable  from 0°-180°.

     (3)  The target holder was a  15.24  cm diameter water-cooled stainless
steel plate.

     (4)  The stainless steel  chamber which housed  the target and substrate
assembles was 61 cm in diameter which allowed for easier loading and removal
of samples.

     These features allowed more samples to be coated simultaneously without
any significant buildup of heat in the system.  We needed, however, to fabri-
cate larger glass targets to accommodate the 15 cm diameter ion beam.  For
this purpose, we cast 17.8 cm  glass disks.  To obtain efficient heat transfer
through the glass, the target  was  ground and polished to a thickness of 0.3 cm
by the NBS glass section.  An  aluminum plate was fabricated in order to heat
sink the glass plate to the water-cooled target stage.  Binding of the glass
target to the aluminum plate was accomplished with a flexible silver-filled
epoxy (17).  A substrate holder was also made at NBS to perform simultaneous
glass deposition on eight 37 mm substrates.  Again experiments were performed
using a pure SiO£ target to assess any elemental contamination of the sputtered
film.  To eliminate contamination we found it necessary to further minimize
any interaction of the incident beam with the sides and bottom of the chamber.
For this reason we fabricated  an aluminum baffle plate placed between the beam
and the glass target in order  to collimate the beam diameter to 12.7 cm.
Although effective collimation was obtained, we found considerable elemental
contamination produced from the collimation plate itself.  This problem was
eliminated by coating the collimator plate with 99.99% A^Os by a flame-spraying
technique.  Subsequent experiments with  a SiOj target showed detectable amounts
of aluminum and copper in the  SiC>2 films but at blank levels that were accept-
able.
5.  GLASS FILM COMPOSITION STUDIES A

     The elemental composition of glass films produced from targets A and B was
determined by EDXRF and by electron probe microanalysis and compared to the bulk
glass composition of the targets.  Standards used for the XRF calibration con-
sisted of:  (a) reground MBS-SRM 1571 orchard leaves on membrane filters (3),
(b) thin films of zinc and lead produced by thermal evaporation of the metals,
and (c) Si02 on polycarbonate prepared by sputtering high purity quartz.  Ti,
Ni, and Mo x-ray targets were used to produce monochromatic excitation of the
samples.  The conditions used for the XRF analysis are summarized in TABLE 4.
                                      17

-------
              TABLE 4.  X-RAY CONDITIONS USED  FOR  ELEMENT ANALYSIS3
 Elements Analyzed           W tube               Exciter             Vacuum
Si, K, Ca
Pb,Zn
Fe
45 KV,
45 KV,
20 KV,
30mA
30mA
30mA
Ti
Mo
Ni
Yes
No
Yes
 rt
  Analysis time was 1000 seconds.

 For microprobe analysis, a square piece of a sputtered glass film  (F-1035) was
 mounted on a flat, polished vitreous carbon square  (1 cm x 1 cm) and was placed
 next to a small slice of the glass target so that the surfaces were coplanar.
 The specimen was then prepared for electron probe microanalysis in a conven-
 tional manner.  The electrons were accelerated at 10 KV to produce a probe
 10 )jm in diameter.  The range of an electron beam of this energy was calcu-
 lated to be about 1.2 microns for a silicate material.  The mass of the sput-
 tered glass film was about 400 yg/cm2 and approximately 2 micrometers thick.
 The results of the x-ray analyses of the thin films prepared with the model
 MIM/TLA 5.5 instrument are given in TABLES 5 and 6 for each of the glass tar-
 gets used.  The first column in each table indicates the time sequence in
 which the glass films were prepared.  The nominal compositions of the respec-
 tive glass targets are included in the tables for comparison.  Examination of
 the results shown in TABLE 5 (target A) indicate that the CaO content of the
 films appears to be the same from sample-to-sample and close to the nominal
 composition of the bulk glass at 18.0 percent.  Greater variation in concen-
 trations was noted for K and Fe, which were significantly below the nominal
 values.  The Si02 composition appeared to be close to the nominal value only
 in those films which were sputtered for the longest times, F-1035 (17 hrs)
 and F-7 (9 hrs).  From the results in TABLE 6 (i.e., target B) little variation
 in the CaO content of these films was noted, but almost twice the nominal value
was obtained.  The PbO to ZnO ratio in the films was comparable to the ratio
 in the bulk glass but the absolute values of the concentration were very low
 compared to the nominal values.   The results of the electron microprobe analysis
 of the F-1035 film prepared from target A were in general agreement with the
values obtained by x-ray analysis for K, Ca, and Fe and were 2 percent, 21
 percent, and 5 percent, respectively.
                                     18

