-------
     Of the  instrumental  monitors  shown  in Table 4.4, only  the  polaro-
graphic  (voltammetric)  analyzers are  designed to  respond to H2S  only.
These systems are  similar  in concept to the classical polarograph except
that solid  electrodes are  used  in place  of the dropping mercury elec-
trodes.    (To  distinguish  them from classical  polarography,  such  instru-
ments are often  referred  to  as "voltammetric"  instruments.)   In  this
instrument  the  gas  diffuses through  a selective  membrane into  a  cell
where it  is oxidized  at  a polarized electrode.  The electrical  current
is proportional  to the rate of diffusion  of H2S to the  electrode,  and
gas  concentration  is therefore proportional  to  current.   Specificity is
attained  through  selective scrubbing of  the  incoming gas,  by selective
diffusion into  the membrane,  and  by setting  the polarizing voltage low
enough to avoid oxidizing concomitant gases.

     Advantages of polarographic systems include simplicity and  ease of
operation.  In  some  cases the same instrument can  be used for more than
one  gas  by  simply  changing  the  electochemical   sensing cell.   Most
systems  are inexpensive  compared  to  other  monitoring systems  and are
often small enough for portable use.  No reagents are required, although
the  sensing cell must be replaced occasionally.

     Disadvantages  of the  polarographic  analyzers   include  the  gradual
deterioration  of  the sensing  cell and  the  associated  requirement for
frequent  calibration.  Materials  which condense onto or  clog the mem-
brane  accelerate  cell deterioration,  may  cause  erratic  results,  and
require  more  frequent cell replacement.  Cell  life under ambient condi-
tions may be  up to two years, although longevity upon exposure to retort
gases has not been reported.

     Of  course, the  polarographic  cell will  also  respond  to any other
gas  which is oxidized as  easily  as H2S and which  can diffuse into the
cell.   While commercially available  instruments have been designed to
reject   interfering   gases  in  ambient  air,   their  rejection   of  the
potential  interferences found  in  retort gas is yet to  be  established,
                                     189

-------
     For polarographic analyzers the gas typically must be conditioned to
100°F  or  less and 60 -  70% RH.   This is somewhat  of  a disadvantage for
retort gas, which would typically require cooling and drying.

     the  dynamic  range  of  the  commercially  available  polarographic
Analyzers  are  usually  more  suited  for  monitoring ambient  air  or for
'industrial  hygiene  applications.    This  fact  probably  reflects market
demands  rather than a technical limitation  of  the polarographic  system,
and  potential  users  are  encouraged  to  contact  the  manufacturer for
instruments  designed for the higher dynamic  range  associated with retort
gas.   Present users are employing dynamic dilution of the retort gas  in
order  to achieve  sufficiently low H2S  concentrations.

      In addition  to  the  polarographic  systems,  the  automated wet  chemical
systems  also are designed  to respond  directly  to H2S.   (See Table  4.4.)
Advantages  and disadvantages of these systems  include all those associ-
ated with the  manual  colorimetric  methods   (see above)  with few excep-
tions.   In  comparison  to  the  manual  colorimetric methods, a  continuous
measurement  is possible  and  labor  costs  are lower.   Nevertheless,  auto-
mated colorimetric methods are reported  to  require extensive maintenance
compared  to  other  instrumental  methods  and  still  demand a  continuous
 supply of reagents.

      Although not shown in Table 4.4,  argentometric coulometric titration
 systems for H2S  are  commercially available for laboratory use.   These
 presumably could be adapted  to field use.   Advantages would include long
 term  stability of  calibration and  low  maintenance  requirements.   Dis-
 advantages include the need to periodically change the scrubbing solution
 and  the  requirement  of  removing  other S   compounds  upstream  from  the
 detector.    Interferences  may be generated  by other  species which  react
 with Ag+ ion or can be readily oxidized.

      Two  instruments  in Table  4.4  operate  by converting  H2S  to S02 and
 then  monitoring  S02,  either  by  pulsed  fluorescence or  UV  absorption.
                                     190

-------
Specificity for the various  sulfur forms can supposedly be gained by the
proper selection  of chemical  filters  and converters.   For  example,  the
pulsed fluorescent  instrument (Thermo  Electron  Corp.)  operates  with an
in-line chemical  sorbent  to  remove S02 followed by a catalytic bed oper-
ated at 300  °C to convert H2S to  S02.   This instrument should therefore
be capable of three determinations:

     (1)  H2S, as discussed above

     (2)  S02, by shunting both the chemical  filter and  converter

     (3)  total S,  by  completely oxidizing the sample  stream

     Advantages  of this approach  include the absence  of reagent  require-
 ments  and-at least  with ambient  air and thoroughly-burned  stack  gas-
 relative  freedom from  interferences  in the  measurement of  S02.   It  must
 be emphasized that this  technique has not been tested  for  retort gases,
 and is  yet  to  be  confirmed for this  application.    Possible  problems
 include spectral interferences or quenching due to organic compounds, or
 altered behavior of  the  selective catalytic oxidizer or chemical filters
 upon exposure to retort gas.

      The  flame  photometric  detector  (FPD)  instruments  in Table  4.4
 (Meloy,  Bendix)  respond to  the  total  S content of  the sample  and gain
 specificity  through  the  use of in-line  sorbents.   In theory they should
 provide a method for  measuring total  S, H2S  and S02 through the judicious
 selection of chemical  filters.  A particular advantage  of the FPD is that
 reagents  are not  required  except for compressed gases.  (See Lucero et
 al.  [1977]  for a  review  of the mechanical,  electrical,  and pneumatic:
 design principles  of the FPD.)

       One shortcoming  of the FPD  instruments  is their  low dynamic  range
  (1-1000  ppb)  which  is  better suited  to  ambient concentrations.   Above
                                      191

-------

-------
this  range,  instrument  response  becomes non-linear  and levels off with
increasing  S levels.   Thus, extensive  sample  dilution  is often required
for FPD  monitors,  a distinct disadvantage for practical reasons, not the
least of which is supplying sufficient clean dilution gas.

.     The  response  of  the  FPD  is  also affected  by  gases other  than  S
compounds.   While  not  normally  problems with  ambient air,  C02 and H2 in
retort gas  present  potential  problems.   Reproducible response of the FPD
to S compounds requires a constant flow rate of H2 (manufacturer's data).
Since H2 is  a major component of retort gas, what effect do variations in
its concentration  have on FPD response?  Von Lehmden (1978) reported the
up  to 40%  suppression  of  the  FPD  response when C02  was present  at  a
thousand fold excess.   The  effect of C02 in retort  gas, which may reach
levels over  20% V/V, is yet to be established.

     Quenching of S emissions by hydrocarbons, another major component of
retort  gas,  has  also been  reported (Thompson,  undated).   While  this
potential problem  has  not  yet  been  investigated  for retort  gas,  Tracer
(1978) reports the suppression of the FPD sulfur response in natural gas.
Varian Associates  claims  to have  minimized  this  problem by  using two
flames in  series,  the  first  to combust  the sample and  the  second as  a
normal FPD  (Thompson,  undated).  Yet it should be emphasized that most
hydrocarbon  quenching  problems  arise  during the  attempt to  determine
trace levels of S  in  a  hydrocarbon matrix, and that retort  gases often
contain S gases as major components.  In addition,  potential users should
note  that  FPD's  have  been  reported with  rejection ratios  as high  as
10,000 for hydrocarbons (Natusch and Thorpe, 1973).

     Because the luminescence in  the FPD is due to the  species S2, theo-
retically the photomultiplier current should vary as  the square of the  S
concentration, (S).   In fact,  the response,  R, varies  as

(4.1)                      R = a+b(S)n
                                    192

-------
most  accurate work  thus  requires  knowledge  of the  chemical  forms of S
which  are present  in the gas  as well  as  the availability of standards
with similar molecular forms.

     Since the  theoretical  value of n in equation 4.1 is 2.0, some FPD's
come  equipped  with  a  "linearizer" which  electronically  completes  the
exponentiation.   As  pointed  out by  Burnett et  al.  (1977), the  use of
linearizers can  lead to errors up  to  400%  for common S gases since n is
not  necessarily  2.0.   These problems were reported for  a single-flame
FPD,, and may be less  severe for dual flame models.

     In  summary,  the FPD  as a  instrument  for measuring  total  S  or H2S
appears to present  several  difficulties. On the other hand, when used as
a  gas  chromatographic detector  most  of these problems  are alleviated.
This use of the FPD is discussed in Section 4.4.

4.1.3  Applications

     Owens and McDonald  (1979)  employed the EPA  Method  11 for measuring
H2S during the  in situ sideburn at site 12 at Rock Springs.  They recom-
mended  adding an  empty impinger  after  the  H202  impinger in  order to
prevent carryover,  since they  claimed that "prior  experience  had shown
that even minute amounts  of this  screening solution could drastically
affect the outcome of this test."