-------
VO
               TABLE 5.  ANALYSIS OF GLASS FILMS FROM SPUTTERING TARGET A (INSTRUMENT MIM/TLA5.5)

                                            WEIGHT PERCENT FOUND

Sample ..
„ Sample
Sequence
1
2



3



4

F-1035
F-4
F-5
F-6
Average
F-ll
F-12
F-13
Average
F-7
F-9
% Si02
32.7
52.6
49.0
51.1
50.9
54.3
50.9
53.7
52.9
38.9
39.8
% CaO
17.9
19.9
19.0
19.3
19.4
18.2
16.9
18.0
17.7
18.5
19.2
% K20
3.0
4.8
4.5
4.7
4.7
4.7
4.5
4.7
4.6
2.7
2.8
% Fe3Cht
7.4
9.3
8.7
9.0
9.0
8.8
8.2
8.7
8.6
7.3
7.5
Wt. of Glass
Film, yg
2,070
997
1,059
1,012

688
702
685

1,604
1,582
Total Mass
pg/cm2
367
177
188
179

122
124
121

284
280
Sputtering
Time (hrs)
17
6



3



9

           Average         39.4       18.9      2.75       7.4


    Nominal Values of
    Bulk Glass:            36.0       18.0     10.0       13.5

-------
             TABLE 6.   ANALYSIS OF GLASS FILMS FROM SPUTTERING TARGET B (INSTRUMENT MIM./TLA5.5)

                                          WEIGHT PERCENT FOUND

Sample _ ,
_ Sample
Sequence
1


2



3



F-17
F-19
Average
F-23
F-24
F-25
Average
F-20
F-21
F-22
Average
% Si02
38.
37.
38.
50.
47.
50.
49.
43.
45.
42.
43.
5
4
0
9
7
5
7
0
4
6
7
% CaO
34.
33.
34.
35.
33.
35.
34.
31.
32.
31.
32.
6
4
0
3
2
0
5
8
9
2
0
% ZnO %
7.
7.
7.
9.
9.
9.
9.
7.
8.
7.
7.
3
0
2
5
5
6
5
3
6
6
8
8
7
8
9
9
9
9
8
9
8
8
PbO
.3
.8
•*
.5
.4
.4
.4
.4
.5
.6
.8
Wt. of Glass
Film, yg
1,568
1,583

692
679
678

583
561
613

Total Mass Sputtering
yg/cm2 Time (hrs)
278 6
281

123 3
120
120

103 ^2
99
109

Nominal Values of
Bulk Glass:
36.0
18.0
20.0
26.0

-------
     In summary, the results show a sizable variation in the elemental composi-
tion of these glass films as compared with the bulk glass using the model
MIM/TLA 5.5 instrument.  Reasonable agreement between the bulk glasses and the
films prepared from target A was obtained for some elements (e.g., Si and Ca)
only for the longest sputtering times.  This suggests that the deposition rate
or sputtering rate is not sufficient for the establishment of conditions in
which the rate of production and/or deposition of sputtered atoms is constant.
The operational parameters for sputtering were maintained nearly identical for
all the glass films prepared and presented in TABLES 5 and 6.  The sputtering
rate was approximately 15-20 A/min.

     Although a sizable variation was noted in the film composition compared to
the bulk glass composition, glass films produced in this manner can still be
potential standard reference samples provided that the film composition obtained
from any particular target is reproducible.  Therefore, a series of experiments
was performed in which instrumental parameters, such as ion-acceleration volt-
age, and ion-beam current, were systematically varied to determine their effect
on the elemental composition of the glass films.  At the same time, it was of
interest to determine  if the differences in the film vs. the bulk glass compo-
sition could be minimized with further modifications in the technique.


6.  GLASS FILM COMPOSITION STUDIES B

     We have focused on two possible causes for the observed differences shown
in TABLES 5 and 6 between the film and bulk glass target compositions.  First,
localized heating of the glass target surface in the model MIM/TLA 5.5 instru-
ment due to ion beam bombardment can cause diffusion of compounds deeper into
the glass target and away from the surface.  Secondly, when compounds are being
removed from the target by sputtering, metal-to-oxygen bonds can break so that
metal atoms rather  than the metal oxides are collected on the substrate.  These
effects can cause the  elemental concentration at the surface to change relative
to'the bulk concentration.  With the  availability of the model MIM/TLA 15 instru-
ment,  the  features  already mentioned  removed any localized heating effects.   In
addition to this change, a small partial pressure of oxygen  (^10-12%) was intro-
duced  into the  argon atmosphere during the sputtering operation.  This was done
to  aid the recombination with oxygen  of any  sputtered atoms  generated at the
target surface.

     Comparison of  the average  film composition obtained with the two instru-
ments  using target  B is presented  in  TABLE 7.   It can be  seen from these results
that  the addition  of oxygen  in  the sputtering  atmosphere  produced films whose
composition for Ca,  Zn,  and  Pb was closer  to  the bulk  glass  composition.  It is
interesting to  note that  although  preferential  sputtering  for the light elements
Si  and Ca, and  a decrease  for Zn and  Pb  still  exists,  the  absolute values of the
relative differences from the bulk composition  are  approximately  the  same.
                                       21

-------
         TABLE 7.  COMPARISON OF THE AVERAGE COMPOSITION OF SPUTTERED
                    FILMS USING TWO DIFFERENT INSTRUMENTS
                                  (TARGET B)
  Instrument         % Si02           % CaO           % ZnO            % PbO
1977 Average1
1978 Average2
Bulk Comp.
% Rel. A, 3)1
% Rel. A, a'2
43.7
46.4
36.0
+21
+29
32.0
22.9
18.0
+78
+27
7.8
14.5
20.0
-61
-28
8.8
19.5
26.0
-66
-25
a
 rercent relative difference between the film and bulk composition, i.e.,

                                Film-Bulk
                                  Bulk
1 --Model MIM/TLA 5.5
2 --Model MIM/TLA 15

     Another important result obtained by the addition of oxygen to the
sputtering atmosphere was that the amount of entrapped argon in the glass
film was substantially reduced.  Therefore, all subsequent work was performed
with 5-10% partial pressure of oxygen added to the sputtering atmosphere.
7.   EFFECT OF SPUTTERING PARAMETERS ON GLASS FILM COMPOSITION

     To study the effects of ion-acceleration voltage and ion-current on the
film composition, six sets of polycarbonate substrates were coated with glass
from target B.  In TABLE 8 is a summary of the instrument conditions employed.
The composition of the glass films measured by EDXRF are presented in TABLE
9.  It appears from the-data that'varying the ion-voltage and ion-current
does not affect the film composition significantly.  The best agreement
between sets was obtained in sets 2 and 3 which were done on the same day
under identical conditions.  The mass loadings of glass deposited varied from
40 pg/cm2 (set 4) to 160 pg/cm2 (set 5).   The best measurement precision
obtained was within set 5 which also corresponded to the set with the highest
mass loading.  The microstructures of the deposited films from target B were
examined by SEM and photomicrographs of two films at mass loadings of 50
Ug/cm2 and 160 pg/cm2 are presented in Figs. 8 and 9, respectively.  In
comparing the figures it can be seen that at higher glass mass loadings more
of the open polycarbonate structure is filled with deposited glass and thus
should be more uniform.  It is also believed that the better precision at the
higher mass loadings is due to the greater mass difference between film and
substrate which can be measured gravimetrically with greater accuracy and
precision.  The amount of glass deposited at 160 pg/cm2 loading corresponds
to about 1.0 mg and can .be measured with an accuracy of ± .01 mg.  To obtain


                                      22

-------
                      TABLE 8.  SUMMARY OF SPUTTERING CONDITIONS EMPLOYED5

Date
5/4/78
5/5/78
5/5/78
6/19/78
6/26/78
6/29/78
Set
No.
1
2
3
4
5
6
No.
in
Set
3
7
6
5
5
5
Time
hrs.
4.2
4.4
4.3
2.5
3
3
Filament
16-24.5/5.5
16-24.5/5.5
16-24.5/5.5
16-24.5/5.5
24.5/5.5
16/5.2
Plasma
Generator
60/2.6
60/2.6
60/2.6
60/2.6
60/2.6
90/.78
Ar
mm Hg
7xlO~5
7xlO~5'
7xlO~5
7x10" 5
7xlO~5
7xlO~5
02
mm Hg
lxlO~5
lxlO~5
IxlO"5
lxlO~5
lxlO~5
1x10 ~5
lon-
Accl.
Voltage
700-800
700-800
700-800
1,000
1000-1500
1,500
Neutralizer
46/2.6
j
46/2.6
46/2.6
46/2.6
46/2.6
46/2.6
Ion-
Current
45 mA
45 mA
45 mA
85 mA
119 mA
70 mA
Ion
Focus
90/40
90/40
90/40
90/40
90/40
150/10
Instrument Model MIM/TLA 15

-------
            TABLE 9.   COMPOSITION OF GLASS FILMS DETERMINED BY

                                    EDXRF


                                 (Target B)

Set No. In
No. Set
1 3
2 7

3 6

4 5

5 5

6 5

Overall Average

Bulk Comp.