     Prien et al.  (1977)  analyzed  retort gas  from the  Paraho  retort
during  both  direct  and  indirect  operation.  Samples were  drawn through
two impingers  containing alkaline  solutions  of CdCl2 at 0°C.  A  third
impinger was  left empty  to condense excess water, and a fourth impinger
contained silica  gel.  In most  cases the analytical error in measuring
the collected sulfide was much less  than the total H2S concentration.   No
gross malfunction of the method,  such  as  the formation of unexplained
deposits or the collection of oily or tarry layers,  was  reported.
                                    193

-------
     One of  the best  confirmations  of a test method  under  field condi-
tions is the  achievement of mass closure after all streams are analyzed.
Such  a  confirmation  was  obtained  when  Goodfellow  and  Atwood  (1974)
analyzed H2S  in  the  product gases from a  Fisher  assay using gas chroma-
tography.  By also analyzing  the oil, spent shale, and product oil  for
total sulfur  they  were  able to achieve 95% closure for S.   Since the gas
included 20%  of the total  sulfur, major analytical errors  would likely
have produced poorer closure.   The  Fischer assay closely simulates  the
TOSCO II retorting process, so that these results suggest the efficacy of
the GC method for TOSCO II product gas.

     Fruchter et al,  (1979)  measured  S02 and  H2S at  the Paraho direct
mode retort using the InterScan polarographic (voltammetric) H2S monitor.
With this  instrument the  polarizing voltage was  changed  to distinguish
between  H2S  and S02.   However,  because the sensor when operated in  the
H2S  mode responds  almost  as  well  to mercaptans,  the results  for  H2S
included  mercaptans.    Other  interferences  listed  by the  manufacturer
include H2 and  unsaturated hydrocarbons,  although it is unlikely that H2
is present  in sufficient  amounts in  retort  gas   to interfere with this
technique.   The  manufacturer claims  that unsaturated hydrocarbons can be
removed by special sorbents.

     Fruchter  et  al.  (1979) also employed  a Del-Mar  continuous color-
imeter  to  monitor  H2S  from the  same source.   On subsequent  days  the
InterScan instrument  and the  Del-Mar  instrument  reported 2800 and 2400
ppm  H2S.  respectively.  Although no samples  were analyzed  in parallel,
these results are encouraging since plant operation was  supposedly uni-
form throughout.

     In  either  case,  no unusual  operating problems  were  reported.   How-
ever, the  retort  gases  did require dilution  in  order to be within the
dynamic  range  of either   instrument.   In  addition,  it  was  not  clear
whether  any cooling  or drying was required  upstream from  the analyzers.
These investigators also measured S in the  raw  and spent shale, product
                                    194

-------
oil, and wastewater, and  were able to obtain 93.5% closure.   Since 12.7%
of the S was partitioned into the gas phase and 80% into the  spent shale,
closure was  limited by the  accuracy of the S measurement in  the latter
two materials.  In  summary,  while the results of Fruchter's  measurements
at  the  Paraho  retort  are  encouraging,  they are  far from  conclusive
regarding  the  viability  of these  two instruments  for measuring  H2S.

     McDonald (1980), at the Laramie Energy Technology Center (LETC), has
been  using a suite  of  instruments  to  monitor H2S,  S02 and  total  S in
product gas from their 150 ton simulated in situ retort.  S02 is monitor-
ed with a Meloy model SA 160-2 and a Dupont model 411.  The Meloy instru-
ment  is  based on  a flame photometric  detector  and achieves specificity
through the use of selective sorbents to remove interfering gases such as
H2S.   The  Dupont  instrument is  based  on  non-dispersive  UV absorption.
H2S  is  monitored  by a Energetics series 7000  Ecolyzer, which  is based on
a  polarographic (voltammetric)  cell.   Total  S  is monitored  by  a Meloy
model  SA  202-2,  which is  also based  on  a  flame photometric detector.
Sample  pretreatment for  all  analyzers  includes  gas cooling and drying.

      The  principal  operational  difficulty reported  by  McDonald is  that
excessive  dilution  of  the  gas  is  required  to maintain concentrations
within  the dynamic range of  the instruments.  (With the exception of the
Dupont  instrument,  the remaining analyzers  are  mainly  for ambient appli-
cations.)   Stability of the  flame  in the Meloy instrument  is  occasionally
disrupted  by the  changing composition  of the retort gas.   In  addition,
negative  concentrations of H2S  have been  observed,  a result  not yet been
explained.
                                     195

-------
4.2  Sulfur Dioxide
                                                                                       •
     Although S02  is  apparently not a major form of S in retort gas, its              !
measurement may  nevertheless  be important for two reasons:   First, total              I
S or  H2S can be oxidized and  detected as S02.   This  interpretation is              j
attractive because of the abundance of commercially  available  S02 moni-
tors which  can   serve  as  the basis for instruments designed more speci-
fically  for  retort gas.   Second,  as a  gas which is not  removed effec-
tively by H2S control  equipment, S02 measurements may  be  important as a
diagnostic tool.

4.2.1  Methods

     The measurement  of S02  in ambient air and  in  thoroughly  combusted
gases  has  been   reviewed  extensively  and  need  not be  repeated here.   I
refer  the  reader  to Barras  (1973) for a review of instrumental  stack
monitors and to  LBL (1976) for a review of commercially available instru-
ments.   LBL  also  reviews the  principles  of  operation  of most  common
instrumental  and manual methods for S02.   APHA (1977)  and  Leithe (1970)
give details  of several  procedures for S02.   (See  Table  4.5  for addi-         [
tional references on methods for S02.)

     Table  4.6   summarizes  some of  the  most  common  methods   for  S02.         j
Because  of the  extensive  reviews which are available, these methods will         '
be  discussed only very briefly,  and with emphasis  on oil  shale.   The         i
reader  should note  that   these methods have  been tested  primarily for         ;
ambient  or  stack  (combusted) gases, and  their extrapolation  to  retort         i
gases  is questionable.    The  interferences  listed  in Table   4.6  thus
represent  potential  but   unconfirmed  problems.   For  example,  although
polarographic  analyzers  respond to  any  gases  which   are more  easily
oxidized than  S02, it is  not clear whether such gases  are a problem in
retort  gas  or whether  the present instrument  design successfully coun-
teracts  such potential   interferences.   On the  other  hand,  there  are
likely  interferences  in retort gases which have not yet been discovered,
simply for lack  of complete testing.
                                    196

-------
                TABLE 4.5.   REFERENCES FOR SULFUR DIOXIDE
Use of  tunable diode lasers

Absorption in H202 solution. EPA Method 6

Commercially available instruments

Collection in  H202 and in perborate

Practical field guide

Selection of  source  monitors

Details of  standardized  methods:
   colorimetric,  titrimetric,  conductimetric,
   amperometric

 Advanced methods

 Details of several manual procedures
                                                    Reid et al.  (1978)
PEDCO (1977)
LBL (1976)
AIHA (undated)
Bench!ey et al. (1974)
 Barras  (1973)
 APHA (1977)
 Stevens & Hoget (1974)
 Leithe (1970)
                                     197

-------
     TABLE 4.6.   OPERATING FEATURES OF

Method         	_j

(1) EPA Method 6            _
  :  S02(g)+H202OO^H20 + S04

  '  S04 by Thorin titration
     (Other methods for
       S04 can be used.)

(2) West-Gaeke Method

     S02 + K2HgCl4+Hg(S02)Cl2 + 2KC1

     pararosaniline +  CH20 + S02-»dye
 (3)  lodometric Method

      2H20+S02 + Is^3r+H2S
        Measure change In I3 (starch)
        absorbance

 (4)  Fluorescence of S02-dye
      complex

 (5)  Conductimetric analyzers
      with selective, in-line,
      chemical filters
 (6) Coulometric (Amperometric)
      Analyzers of various design
      e.g. 2Br~-»Br2 + 2e
      2H20 + Br2 + S02-»H2S04+2HBr


 (7) Polarographic (voltammetric)
      analyzers

      chemical oxidation at  a
      "polarized" electrode
METHODS FOR MEASURING SULFUR DIOXIDE

Comments  	•

Range : 3- 90,000 mg/m3
Interferences: NH3, sulfides, soluble
  sulfates, Wackenroder's
  reactions.
Application: stack monitoring


EPA  reference method for ambient air.