AV.
s a
X
AV.
S
X
AV.
S
X
AV.
S
X
AV.
S
X
AV.
S
X

sAVb

% Si02
48.7
3.0
46.6
1.0
46.4
1.1
43.5
0.9
45.6
0.2
47.5
0.6
46.4
0.7
36.0
% CaO
21.4
1.2
22.4
0.5
21.7
0.4
22.4
0.5
23.8
0.1
25.4
0.3
22.9
0.6
18.0
% ZnO
14.5
0.5
14.9
0.3
14.7
0.3
14.2
0.3
13.3
0.1
15.4
0.2
14.5
0.3
20.0
% PbO
21.2
0.7
20.8
0.5
20.3
0.4
18.4
0.3
16.7
0.1
19.5
0.2
19.5
0.7
26.0
Q

 S  is one standard deviation of the average value of each set ( i.e., S//n)
  A


 where S is the standard deviation of a single measurement.





 S   is one standard deviation of the overall average of all the sets (i.e.,



 n = 31) .
                                      24

-------
    ,<*•'*•
    f.
I

  Figure 8. SEM photo of target B film: Mass loading = 50 pg/cm2.

-------
Figure 9.  SEM photo of target B  film:  Mass loadings = 160 ug/cm2.
                                 26

-------
more information at the higher mass loadings three additional sets were pre-
pared where the sputtering time was about 5 1/2 hours each with identical
sputtering parameters.  The mass loading varied from 167 to 185 yg/cm2.
These results are tabulated in TABLE 10.  It can be seen from the data that
the within set precision is again comparable to that found for set 5 in TABLE
9, and corresponds to a R.S.D. of the" mean of 0.7% for Si(>2, 0.8% for CaO, 2%
for ZnO, and 1.3% for PbO.  The maximum variation between sets for each
element in TABLE 10 is on the order of 10% for Si02, 6% for CaO, 7% for ZnO,
and 7% for PbO.  The maximum variation between sets for each element in TABLE
9 is 12% for Si02, 19% for CaO, 16% for ZnO, and 27% for PbO.  Hence it can
be concluded that variations in sputtering parameters does not seem to affect
the Si02 composition but does introduce two to three times more variability
for Ca, Zn, and Pb.  In addition, good film reproducibility within sets was
obtained with glass mass loadings from 160-190 yg/cm2.

     In Figures 10 and 11 are EDXRF spectra of glass films produced from
targets C and D.  The mass loading of the glass films in Figures 10 and 11
are approximately 73 yg/cm2 and 122 yg/cm2, respectively.  These results
indicate that sulfur- and phosphorus-containing glass films can be fabricated
by this sputtering technique.  Such films, however, would require relatively
large x-ray absorption corrections for sulfur and phosphorus because of the
absorption by the glass film matrix.  These corrections, however, can in
principle be determined. For example, a film about 180 yg/cm2 the attenuation
for sulfur assuming the glass film composition is the same as target C would
be about 20%.  Because of the relatively large corrections involved for
sulfur and phosphorus, the targets employed for the fabrication of such films
should ideally contain a minimum number of matrix elements of low atomic
number.  Although the composition of target C is homogeneous, the large
amount of Na£0 present makes it water soluble.  Therefore, films produced
from it are moisture sensitive.  Current research efforts are 'being directed
toward fabricating a low atomic number glass composition containing sulfur
which is less moisture sensitive.  Target D, on the other hand, appears
suitable for fabricating moisture-resistant phosphorous-containing films.

8.  RADIAL COMPOSITION OF GLASS FILMS

     Another experiment was performed to characterize the radial film
composition of the glass films across the diameter of the new 15.3 cm substrate
holder designed to hold up to eight substrates in the MIM/TLA 15.  The same
geometry as shown in Figure 2 was used except a vertical 23 cm strip holder
was fastened to the substrate holder and mounted vertically and parallel with
respect to the target.  This stationary vertical strip held five 37 mm sub-
strates, three within the 15.3 cm diameter of the substrate holder and two
outside as shown in Figure 12.  The substrates were numbered consecutively as
shown in the figure.  Target B was sputtered for about five hours and the
resultant films were analyzed by EDXRF as described previously. The data
obtained are summarized in TABLE 11.
                                      27

-------
               TABLE 10.  COMPOSITION OF GLASS FILMS AT HIGHER MASS LOADINGS
                                         (Target B)
Set Film
No. No.
1 1
2
3
4
5
AV.
sb
X
2 1
2
3
4
AV.
S
X
3 1
2
3
4
AV.
S
X
Mass Loading
Glass Film
pg/cm2
180.4
184.0
184.2
181.0
179.6

""""*
179.4
178.0
185.4
178.8


167.1
170.0
182.6
173.2


ZSi02
43.1
42.1
41.5
43.6
43.4
42.7
0.4
45.4
45.0
44.9
44.7
45.0
0.2
47.0
47.4
46.2
47.2
47.0
0.3
% CaO
23.3
22.7
22.5
23.5
23.3
23.1
0.2
24.5
24.4
23.7
24.0
24.2
0.2
24.9
24.8
24.0
24.4
24.5
0.2
% ZnO
12.8
12.3
12.4
12.5
12.8
12.6
0.1
12.6
12.6
11.9
12.5
12.4
0.2
12.4
11.5
11.2
11.5
11.7
0.3
% PbO
16.6
16.1
16.0
16.7
16.7
16.4
0.2
15.7
15.7
15.0
15.5
15.5
0.2
15.4
15.2
15.2
15.1
15.2
0.1
Ion-acceleration voltage = 1000V; Ion-current = 90mA; Time = 5 1/2 hours.

Standard Deviation of the mean value i.e., (S/v^n) where S is the standard deviation of a single
measurement.

-------
NJ
    CO
    z

    8
                                                                                              Ti sec
                               Figure 10.  EDXRF spectrum of film from target  C.

-------
U)

o
      o
      o
      CO
                            Figure 11.
  2.0    keV    3.0             4.0

EDXRF spectrum of  film from target D.
                                                                                   4.5     5.0

-------
                         Strip Holder
                          (23 cm)
                          Substrate Holder
                             (15.3cm)
Figure 12.  Al strip holder.
          31

-------
         TABLE 11.   COMPOSITION OF GLASS FILMS DEPOSITED ON SUBSTRATES
                         HELD BY STATIONARY STRIP HOLDER
 Film #
    Glass
Mass Loading
   Ug/cm2
% Si02
% CaO
% ZnO
% PbO
1
2
3
4
5
Bulk Comp. , %
160.0
169.0
180.0
161.0
144.0

41.5
44.1
40.4
42.9
39.1
36.0
25.0
26.7
23.0
22.6
17.1
18.0
8.0
9.7
9.6
11.3
10.4
20.0
9.0
12.2
13.9
17.5
17.6
26.0
     These data suggest that within the 15.3 cm diameter substrate holder
(i.e., films 2, 3, and 4) the glass composition is not significantly different
for Si02, CaO, and ZnO, but for PbO there appears to be a concentration
gradient which is quite marked especially if one includes films one and five.
Rotation of the substrates, however, will tend to average out this effect.
                                       32

-------
9.  SUPPLEMENTARY STUDIES

A.  Thin Metal Films Prepared by Thermal Evaporation

     To assess the state-of-the-art for producing thin films of metals thermally
evaporated on polycarbonate substrates, a few selected thin films were purchased
from a commercial supplier and characterized.  These consisted of Ti, Fe, Pb
and Zn, each at three mass loadings ranging from 50 to 100 ug/cm2.  We have
characterized these films by independently measuring the mass loading at NBS by
gravimetry and by x-ray fluorescence spectrometry.

1.  Mass Loading Measurements
     A number of filters preweighed at NBS were sent to Micromatter Co. for
coating with high purity Ti, Fe, Pb, and Zn.  These metals were chosen because
they cover a large range in atomic number.  Three films of each metal from 50
to 180 pg/cm2 were deposited on these  substrates by thermal evaporation.  A
film of gold was also deposited on a preweighed filter to serve as a control.
The mass loading of each metal was determined at Micromatter by weighing.
The filters were mounted on removable  lucite rings and returned to NBS.  A
comparison of the mass  loadings as determined by Micromatter and  NBS are
presented in TABLE  12.  As seen from the table, the best agreement between
the two  laboratories was obtained with the  Ti films.  With both the  lead and
zinc  films, the NBS results were systematically higher.  This  is  probably due
to oxidation  of the metal  films after  removal from the evaporation chamber
which continues for a  considerable  time after leaving the Micromatter  Co.
This  points out the difficulty  in  obtaining independent  accurate  mass  loading
'measurements  to better than  five percent by gravimetric  means  which  is the
 simplest and  least  expensive method.   It can be  seen  from  the  data that in
 order to obtain an  accuracy  of  better  than  five  percent  in the mass  loading
by gravimetric  methods, additional precautions must be  taken.   For easily
 oxidizable  metals,  the weighing should preferably be  done  in situ or alter-
 natively by protection of  the evaporated film  in an inert  environment prior
 to and during the weighing operation.   Some laboratories have  balances
 incorporated  in vacuum systems  but these are few, and it is a  relatively
 expensive proposition for  commercial suppliers  to build such systems.
                                        33