Low  bias  in results

Range:  0.01-^3 ppm
Optimized for ambient  air
Requires  skilled operator
Tempermental
Interferences: NOX
Time consuming
 Interferences:  oxidants and
                reductants
 Highly sensitive
 For ambient or pristine atmospheres

 Range: 0.01-8,000 ppm
 Interferences: any species
   forming or removing ions
   (e.g. NH3, H2S, organic acids)

 Stable calibration
 Minimal maintenance
 Interference: oxidizing and
   reducing agents
  See  Section 4.1.2
  Simple  operation
  No reagent requirements
  Selectivity depends  on
   Chemical filters,  membrane
   selectivity,  and selection
   of the polarizing  voltage
                                        198

-------
TABLE 4.6  (continued)
Method
                                       Comments
 (8) Flame Photometry
     emission of the S2
     species at 394 nm
 (9)  Non-dispersive infrared
      spectrometers
 (10) Non-dispersive ultraviolet
      spectrometry
 (11) Dispersive infrared
      spectrometers
 (12) Second derivative UV
      spectrometers

 (13) Ultraviolet fluorescence

      Fluorescence at  340 or 415  nm
Interferences:  compounds
  more easily oxidized than
  S02
Most are designed for ambient air

See Section 4.1.2
Sensitive
Responds to all S compounds
Interferences: H2, C02, hydrocarbons,
  other S-containing compounds

No reagents
Interferences: other IR
  absorbing species
Range: > 10 ppm                    .
Restricted mainly to stack monitoring

For  stack monitoring
Interferences: Species  absorbing
   near 280  nm

 Range:  0.1-3000  ppm
 Capable of  multi-species  monitoring
 Interferences:  other IR absorbing
   species
 High calibration drift

 Multispecies capability
 Available as either source
   or ambient monitors
 Relatively free of interferences
   under normal ambi ent or
   stack conditions
 Interferences: compounds
   fluorescent at 540 or 415 mm
   fluorescent quenchers such as H20
                                        199

-------
     Method (1) in Table  4.6 is widely employed as the compliance method
for S02  stack emissions.   The  sampling train called for in  this method
consists of a  probe  followed by four bubblers  cooled  to  0°C.  The first
bubbler  contains  80%  isopropanol  to  remove S03,  the third and fourth
contain  3% H202  for collecting S02,  and the  fourth  bubbler  is empty.
Although the EPA  protocol  calls for determining the collected S02 (after
conversion to  S04~)  by  the Thorin  titration method, any  number of other
finishes,  ranging from acid/base  titrations  to ion chromatography,  are
technically feasible.
                                   '                                    •
                                              -     •
     One component  of retort gas  which  is  likely to  interfere  in this
method  is  NH3, which can  form (NH2)2S03  in  the isopropanol  bubbler.
Other sulfides, such  as H2S, can also be oxidized to sulfate in the H202
bubbler  to  give  a  positive  interference.    Of  course,  Wackenroder's
reaction (Section 3.1)  likewise may remove S02 in  the  isopropanal  solu-
tion or convert H2S to sulfur oxides in the H202 solution.

     The second method in Table 4.6 is  designed primarily  for  ambient
air, and even under  such conditions has  a reputation for being highly
sensitive  to  the  skill of the  operator.  Its  application  to retort gas
has no apparent benefit.

     The  remaining  comments  in Table  4.6  should be  self explanatory.
Based on the available data, no single method can be selected as superior
or uniquely suited  for retort gas, and  each  method has potential inter-
ferences.   Obviously, methods   which  require  no  reagents  and  minimum
sample pretreatment are preferred.   Especially because of the presence of
NH3 in  retort  gas,  exposure to liquid water in the sample train (e.g. in
a  cooler/dryer)  is  likely  to remove S02 as  (NH3)2S03.   Analyzers which
depend on  a non-specific  response, such as conductimetric analyzers, may
work well  under ambient  or stack  conditions, but  their applicability to
retort  gas is doubtful because of the  large number of potential inter-
ferences not found in other gas streams.
                                    200

-------
4.2.2  Applications

     Prien et  al.  (1977)  used the standard  EPA  method 6 for determining
S02 in both  direct and indirect process gas at  the Paraho retort.   They
concluded that the analytical  errors in measuring the collected S02 were
minor compared to  the concentrations they encountered.  Parallel samples
were not  collected so that  no statement can be made  regarding the pre-
cision of  the method.  During continuous  operation of  the  plant  over a
four day  period,  the  measured S02 values varied from 14-23 ppmv,  which
can either be  attributed  to actual variations in S02 or to random errors
in the method.  Interferences by NH3 were not addressed.

     Fruchter et al. (1979) measured S02 using an interscan polarographic
H2S monitor  in  which they  varied the  retarding  potential  in  order  to
distinguish H2S  and S02.   With and without  prior dilution  of the sample
stream they achieved similar results.

     Owen and  McDonald (1979)  measured the S02  in  the  stack gases pro-
duced by the  in  situ  retort at Rock Springs site 12.  In order to remove
the interference  due  to  NH3,  a bubbler containing 0.1  M HC1  was  placed
upstream from  the  normal   EPA  method 6  sampling train.  No operational
problems, such as the formation of oily layers or dark precipitates, were
reported.  Because  no  samples  were collected in  parallel,  no information
is  available  on  the  precision of  this method  under  field conditions.

     The Laramie Energy Technology Center  (McDonald, 1979) uses a DuPont
model  411  S02 analyzer and  a  Me Toy model  SA 160-2 to  determine  S02  in
retort gases from  their simulated  in situ retorts.  The DuPont analyzer
is  based on  non-dispersive UV absorption  and  the  Meloy  uses a  flame
photometric  detector  with  appropriate  chemical  filters.    The compara-
bility of these two instruments has not yet been  reported.

     Other  developers  are  currently   using gas  chromatography  with
S-selective detectors for measuring  S02 as well  as other S gases.   These
are discussed in section 4.4.
                                    201

-------
                TABLE 4.7.   AMMONIA AND OTHER N-CONTAINING GASES
General Methods

Manual, standardized methods for
  ambient and industrial atmospheres


Pyridene-pyrazolene colon"metric
  method.  Collection with H2S04-impregnated
  filters

Collection on Ag/Mn impregnated
  filters

Automatic, continuous monitor based
  on  pH  electrode  and aqueous scrubber

Direct UV  absorption at 201  nm

Direct UV  absorption

UV  absorption  with a diode  laser
   source

Monitoring over long pathlengths by
   absorption of C02 laser radiation

 Microwave absorption

 Potentiometric detector for gases

 Conversion to NO with chemiluminescence
   finish

 Conversion of NH? to NO on Pt catalyst;
   IR or chemiluminescence finish

 NH emissions from a N2/Ar flame

 Gas  chromatography with a Hg-sensitized
    luminescence  detector

 Gas  chromatography of  furnace
    gases

 Gas  chromatography with a
    thermal conductivity detector
APHA (1977)
Leithe (1977)
NIOSH (1977)

Okita & Kanamori
(1971)
Eguchi & Hirofumi
(1977)

Steczkowski (1979)
 Cresser  (1977)

 Gunther  et al.  (1956)

 Reid et  al.  (1978)


 Artemov  (1977)


 Hrubresh (1977)

 Nicholas (1973)

 Aneja et al. (1978)


 Herose et.aV. (1979)


 Butcher & Kirhfught (1978)

 Harker  (1975)


 Derge (1965)


 Sims (1974)
                                        203

-------
TABLE 4.7 (continued)

GC detector for NH3
NH3 at the Paraho Retort
The effect of retorting conditions on
  ammonia formation
Ammonia  in a true  in situ  retort

 N-Containing Gases
 NO  in combustion streams  by
   collection in H202/H2S04
   N02, NO, HCN in the atmosphere:
   a review of methods
 NO  in combustion streams:
   practical suggestions
 HCN by collection in NaOH solution
 N02 and NO: review  of commercial
    instruments
  HCN,  CS2  in the workplace
Poppet et al.  (1978)
Cotter et al.  (1978)
Maier et al. (1924)

Owen & McDonald (1979)
 PEDCO (1977)

 Leithe (1970)

 Benchley (1974)

 AIHA  (undated)
 LBL  (1977)

 NIOSH (1979)
                                         204

-------
      Steczhowski  ef.al.  (1975)  automated a  wet chemical  technique to
 provide  continuous  process  control.  Their device consisted simply of an
 aqueous  scrubber and  pH  meter.  Adaptation of this technique to retort
 gas  would presumably  require  consideration of other water soluble gases
 such as  C02,  S02, H2S, and  organic  acids.

'  ''    UV   absorption   appears   an  attractive—and   as  yet   untested-
 alternative for measuring  NH3  in  retort gas since sample  drying  or cool-
 ing  could  presumably  be  avoided  and the absorption  spectrum could be
 recorded directly.   NH3 absorbs  strongly Jn a  series  of  sharp bands
 between  188  and 224  nm (e= 10-4000 1  mole"  cm"  ),  which  should  be ideal
 for  measuring  NH3  in  the  concentration  range 10 ppmv to 1%  (Gunther et
 a!., 1956).  Of course, many  other compounds  also absorb  in this region,
 and the  applicability of this  technique for retort gas must yet be estab-
 lished.    The  sharp   absorption  spectrum  in   NH3   ranges is  especially
 valuable,  since this aids  in  the  rejection  of  broadband  background
 absorption by such techniques as second derivative spectroscopy.