-------
          TABLE  12.   COMPARISON  OF MASS  LOADINGS  OF VARIOUS METAL FILMS
                    ON POLYCARBONATE  SUBSTRATES BY GRAVIMETRY

Sample
Blank Filter
Au Deposit
11 Pb
#2 Pb
#3 Pb
#1 Fe
#2 Fe
#3 Fe
#1 Ti
#2 Ti
#3 Ti
#1 Zn
n Zn
#3 Zn
Micromatter
vg
7,882
676
458
701
375
377
968
724
491
407
732
248
906
614
NBS
yg
7,878
693
505
753
405
352
1,003
691
510
402
730
280
930
651
Percent
Differenc0
- 0.05
+ 2.5
4- 9.0
4- 7.0
4- 8.0
- 6.6
+ 3.5
- 5.0
+ 4.0
- 1.3
- 0.3
4-12.0
4- 3.0
4- 6.0
2.  X-Ray Fluorescence Measurements
     The above films were mounted between two flat lucite rings and measured
with EDXRF.  The x-ray fluorescence measurements did not reveal appreciable
amounts of any elements in the films other than the element of interest.  The
net intensities of the Ti Rot, Fe Ka, Zn Ka, and Pb La were measured.  In
TABLE 13 is a summary of the net intensity of each film divided by the mass
loading (Micromatter values) in ug/cm2.  Since the x-ray fluorescence inten-
sity is directly proportional to the mass loading for these thin films, the
ratio of intensity to mass loading should be a constant. The estimated
standard deviations of each set were also calculated and are included in the
table.

From TABLE 13, the variability of the Ti, Fe, and Pb films are comparable, but
again the largest variation was noted with the zinc films.  The conclusion of
this limited study is that there may exist some serious problems in obtaining
accurate mass loadings of some thermally deposited metal films.
                                      34

-------
TABLE 13.  RATIO OF INTENSITY  TO MASS  LOADING  (I/M)  FOR  Zn, Pb,  Fe, AND Ti  FILMS
                (s =  standard  deviation  of  a single  measurement)
Sample
#1 Ti
#2 Ti
#3 Ti


//I Fe
#2 Fe
#3 Fe


//I Pb
n Pb
//3 Pb


#1 Zn
#2 Zn
#3 Zn

I/M (counts/yg/cm2)
159
156
149
Average = 155
s = 5.1 (3.
420
399
414
Average = 411
s = 10.8 (2.
37.2
39.9
38.6
Average = 38 . 6
s = 1.4 (3.
884
722
901
Average = 836





37,)




6%)




5%)




s = 99 (12%)
                                       35

-------
B.  Thin Aluminum Foils

     Some materials which are already commercially available can serve as thin
calibration standards provided that their trace element composition can be
measured.  Four square sections (^5 cm2) from a commercial aluminum foil (6.3
ym, ^2 mg/cm2) have been examined for homogeneity by EDXRF.  The elements Fe
and Ga were examined at approximately the 6 yg/cm2 (3000 ppm) and 0.1 yg/cm2
(50 ppm) levels, respectively.  The estimates of relative standard deviations
of a single measurement for Al, Fe, and Ga based on four sections of foil
were less than 3.0 percent.  Although this material seems homogeneous, it
does appear that thin Al foils would find few applications as a. calibration
standard because of the limited choice of elemental constituents.

C.  NBS-SRM 1648 Urban Particulate Deposited on Membrane Filters

     We have attempted to deposit NBS-SRM 1648 Urban Particulate on membrane
filters to serve as a potential thin calibration standard using essentially the
same procedure developed for depositing reground orchard leaves (3).  We found,
however, that many of the particles tended to clump together rather than dis-
perse uniformly on the filters.  For this reason, this work was discontinued.
It appears that materials of this type may be deposited more uniformly with an
air filter facility as described by Olson (10), but this may require reanalysis
of the deposited material.

P.  Synthetic Ground Glasses Deposited on Filters

     Another type of potential synthetic particulate standard was investigated
and consisted of synthetic ground glass of known particle size deposited on mem-
brane filters.  We have investigated three types of grinders in order to reduce
the particle size below 2 ym.  These consisted of grinding in (a) an alumina
ball mill, (b) a "puck and ring" commercial mill, and (c) a commercial automated
mortar and pestle apparatus.  A sedimentation procedure was used to separate the
fine particles from coarser ones greater than 2 ym.  Although this resulted in
the separation of a large fraction of particles below about 5 ym, it is difficult
to obtain a size distribution where 95-100% of the particles are below 2 ym.
One major problem with grinding is that elemental contamination can easily be
introduced although this would be minimized using boron carbide grinding mate-
rials.

     Results obtained thus far on the ground synthetic classes demonstrate that
such a particulate thin standard is feasible.  Additional work is required to
obtain a sufficient supply of glass particles small enough for deposition on
membrane filters.  The size, of course, is critical in order to minimize
particle-size x-ray absorption effects.  In addition, to obtain sample integrity
a suitable adhesive would be necessary to keep the glass deposit fixed to the
substrate.  Binding of the deposit to the filter with a collodion film layer as
described by Semmlar et al. (11)  may be appropriate and will require further
investigation.

10.  CONCLUSION

     From evaluation of the results obtained multi-element thin standards
suitable for calibration purposes can be fabricated by the focused ion-beam

                                      36

-------
sputtering technique.  In order to include all the elements of interest in
any particular film, several glass targets would need to be fabricated.  Because
there exists a difference in the film composition compared to the bulk target,
the elemental composition of these films can be measured and certified by NBS
using x-ray spectrometric and other instrumental analysis techniques.
                                      37

-------
                                  REFERENCES


 1.   Columbia Scientific Industries, P. 0. Box 6190, Austin,  Texas  78762.

 2.   Micromatter Co., 197-34th Avenue, East Seattle, WA 98112.

 3.   Bella, P. A., E. C. Kuehner, W. A. Cassatt.  Development of  Thin
     Calibration Standards for X-Ray Fluorescence Analysis.   EPA  Report  No.
     600/2-76-126, 1976.

 4.   Pradzynski, A. H. and J. R. Rhodes.  Development of Synthetic  Standard
     Samples for Trace Analysis of Air Particulates.  American Society for
     Testing and Materials Publication 598:320-336, 1976.

 5.   Camp, D. C. , A. L. VanLehn, J. R. Rhodes, and A. H. Pradzynski.
     Intercomparison of Trace Element Determinations in Simulated and Real Air
     Particulate Samples.  X-Ray Spectrom., 4:123-137, 1975.

 6.   Baum, R. M. , W. F. Gutnecht, R. D. Willis, and R. L.  Walter.  Preparation
     of Standard Targets for X-Ray Analysis.  Anal. Chem., 47:1727-1728, 1975.

 7.   Stiles, A. R, , T. G. Dzubay, R. M. Baum, R. L. Walter,  R.  D. Willis,
     L. J. Moore, E. L. Garner, J. W. Gramlich, and L. A.  Machlan.   Calibration
     of an Energy Dispersive X-Ray Fluorescence Spectrometer.  In:   Advances in
     X-Ray Analysis, 19, R. W. Gould, C. S. Barrett, J. B. Newkirk, and
     C. 0. Rund, eds.  Kendall/Hunt, 1976.  pp. 473-486.

 8.   Giauque, R. D. , R. B. Garrett, and L. Y. Goda.  Calibration  of Energy
     Dispersive X-Ray Spectrometers for Analysis of Thin Environmental Samples.
     In:  X-Ray Fluorescence Methods for Analysis of Environmental  Samples,
     Chpt. 11. Ann Arbor Science, 1976.

 9.   Dzubay, T. G. and P. J. Lamothe.  Polymer Films as Calibration Standards
     for X-Ray Fluorescence Analysis.  In:  Advances in X-Ray Analysis,  20,
     H. F. McMurdie, C. S. Barrett, J. B. Newkirk, and C.  0.  Rund,  eds.   Plenum,
     1977.  pp. 411-421.