      Table 4.7  lists  four  studies of the measurement of NH3 by direct UV
 absorption.  Gunther  et al.,  (1956)  used a  simple Beckman  DU  spectro-
 photometer tuned to 204.3  nm to measure NH3 in the gas phase in the range
 7-1000 ppmv.  Cresser (1977) similarly was able to quantitate NH3 evolved
 from soil  samples  in  the  range 0-500 ng.  Reid et al. (1978) and Artemov
 et  al.  (1977) both used laser  sources for detecting NH3 in ambient air by
 UV  absorption.

       Microwave  spectroscopy has been  proposed as a method for NH3 because
  of  the  easily resolved series of  lines  centered at 23 GHz, which are due
  to  an  inversion  oscillation.   Using the absorption  line  at 23.870 GHz
  Hrubresh (1977) designed a field  model  suitable  for monitoring NH3 in the
  industrial  workplace  (range = 0.1-100  ppm).   This  instrument required a
  semi-permeable membrane   to  preferentially  enrich  NH3  over  the   other
  components  in  ambient air.   (Microwave spectroscopy is discussed more
  fully in Section  4.5.2.)
                                      206

-------
     NH3  can  also  be determined  by chemiluminescence  after catalytic
conversion to  NO (Aneja et  .1..  1978).   The commercial availability of
such instructs, which  can  then  also measure NO and N02   has  increased
the  popularity of  this  approach,  at least  for combusted stack  gases.

 '~   Commercially available instruments currently cover the range 0.1 ppb
to  K,  more than  adequate for monitoring  retort gas.  Water  vapor and
C02   have been  reported  as interferences,  although  this has  not been
consistently observed (LBL, 1976).  As with other methods which have been
optimized for  criteria pollutants in ambient air,  little  data is «.,!-
able establishing  the validity  of  this method for retort gas.

      In  a  N,/H2  diffusion  flame  the  introduction  of NH3  generates a
 broadband  emission  at   336.0  nm,  presumably  due  to chemiluminescent
 emissions from  the  species  NH.   Bulcher and Kirkbright used thls  pheno-
 menon  to measure  NH3 so!ution  at levels down to 0.2 ng/ml, but the suit-
 ability  of the  technique is clearly not established for  complex samples
 such as  retort gas.

      NH3  in  an inert carrier gas has also been determined by Hg enhanced
  luminescence  (Marker, 1975), based on the broadband emission centered at
  345 nm which occurs  when a mixture  of Hg vapor and NH3 are ,rrad,ate din
  the UV region.  A similar phenomenon occurs  with Cd vapor.  TMs approach
  is  highly  subject  to  interferences due to hydrocarbons, amines,  alde-
  hydes,  and  water,  and would  therefore require  a clean  chromatograph,c
  separation prior to detection.

        The gas chromatographic separation of NH3 has been difficult because
  of the tendency  of this substance to disappear onto  column walls, pack-
  ings, and  tubing.   Nevertheless, several  workers have reported successes
  „„ various  substrates.  For example, Sims  (1974)  separated  NH3 ,n water
  samples on a specially treated  column of Carbowax 20H on Teflon  and was
  able  to quantitate  NH3  in  the range 0.1-6% v/v.  For NH3 in a mixture o
   ethylene and N2,  he used  phasepack Q packing neated with 8% w/w KOH and
                                       207

-------
4.3  Ammonia (and other N-containing gases)

     Ammonia is  important  in  retort gas because of  its  commercial  value
and its  deleterious  effect on certain types of  S  control  equipment.   In
addition,  the  presence  of ammonia and  other N  compounds in  fuel  gas
increases  formation  of NOX  upon  combustion.  Consequently,  currently
proposed  regulations on NOX  emissions take into account the  level  of N
compounds  (other than N2)  in the fuel gas.

4.3.1  Methods

     Table 4.7 summarizes a  number of references  which describe methods
for  measuring  gaseous  NH3.   (Methods  for aqueous NH3  are discussed in
Chapter  3.)  Many  of the methods described in  Table  4.7  are optimized  for
ambient  or near ambient levels of  NH3  with the  corresponding  emphasis on
low  detection  limits and preservation of sample  integrity.  However, when
the  sampling  and interference problems associated with  retort gas  are
properly accounted for, it is  likely that at  least  some of these methods
can   be   adapted.    Table  4.7   also  contains  references   for   other
 N-containing  compounds,   although  these  will  not  be  discussed further
 here.

      The  most  common manual methods begin by absorbing NH3  in a  dilute
 acid, typically 0.1 N H2S04, followed by a variety of possible finishes.
 For example,  the NIOSH (1977) calls for the  collection of NH3 in  0.1  N
 H2S04,  followed the potentiometric (i.e. ion selective electrode)  detec-
 tion  of the collected  NH3.   Assuming a  working  range  of  0.1-1000 mg/ml
 for the  ammonia electrode, a solution volume of 10 ml, a sample flow rate
 of  100  ml/min, and a sampling time of  100  minutes leads to a working
 range  of  0.1-1000  mg/m3  in the  gas  sample,  adequate for  retort gas,
 especially  since  the  absorbing solution  can be diluted.  Major  advantages
 of  the  ion  selective  electrode (ISE) are ease of  use and relative freedom
 from  interferences.  (The major interferences  with the ISE are volatile
 amines,  which  are insignificant compared to NH3 in  retort  gas.) However,

                                      202

-------
10% w/w Versamide  900.   Using a thermo-conductivity detector he was able
to  operate  in  the range  2-10% v/v  NH3,  although  better  detectability
should be feasible using modern thermo-conductivity detectors.

     For  ambient  air levels  of NH3, Marker  (1975) recommends  a Teflon
column packed  with chromosorb 104 treated with  THEED  (optional).  Derge
(1965)  also  discusses   the   application  of  gas  chromatography to  the
analyses of furnace atmospheres.

     While  the thermo-conductivity  detector  may be somewhat insensitive
as  a detector  for NH3   at the levels  expected  in retort gas  (detection
limit s  0.1% v/v), newer N-selective detectors  should provide more than
adequate  detectability.  For  example, electrolytic  conductivity cells can
detect  less than  a nanogram of N  and therefore  should be adequate to
detect NH3  in  the  ppm range.

4.3.2  Applications

      NH3  is a major component of  retort gas  and in some processes  may be
a commercially valuable by-product.  For  this reason  its production  as  a
function  of retorting conditions  has  been widely  studied (Maier,  et al.
1924;  Jones,  1976; Owen and  McDonald,  1979).   While the methods of  analy-
 sis are  not  normally described in  detail,  it is  usually safe to  assume
 that the manual,  wet chemical  methods have  been  applied,  especially  in
 the older studies.

      Cotter et al.  (1978),  however, did describe in somewhat more detail
 the  sampling  and analysis   procedures  which  they used  at  the  Paraho
 retort.   They employed four impingers in series  cooled in an ice bath.
 The first  typically  contained d.i.  H20, the  second 5% HC1, the third was
 empty to catch any spill-over, and the last impinger contained  silica gel
 to dry the gas stream.   Sampling rate was 200 ml/min.   They reported that
 the analytical error in measuring the  captured NH3  was much less than the
 total ammonia concentration.  No data was available  on the precision of
 the total  sampling and  analysis scheme,
                                     208

-------
the membrane in the  ammonia is easily fouled by  oils  and tars,  and this
feature remains a potential problem.

     Other  common methods  for measuring NH3  in  the  absorbing  solution
include  the nitrite  method, the indophenol method,  and Nessler's method
CAPHA,  1977;  Leithe, 1971:  Okita  and Kanamori,  1971).  In  the nitrite
method,  NH3 is oxidized to  N02  with HOC!,  and the  nitrite  is  then con-
verted  to  an azo dye and measured colorimetrically.  Known interferences
include  CHO,  N02, hydrolyzable  amino compounds  and alkali  earth salts.
The  range   of  this  method  is  approximately 14-220  ng/m3;  higher values
should  be  accessible by diluting the  sample collection  solution.

      In the indophenol method  NH3  is  converted to a  indophenol dye, which
is then measured by spectrophotometry.   Interferences include N02, S02,
CHO,  Fe, Ci, Mn, and Cu,  although several  of these can  be  masked  by the
addition of EDTA.  A typical  range  for  this method (with no dilution  of
the  absorbing  solution)  is 20-700  ng/m3.

      Nessler's reagent  has been widely  used  in  the past for the  deter-
 mination  of  collected  NH3 but  has now been  largely replaced.   (APHA,
 1977;  Leithe,  1971).   Interferences include  H2S,  and CHO.  The working
 range has been reported as 0.02 to 80 mg/m3.

      In place of  aqueous  solutions and  impingers,   solid  sorbents  are
 occasionally used for sample collection.  For example, Okita and Kanamori
 (1971)  selected  H2S02  impregnated  filters for  the collection  of NH3  in
 the  urban atmosphere,  a  preference based on their observation that the
 collection efficiency  of  H2S04  bubblers gradually decreased  during the
 course of  sample  collection.   Eguchi   (1977)  also  describes  a passive
 sampling  device  consisting of a  filter  medium which  had been loaded with
 Ag  and Mn salts as well  as glycerin.   The  effectiveness  of the latter
 combination is  somewhat  doubtful  considering the  high  levels  of H2S in
 retort gas.  Advantages  ascribed to solid  samples  are  ease  of use  and
 higher sample collection  rate.
                                      205

-------
4.4  Gas Chromatography

     Gas  Chromatography  has been used perhaps  more extensively than any
other  technique for  analysis  of the various S-gases  in the atmosphere.
This situation  arises because of the availability of highly selective and
sensitive  S detectors which distinguish  S and non-S  containing gases.
The  advent of  highly inert materials,  such as  Teflon,  also permits the
handling of  gases  such as H2S, which otherwise would  react irreversibly
with  columns and  sample  lines.   The general  simplicity of  gas chroma-
tographs  as well  as  widespread  familiarity with this  technology also
encourages such application.

     Table 4.8  lists  references describing the gas chromatographic deter-
mination of S  gases  in  various environmental  samples.   As  can be seen,
most  references treat the  determination  of very low  levels  of S gases,
such  as  those  occurring  in the stratosphere  or in ambient air.  These
investigators  thus  emphasize  preconcentrating  the  sample,  improving
detectability,  and  avoiding  loss  of  the  minute  mass  of  collected
material.

4.4.1  Sampling

     Adapting  such methods  to retort  gas  presents  quite  the opposite
problem.    The  high levels  of  S compounds can  easily  exceed the dynamic
range  of  the   detector,  and  methods of  injecting  sufficiently  small
samples must be considered.   Because gases such as S02  can  easily dis-
solve  in  the basic condensate  produced  by  cooling the  gas,  retort gas
should be  injected directly into  the  GC with  no pretreatment.   As dis-
cussed below, a heated sample line and a  sample  loop should be sufficient
for most retort gases.

     When  retort gas  cannot be injected  into the GC,  it may be necessary
to use a sample container for transfer and  temporary  storage.   Such gas
sampling techniques described in the literature  include:
                                    209

-------
     o    cryogenic traps
     o    rigid glass or steel containers
     o    inflatable plastic bags
     o    sorbent cartridges

     Cryogenic trapping  has been widely used for  sampling  pristine air,
Such as is found in the stratosphere.  For example, Bramman et al.  (1978)
collected H2S and other S compounds in air cryogenically with gold-coated
glass  beads.   Using similar  cryogenic  techniques, Inn  (1979)  collected
COS and CS2 in ambient air.  Inn reported that COS was stable for several
days when kept in the cryostat.

     Farwell  (1979) used  a  cryogenic  U-tube filled with  polysiloxane-
coated glass  beads  at  liquid 02 temperature.  He  also  reported that the
anti-seizing  compounds found  on fittings are a strong sorbent for S com-
pounds and  must be  removed.   FEP Teflon loops and column  packing beads
were  also  evaluated as  trap  materials,  but exhibited  unsatisfactory
memory effects.   Each  of  these sampling techniques  was designed  to  be
compatible with a GC injection system for final analysis.

     In western oil  shale  areas the levels of S compounds in ambient air
are typically below the  detection limits of  most  commercially  available
instruments.   Cryogenic  enrichment  may therefore  be  appropriate  for
monitoring ambient  air in  these locations.   However,  this approach seems
particularly  inappropriate  for retort gases because of  the  large  volume
of water  vapor  contained in retort gas  and  because of  the need to avoid
exposing the gas to liquid water.

     Gangwal  et al.  (1979)  report  on  the  use  of glass containers  to
collect samples  from stacks and process  lines, and include  descriptions
of  valves,  connectors,  and  tubing.   Special  features  of  their  system
include the  ability  to  sample stacks  at  either sub-  or  super-ambient
pressures and the  capacity to store samples  at 50°C  until  analysis.   By
means of a special sample transfer system, samples were injected directly
from the glass buret into a GC without dilutions with  air.
                                    210

-------
       TABLE 4.8.   GAS CHROMATOGRAPHIC TECHNIQUES FOR S-CONTAINING  GASES
Gas_Chromat"Taphic Techniques
S gases In the atmosphere with FPD detector      APHA (1977)
Description of the FPD detector
Commerci ally  avai1able i nstruments
Element selective  detectors
   dure  sulfur compounds in air;
   Cryogenic trapping and capillary columns
 Gaseous sulfur compounds from fossil fuel
   conversion
 General discussion: instrument requirements
 CS2  in  air
 COS  in  the  stratosphere;  collection &
   analysis
 Various S  gases in the atmosphere
  H2S, COS,  CS2 & S02 in hydrocarbon streams
  COS and CS2 in the atmosphere
  H2S in air; preconcentration during  sampling
   H2S, COS,  S02, CH3SCH3 in air
   Electrochemical  detector for S02,  S03,  H2S,
     CH3SH,  and COS
   Electrochemical  detector sensor for S02, NO,
     CO,  and N02
   Column and FPD improvements
   Comparison of GC and  classical methods
   CS2 with  an  EC  detector
Lucero and Pal jug (1974)
LBL (1976)
Natusch & Thorpe (1973)
Farrell et al. (1979)
 Gangwal  et  al.  (1979)

 Thompson (undated)
 Godin (1979)
 Inn (1979)

 de Souza & Bhatia (1976)
 Pearson &  Mines  (1977)
 Sandal!s & Penkett  (1977)
  Braman et al.  (1978)
  Walker (1978)
  Chamber!and & Gauther
  (1977)
  Blurton & Stetter  (1978)

  Bruner et al.  (1976)
   Fradkin & Petrov (1978)
   Pinigina (1977)
                                        211

-------
TABLE 4.8 (continued)
Sulfur compounds and C02 in hydrocarbon
  streams with an electrolytic conductivity
  detector   .
Effect of S compound type on FPD sensitivity
Errors in using the FPD
H2S at trace levels
Trace sulfur compounds by standard
  additions
Interferences in the coulometric
  S detector
Sulfur compounds with a coulometric
  detector
S gases  in the air by GC/FPD
CS2 using a PID
Schiller & Bronsky (1977)

Doehler et al. (1977)
Burnett et al. (1977)
Steller et al. (1977)
Marcel in (1977)

Cedergren & Suden (1977)

Martin & Grant (1965)

Stevens & O'Keefe (1970)
Smith & Krause (1978)
 FPD =  flame photometric detector
 EC = electron  capture detector
 PID =  photoionization detector
                                     212

-------
     Gangwal et al. tested  the stability of COS, CH3SH, CS2,  C2H6S,  and
thiophene in coal  gas using this system.   Over a 100 hour period  changes
were  less  than  5%.   Similar  data for  H2S,  S02,  COS,  and  CH3SH  are
reported  for ambient air  samples,  although H2S  did decay signi^cantly
after 100 hours.

:    Plastic bags, typically  constructed  of  laminated materials such as
Teflon,  aluminum,  and Mylar are recommended  because of convenience and
ease of handling.   However,  several problems  remain, including the  con-
densation of water vapor and  concomitant  dissolved  species,.diffusion of
 sample through  the walls,  and reactions  with the  walls  (Krun et  al
 1979).   In  my opinion,  polymeric  bags should be considered unproven for
 sampling reactive or trace components  of retort gas.

      Porous polymers such  as XAD-2 and TENAX-GC have been widely used as
 sorbent traps, especially for organic compounds.  The principal advantage
 to  this  system is  ease of sample collection  and  shipment.   A number of
 investigators have examined the retention  of individual organic compounds
 by  porous  polymers;  however, the  behavior of  such polymers when exposed
 to  a gas  mixture is less clear (Pellizari  et al., 1975; Russell,  1975;
 Russell  et al, 1977).

       For example, Stephen and Smith (1977) have shown that porous polymer
  resin can  retain non-polar  compounds preferentially while allowing  polar
  compounds  to escape.   Although porous  polymers may be adequate for non-
  reactive  organic materials, they  would  not  be recommended  for reactive
  gases such as H2S,  SO*, and mercaptans.  Indeed,  a major problem in the
  gas  chromatographic  analysis of  such compounds is their reactivity with
  support  and column materials,  and  similar  reactions with sorbent  mate-
  rials are likely (Stevens and O'Keefe,  1970).  Thus, sorbent resins may
  be convenient  for  some  classes of  compounds  but do  not 'nec.ss.ri ly
  represent a universal, non-selective collection media.
                                       213

-------
     On the  other hand, the  use  of sorbent tubes for the  collection  of
specific compounds in  industrial  atmospheres  is common.   Thus,  Smith and
Krause  (1978)  discuss  the  use  of charcoal  tubes for the collection and
analysis of  CS2  in the industrial atmosphere, and Godin (1979)  discusses
the use of hexamethylphosphorotriamide and sodium azide on  chromosorb W
for the same purpose.    However,  in such  cases the sorbent  material  is
tested  for  a  specific  compound  and  is  not  necessarily  applicable  to
others.