10.   Olson, K. W. and V. A. Fassel.  Simultaneous Multielement Preparation of
     Air Particulate Standard Reference Materials and Their Application  in the
     Inter-Calibration of Analytical Instrumentation.  In:  The 29th Pittsburgh
     Conference, Cleveland, Ohio, 1978.  Paper #648.
                                       38

-------
11.  Semmler, R. A, and R. G. Braftz.  Calibration Standards for X-Ray Spectro-
     meters Used for Pollution Sample Analysis.  EPA Report No. 600/2-78-197,
     1978.

12.  Costellano, R, N., M. R. Notis, and G. W. Simmons.  Vacuum, 27:109, 1977.

13.  Amano, J.  Thin Film Deposition Using Low Energy Ion Beams.  J. Vac. Sci.
     and Tech., 3(14);831-836, 1977.

14.  Amano, J.  J. Vac. Sci. and Tech., 1(15):118, 1978.

15.  Weissmantel, C.  Trends in Thin-Film Deposition Methods.  In:  Proceedings
     of 7th International Vacuum Congress, Vienna, Austria, 1977.

16.  Technics, Inc., 7950 Cluny Ct., Springfield, Va. 22153.

17.  Epoxy Technology, Inc., 65 Grove St., Watertown, Mass. 02172.
                                      39

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
1. REPORT NO.
                              2.
                                                           3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
  The Development of Potential  Thin Standards for
  Calibration of X-Ray  Fluorescence Spectrometry
             5. REPORT DATE
                June 1980
             6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
                                                           8. PERFORMING ORGANIZATION REPORT NO.
            P. A. Pella
9. PERFORMING ORGANIZATION NAME AND ADDRESS
      Center for Analytical  Chemistry
      National Bureau of  Standards
      Washington, DC  20234
             10. PROGRAM ELEMENT NO.

                625 BE	
             11. CONTRACT/GRANT NO.
                EPA-IAG-D8-E684
12. SPONSORING AGENCY NAME AND ADDRESS
      Environmental  Protection Agency
      Office of Research  and Development
      Office of Environmental  Engineering & Technology
      Washington. DC 20460                        	
             13. TYPE OF REPORT AND PERIOD COVERED
                  Final   	.
             14. SPONSORING AGENCY CODE
                EPA/ORD/17
15. SUPPLEMENTARY NOTES
16. ABSTRACT

       Thin films containing known  concentrations of metals are  important for the
   calibration of X-ray Fluorescence Spectrometry (XRF), especially  for the analysis
   of  collected airborne particulate matter.   A focused ion-beam  sputtering technique
   has been investigated as a candidate method for fabricating thin  glass films contain-
   ing known concentrations of metals on  polycarbonate substrates.   Glass targets were
   fabricated at NBS for these studies, and parameters such as ion-acceleration
   voltage and ion current were systematically varied to determine any changes in film
   composition.   It was found that rather severe changes in instrumental parameters do
   not affect the elemental composition of the films appreciably.  Up  to eight substrates
   were coated at one time and the compositional reproducibility  as  measured by XRF
   for Si,  Ca, Zn, and Pb for 13 samples  was  within five percent  relative standard
   deviation at mass loadings of glass from 160 - 190 yg/cm2.  Glass films containing
   phosphorous and sulfur were also  prepared  to demonstrate the feasibility of prepar-
   ing glass films containing such elements of low atomic number.  Additional studies
   consisted of the deposition of finely  ground synthetic glasses on membrane filters,
   and the characterization of some  selected  commercial thin films prepared by thermal
   evaporation.
17.
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
                                              b.IDENTIFIERS/OPEN ENDED TERMS
                           c. COSATI Field/Group
      Thin  Films
      Calibration
      X-Ray Fluorescence Spectrometry
      Aerosol  Analysis
   Air Pollution Control
                                                                             7B
18. DISTRIBUTION STATEMENT
       Release to Public
19. SECURITY CLASS (ThisReport)
 Unclassified
21. NO. OF PAGES
      47
                                              20. SECURITY CLASS (Thispage)
                                               Unclassified
                           22. PRICE
                              $6.00
EPA Form 2220-1 (Rev. 4-77)   PREVIOUS EDITION is OBSOLETE

-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA-335
Official Business
Penalty for Private Use, S30O
                                                                                                             Special Fourth-Class Rate
                                                                                                                       Book
                                                                 Please make all necessary changes on the above label.
                                                                 detach or copy, and return to the address in the upper
                                                                 left-hand corner

                                                                 If you do not wish to receive these reports CHECK HERE a.
                                                                 detach, or copy this cover, and return to the address in the
                                                                 upper left hand corner
                                                                         EPA-60G/7-SO-123

-------