4.4.2  Column Materials and Conditions

     Gas chromatographic separations of the major inert gases such as N2,
CO, C02 are  well  established and need not be discussed here.   Similarly,
molecular  weight profiling  of light  hydrocarbons is  one  of  the  first
applications  of GC,  and is  a  technique which should work  readily for
retort gases as well.

     Of special  interest in this  section will be separations of S gases,
principally  for detection by  a flame photometric  detector (FPD)  or  in
some  by other  S selective  detectors.   In reviewing  the  case  studies
described  below,  the following  guidelines which columns must meet should
be considered:

      (1)  S species must be separated from each other; H2S and
           COS  (bp  =   -61.8  and  48°C)  tend  to co-el ute  and
           therefore present a special problem.

      (2)  S species must be separated from major peaks of C02,
           CO,  or hydrocarbons to avoid interferences  in  the
           FPD.

      (3)  Special  valves,  tubing, sample  lines,  and  column
           materials must  be selected to  avoid sample losses.
                                    214

-------
      (4)  Sudden bursts  of H2 or  hydrocarbons  can extinguish
           the FPD--thereby requiring recall*bration—and should
           therefore be avoided.

     Besides the S  gases,  reactive  N gases such  as NO and NH3 have also
been separated  by gas  chromatography.   However, this  is  a  rather minor
application and will not be discussed extensively here.

     In one of the earlier descriptions of the gas chromatographic deter-
minations  of S  compounds with  specific  element detectors,  Martin  and
Grant  (1965)  used a  silicon  rubber column for measuring  S  in petroleum
products.  Although a S profile was  obtained up to C2o» only the lightest
compounds could be resolved.

     In 1970 Stevens and O'Keefe discussed the difficulties in developing
sufficiently inert columns, valving,  and a sampling system for measuring
low  ppb levels of  S02, H2,  CH3SH,  and CH3CH2SH  in ambient  air.   They
selected a 34'  x  0.085 id FEP Teflon column packed with polyphenyl ether
and  phosphoric acid   on  40/60 M  Teflon.  Temperature programming  was
apparently unnecessary.

     De Souza and Bhatia  (1976) describe a GC system for separating H2S,
COS, S02,  CH3SH,  and  (CH3)2S2 at ambient levels.   Their column of choice
was  3  m x 3.2  mm Teflon  packed  with pre-treated  Porapak QS.   Sampling
valves were  of  carpenter  20  stainless steel.   For  the separation of H2S
and COS  the  column  was kept at 140°C, and then programmed at 40°C/min to
230°C for heavier compounds.   S02  was not successfully determined because
of irreversible losses.

     Sandalls  and  Penkett  (1977)  measured  CS2  and COS  in  atmospheric
samples using three separate  columns.   COS was separated using a 50 cm x
4 mm id glass  column  of 80/100 M Porapak QS at 50°C.  CS2 was determined
on  two  separate columns:   a  2 m x 4 mm id glass  column  containing  10%
Triton X 305  +0.57 H3Po4 on 70/80  M acid washed diatomite M at 35°C and
                                    215

-------
(2)  a 1 m  x 4  mm id glass  column with  25%  1,2,3-Tris (2 cyanoethoxy)
propane  on  60/80 M  acid-washed  diatomite  C  at 40°C.   Also,  in  1977
Sheller and  Brovsky  described the use of  two  columns  for separating H2S
through butyl mercaptans  in petroleum products.  The first column was 3'
x 1/4" od  stainless  steel with acetone washed Porapak QS 80/100 M, which
was temperature programmed from 60°C to 160°C.   The second column was 36'
x 1/8"  Teflon containing  polyphenyl  ether and H3P04  on  40/60 M Teflon
powder operated isothermally at 80°C.

     Pearson  and  Nines  (1977)  developed  a  master scheme  for measuring
H2S,  COS,  CS2,  and S02  in a variety  a  samples  including specifically
inert  gases  and  water  vapor,  methane-ethylene  mixtures, and propane-
butadiene mixtures.  Quantitative analyses required the separation of the
S compounds from the hydrocarbon materials in order to avoid interference
in  the  flame photometric detector (FPD).  In  addition,  the hydrocarbons
would  often   extinguish  the flame  in  the FPD  which would then require
recalibration.

     For this range  of samples  Pearson and Hines  required four separate
columms, each made of  3 mm od stainless steel  which had been deactivated
with Siliclad 3%:
      o    3m,   with  5%  polyphenyl  ether  +  0.4%  H3P04  on
           chromosorb G,  80/100 M

      o    1.8 m with treated silica gel

      o    1.8 m with 5% silicone QFI-6500 on Porapak QS
      o    7.3  m with  10% polyphenyl  ether +  0.4% H3P04  on
           Chromosorb G
                                    216

-------
                                           ':	        "             "       i
     The  first  column is  programmed from 70-200°C and  serves  to quali-               j
tatively  screen  the sample  for compounds up  to diamyl  disulfide.   The
second column is most widely used and can separate H2S, COS, CS2, and S02               '
from each other  but not necessarily from  hydrocarbons ^ C3.   A combina-              ^
tion of columns  is  required to quantitate all  four  species in the three
matrices  studied.    This   paper amply  illustrates  the   difficulties  in
adequately  separating  hydrocarbons  from  the S  compounds  in hydrocarbon
matrices.    Of  particular  value was  a dual  FID/FPD  detector which  per-
mitted the  detection of  hydrocarbons as well as S-compounds.   With arti-
ficial gas mixtures Pearson and Nines claimed a precision of 3.3% for H2S
and 1.67% for COS.

     This system was primarily designed  for S  gases in  the  range  1-50
ppm.  Above this range  the FPD responded  poorly, while  below this range
stainless steel  columns  are not normally  used.  Since the S hydrocarbon               ;
ratio  in  retort gas  is  much  higher than  the  ratios   investigated  by
Pearson  and Mines, hydrocarbon  interferences  should  be less  severe.

     Walker (1978)  used aim x  1/4"  od FEP Teflon  column  packed  with
Tenax GC  to separate H2S, COS, S02,  CH3 SH, and (CH3)2S2  at  levels  of
1-100 ppm in air.   He claimed that  this  column  was  particularly easy to
pack and required no pretreatment.   Nevertheless, several injections  were
required before a reproducible response was achieved.

     In 1977  Stetter used a 6' x  1/8"  od FEP Teflon  column  packed  with
chromosil  310 for the  measurement  of H2S in air.  In addition,  he recom-
mended the  use  of  all  Teflon gas  sampling  valves and associated parts.
Inn et  al.  (1979)  also  used a chromosil 310 column  to separate COS  in
stratospheric samples, as  well  as  a Porapak QS in a 1/8" stainless steel
column  and  a  Porapak QS  with H3Po4  in  a Teflon  column for  the  same
purpose.

     Farwell et  al.  (1979) attacked the  problem of  hydrocarbon inter-
ferences  in  the FPD by  using a  higher   resolution,  wall  coated  open

                                    217

-------
tubular  column  (WCOT)  in place  of  the packed  columns  which have  been
described above.  They first evaluated a 30 m x 0.25 mm WCOT columns  with
SD-30,  carbowax 20M  and OV-17,  but  each  exhibited  excessive  tailing.
WCOT  glass  columns of .30-38  m length were finally  selected.  All  glass
parts  and  columns required  deactivation  in  order  to  avoid  sorptive
losses.  This  column  separated  H2S,  COS,  CH3SH, CH3SCH3,  CS2,  (CH3)2S2
and  other  compounds from each other as well as  from  hydrocarbon  inter-
ferences  in ambient  air samples.   Column temperatures were  programmed
from  -70 to 100°C, and about 15  minutes was  required  for each chromato-
gram.

     Harwell evaluated  this column  in comparison to  column  packings  of
Triton x-305, Teflon  powder with polyphenyl ether and H3P04,  deactivated
silica  gel, graphitized  carbon  black with  H3P04}  and acetone  treated
Porapak QS.  Advantages  he  claimed for the WCOT columns included smaller
surface area (hence fewer sorptive losses), better separation from hydro-
carbon  interferences,  lower pressure drops, and the better resolution of
H2S and COS.

      In  summary,  it is  not clear that a single  column will  suffice for
separating  both the minor and major sulfur species in retort gas as well
as  the  interfering hydrocarbons, carbon  monoxide, and carbon  dioxide.
Nevertheless, the  number of columns and the number of successful separa-
tions with  similar  samples  is quite encouraging.

4.4.3  Detectors

      The flame  photometric  detector (FPD) is clearly the most common type
of S-selective  GC  detectors.  This detector has already been discussed in
section 4.1.2 as a total S  monitor, and the same comments apply here.  By
way  of  review,  salient features  include:

      o    low dynamic range (10~l6 - 10"7 g/sec)
                                    218

-------
    o    freedom  from  reagent  requirements  except  for com-
         pressed  gases

    o    fast  response compared  to  most other  S  detectors

    o    instability of the flame upon exposure to combustible
         gases such as H2, CO, and hydrocarbons

     o    selectivity  ratios for  S compounds  (as  compared to
         hydrocarbons) of  up to 10,000

     o    suppression  of the  S signal  by C02 or  hydrocarbons

     o    dependence of S response upon molecular form


     The low dynamic range is best suited for ambient levels of * and for
 retort  gas.   Care  must  be titan  to  achieve sufficiently «T^-P £.
 The  freedom   from  reagent  requirements  is  especially  desirable  when
 operat  ng  under  field  conditions,  although compressed H2 cylinders may be
 pro  b  ted in some locations.   The  comparatively fast response per.it,
The use of WCOT  or capillary columns  whenever they are  needed to  separate
 interfering hydrocarbons from the S compounds.

      The  instability  of  the  flame  upon exposure to  combustible  gases
 ariselfr m  the requirement of operating the FPD with a fuel-rich flame
        v    ng on extinction.  The addition of small amounts of other fuel
 ga e   can therefore disrupt the  flame equilibrium, thereby changing it,
    ntivity  or  even extinguishing the  flame  altogether,  in wh.c  case  an
                                                                     «
      re                 -y  be  necessary.    Since  retort
  includes H2, CO, and hydrocarbons in amounts ranging  from 0.1 to  25%,  the
  effects of these major peaks must be considered.
                                      219

-------
     The selectivity of the FPD means that up to 10,000 parts of a hydro-
carbon gives  the same response as 1 part of S.  However, at  much lower
levels hydrocarbons  and C02  quench  the S  signal,  implying that hydro-
carbon and S  peaks  must be cleanly  separated  for  quantitative analyses.

:     The dependence of sensitivity upon molecular form requires that each
S  compound  be  calibrated  separately  for quantitative  work.   Varian
Associates has  attempted  to  minimize this  inconvenience  by placing two
flames in  series,  the first to combust  all  S  forms to common fragments,
and  the  second  to provide the  luminescent  signal.   This arrangement was
also designed to minimize  flame extinction and  hydrocarbon interferences.
However,  Gangwal  et  al.  (1979) report  that even with  the Varian dual -
flame  detector, the FPD remains measurably  sensitive to molecular form.

     Pearson  and Hines  approached  this same  problem by  using a FPD  in
combination  with  a flame ionization detector (FID).   This  arrangement
clearly  indicated the  presence of  co-eluting hydrocarbons, which could
then  be   removed   by   column   modification.    They   reported  that
re-calibration  of the FPD was  required  every  four  hours for quantitative
work.  With  a one ml  sample  loop, gases  containing  over 50 ppm of S  gases
required dilution in order to remain within the dynamic range  of the FPD.
They reported  a precision of  3.3%  RSD for H2S and 1.6% RSD for COS  in
 synthetic mixtures of light hydrocarbons (Ct - C4).

      In  addition to  the  FPD, other  GC detectors  which  operate on  the
 principles  of  polarography  (voltammetry),  electrolytic  conductivity,
 photo-ionization, and coulometry have been reported.  Thus Stetter (1976)
 and Blurten  et al.  (1978)  used  a  PTFE diffusion bonded  electrode  for a
 polarographic  GC  detector.   The  lower detectable limit was 5 ppb (HT12
 g with  a 1 ml sample  loop)  for  H2S, which is  approximately  the same as
 for the  FPD.   On the other  hand, H2S was  determined at levels up to 100
 ppm,  demonstrating a dynamic  range  somewhat  larger  than the FPD.   Like
 other  polarographic  detectors,  this  detector also  responds  to  other
 oxidizable gases  such as  NO, CO, S02,  and N02 and can be  used for their
                                     220

-------
                                          r;:

lence of use.

    However  several major disadvantages are also  apparent.   Response
time is sTow  causing peats to tail for approximately a minute, behavior
«   h  is un cceptable especially for WCOT  or capillary
^ tor response depends  on molecular for.,  so  that each gas »ust
 calibrated separately. Thus, it is not clear whether .ercaptans, COS,
 other stable forms of organic sulfur can be detected.
The e
          lectrolytic  conductivity detector (CD) operates
  wUh the CD is approximately 10-g for S,  poorer than the FPD but
  tainly adequate for retort gas.

      Advantages of the CD  over the FPD for retort gas include a higher
  dynamic rangl  a linear  response to S, and less  -ere interferences
  to hydrocarbons.  On  the other hand, the operate of the CD u  1  . s
  convenient than the FPD, requiring reagents, solvent pumps  nd c  c  a
  ting fluids, an ion exchange bed, and a gas/Hqmd scrubber   The t ,e
   . s onse is poorer than the FPD, and  the chronograms .^ ^ S
  exhibit tailing for approximately a  minute.  Ideally  >f a11 S  6

  rr^TK^*   --- -  - -- -
  Ibustion efficiency depends on the chemical form of S, and the response
                                  221

-------
of the CD  is  determined by furnace temperature, the  solvent and carrier
gas flow  rates,  the molecular form of  S,  and the flow rate  of  the  sol-
vent.   Each  chemical  form  of S thus requires  separate standardization.
Although  either  oxidation  or hydrogeneration  can be  completed  in  the
reactors,   Schiller  et  al.  found  that for  the analysis  of hydrocarbon
mixtures a stable baseline  could be achieved only in the oxidation mode.

     In summary,  the CD  appears to be  a promising  detector for retort
gases.  Although  the response  time is not  suitable  for  high resolution
columns,  its  superior freedom from hydrocarbon  interferences makes  this
less  of  a problem  than for  the FPD.   Its  higher dynamic range is  more
suitable than the FPD for retort gas.   The principal  disadvantage of the
CD  is  its rather complex circulating  and cleaning system and the asso-
ciated time  lag and  inconvenience of  operation,  especially under field
conditions.

     The photoionization  detector  (PID) operates by exposing the eluting
gases  to  a  UV  (e.g.  10.2  eV)  light  source and measuring  the induced
photoelectric current.  Therefore  they  detect any compound with a suffi-
ciently low  work function rather  than  specific  elements.   Advantages of
the PID include lack of fuel  requirements, a  linear range  of 106 compared
to  103 for the FPD,  ease of operation, and  low cost (Smith and Krause,
1978).   Disadvantages  include  the lack  of  selectivity  and compound-
dependent sensitivity.

     Smith and  Krause (1978) report on the  application of the PID to the
determination of  CS2  in the  industrial atmosphere.  These authors report
a  range  of 0.001-  50,000 ng for CS2.   However, the  PID  has clearly not
found  wide application  in pollution measurements.

     The  microcoulometric detector (MCD) has also found limited applica-
tion  for  the selective measurement of S  compounds  in complex mixtures.
These  systems  operate  by  combusting  or  hydrogenerating  the  effluent
stream,  scrubbing  interfering  gases,  and  titrating the  final  product
                                    222

-------
coulometrically.   (See Wallace et  al.  [1970] and Martin and Grant [1965]
for more  complete  descriptions.)  For S compounds  in  petroleum products
Martin and  Grant report  an operating range  of 10~8  -  1(T5  g.   Special
advantages of  the  MCD includes stable calibration  and response to all  S
compounds.   Disadvantages   include  complex   instrumentation,   reagent
requirements,  sensitivity to flow and reactor  conditions,  and poor time
resolution.

4.4.4  Applications

     Skogen  (1980)  has  regularly  used gas  chromatography  with flame
photometric  detection  (Hewlett  Packard  Model  5700/5710)  since 1976 to
monitor  gaseous  sulfur species  at  Occidental Oil  Shale,  Inc.  modified in
situ  retorts.  A sample splitter using ratios between  1:1 to  1:5 has  been
used  in  conjunction with a 1/8 ml  sample loop to monitor H2S  in the range
of 100  ppm  to 3000 ppm.   Both  polyphenylether and a 6' x 1/8"  diameter
Teflon column packed with  5% QF-1 on  80-100  mesh Poropak  QS  have  been
 used as  columns.   The  latter is preferred due  to better peak separation
 and less tailing.   Temperature  programming starts at 65°C  and increases
 to 140°C at a rate of 8° per minute.

      Skogen reports that this system is operated  in a mine and has proven
 reliable  under  severe   environmental   conditions.    These  conditions
 included shocks from blasting,  dust, and control  room temperatures up to
 100°F.   Water and oil  are  removed  from the sample by traps  and filters.

      A   operational  difficulty  is  finding  sample loops  small  enough to
 accomodate  the  low dymanic  range  of the flame  photometric detector.  The
 analyses of direct process  gases,  which are typically at  least ten  times
 more  concentrated in H2S,  may therefore require a less sensitive detector
 or a smaller  sample injection system.

       At the  Oil  Shale Symposium:  Sampling, Analysis, and Quality Assur-
  ance (March, 1979), Gangwal reported  on a sampling  and  analysis scheme
                                      223

-------
for measuring  H2S, COS,  S02,  CH3SH, C2H6S,  CS2, and  thiophene using a
Porapak N column  and  a dual flame  FPD.   With coal gas, H2S and COS were
only partially separated.  Even using the dual flame detector, most gases
required separate  calibration,  although H2S,  COS, and S02 collapsed onto
a single calibration curve when plotted against A/Vh~, where A is the peak
area and his the peak height.

     Gangwal's data clearly  indicated the virtue  of the  dual  flame FPD.
With  the  single  flame detector  the response to CH3SH,  C2H6S,  and CS2
varied by  75% due  to hydrocarbons  co-eluting  from previous injections.

     In summary,  field experience  has  provided  data  on  the reliability
and operational problems  associated with several manual and instrumental
techniques.   However, data  on  the accuracy and  precision  of the methods
are not  available.  Establishment  of these  parameters will require the
analysis of  known  samples and the  simultaneous  analysis  of the same gas
stream by several techniques.
                                    224

-------
4.5   Additional Spectral Techniques

     Spectral techniques are  attractive for the analysis of retort gases
for two primary reasons.   First, in situ measurements are at least theo-
retically  possible,  thereby   avoiding  sample  conditioning  (and  alter-
ation).   Second,  several gas  components may  be measured simultaneously.
This could yield dividends in the form of better diagnostics and controls
for both the retorting and pollution control operations.

     Spectral techniques are  obviously widely used  for monitoring cri-
teria pollutants such as S02 and N0"x, but most commercial instruments for
this  purpose are  limited  to  a single  species  (e.g.  UV fluorescence for
S02).  Similarly,  techniques such as 1R spectrometry are obviously widely
used  in  the laboratory for in situ analysis of multi-component mixtures.
However,  the application of such approaches to the analyses of retort gas
is not widespread  and is essentially absent from the literature.

     This  section  will therefore briefly review the possibility of apply-
ing  techniques  such  as  microwave and  UV  spectroscopy  to  the multi-
component analysis of retort  gases.  The  literature  on any one of  these
techniques is  large,  and  I make no effort to review the entire subject
here  but attempt  only to  highlight those  features especially relevant  to
the analyses of retort gas.

4.5.1  Second Derivative Spectroscopy

      Two principal types of second  derivative spectrometers  are described
 in the literature: those which scan the absorption spectrum with  time and
 then  calculate  the various derivatives, and those which obtain the  second
 derivative  directly  as an electronic  signal  from the  photomultiplier.
 The latter  spectrometers  operate  by  modulating the  incident wavelength
 sinusoidally about the  absorption  peak maxima, and recording  the  second
 harmonic from the detector (Hager,  1973; Cahi11,  1979).
                                     225

-------
     In comparison to the more common spectrometers which measure optical
absorption,  optically modulated  second derivative  spectrometers  (OMDS)
often exhibit a signal-to-noise ratio which is improved several orders of
magnitude, and are  therefore  attractive for the detection  of  trace com-
ponents.  On the  other hand,  electronic derivative spectrometers produce
quite  the opposite  effect and  will  not  be  discussed further in this
report.

     The  OMDS produces  a signal  S which is  related to the concentration
by equation 4.2 (Hager, 1973).
             S     4      11
      c = - _ •  _— •_  • 	                         (4.2)
             I   (A\)2    Z    (d2a/dX2)
where     C = concentration of analyte

          S = the second harmonic signal

          I = the average intensity of the signal

          A, = wavelength

         AA. = the wavelength modulation amplitude

          £ = absorption path length

          a = absorbance of the analyte
      In  this  equation,  the first four terms are under the control of the
 operator and  instrument designer, and the last term is a physical charac-
 teristic of the molecule.
                                    226

-------
     Three features are  of particular interest in this equation.   First,
OMDS clearly favors the  detection of those species with sharp absorption
lines,  typical  of  simple  inorganic  molecules,  while  discriminating
against broadband absorption.   Thus,  OMDS provides a means of compensat-
ing for background due  to broad-band absorption and scattering.   Second,
unlike  normal  absorption  spectrometers where  concentration  is  related
logarithmically to intensity,  OMSD signal is related  linearly to inten-
sity.   This feature favors a wider dynamic range.  Third, the OMSD tech-
nique  is  relatively  insensitive  to  drift in the  light  source and asso-
ciated electronics,  since S and  I are both measured  at  the  same photo-
tube.    Unlike  other  techniques  designed to  compensate for  background
absorption, the OMSD spectrometer has only one light source,  one optical
path,  and one photo-detector.

     OMSD  spectrometers  normally  operate  in  the  UV  region in order  to
take advantage of the  higher absorption coefficients in this part of the
spectrum.   At  these wavelengths matrix gases such as N2, 02,  H2,  and H20
do not absorb,  so that sample modification is not required.   This feature
is particularly  attractive for  retort gases where removal of water may
inadvertantly  remove other soluble gases such as C02, NH3, H2S,  and S02.

     Aside  from  these  general comments,  I  am unaware  of any  studies
detailing  the  application of OMSD spectroscopy to retort gas.  Clearly,
the availability  of  interference-free  spectral windows  and sufficiently
sharp  spectral lines should be one of  the  first  areas of investigation.
Table  4.9  lists  interference-free detection  limits for  a few gases  as
measured by OMSD  spectroscopy.  While these  numbers  are  encouraging for
NH3, •  NO,   N02,  S02,  and Hg vapor,  the  feasibility of  measuring these
compounds  in retort gas must still be determined experimentally.
                                    227

-------
 TABLE 4.9.



   Compound


;      NO


      N02


      S02


      03


      NH3


 Benzene


 Toluene


 Xylene


 Styrene


 Formaldehyde


 Mercury Vapor
DETECTION LIMITS FOR VARIOUS GASES BY SECOND DERIVATIVE
SPECTROSCOPY (Hager, 1973)
                              ppb(v/v)
                                 40


                                  1


                                 40


                                  1


                                 25


                                 50


                                100


                                100


                                200


                                  0.5
                                    228

-------
4.5.2   Millimeter Wave Spectroscopy

     Microwave rotational spectroscopy can serve as a specific and sensi-
tive tool for monitoring gaseous constituents in the environment and also
for  controlling  gas treatment  technology (Hrubesh, 1973;  Kolbe et a!.,
1977; Leskovar and Kolbe, 1978).  The physical process leading to absorp-
tion is  the  coupling of incident microwave  radiation  with  the permanent
dipole moment of a rotating molecule.

     The  advantages  of microwave rotational  spectroscopy as a technique
for gas detection and measurement include the following:
           Spectra can  be  measured with  a  high  frequency  reso-
           lution of  one  part in 10s to 106.   Thus,  the  mea-
           surement of a single line usually permits the deter-
           mination of a constituent in a mixture.

           A single  instrument can be used to monitor a number
           of  constituents  by  automatic   tuning of  the
           frequency.

           Recent progress  in mm wave technology permits the
           development of  portable,  all  solid-state,  spectro-
           scopic instruments.
The peak  value  of the gas absorption coefficient  can  be approximated by
                               p = Av2                            (4.3)
where A  is  a quantity which depends on such gas parameters as concentra-
tion, pressure,  temperature, rotational transition mode  and  fraction of
the molecules in the transition state of interest.  The frequency of peak
absorption is v (Gordy and Cook, 1970).
                                    229

-------
     Since the  absorption  coefficients  of typical molecules are small (y
= 7 x 106 cm'1 for S02  at 20 GHz) and are also proportional to v2, it is
important to  select rotational  transition  modes at  frequencies  as high
as  possible.    At  higher  frequencies  there  is  also  a larger  thermal
population difference, which helps to increase the quantity A in equation
(4.3) and thus the absorption coefficient.

     Because of the small  absorption coefficients and because saturation
occurs  at relatively  small  drive  powers  P (=lw),  a  highly  sensitive
detector  is required.   Up  to now the unavailability  of lockable,  solid-
state oscillators and sensitive detectors has restricted the operation of
such mm wave spectrometers to frequencies below about 80 GHz (Zoellner et
a!., 1979).  Recently,  work  has been initiated to  raise this upper fre-
quency  limit  to 145  GHz  using  specially designed  upconverters (Edrich,
1979) and improved mixing techniques with higher efficiencies (Carlson et
al.,  1978;  Edrich   et  al.,  1978;  Kerr,  1979;  Edrich,  1979).   These
improvements are expected to lead to (a) higher detection sensitivity and
selectivity, and  (b)  to coverage of more line  frequencies,  (more molec-
ular species with one single spectrometer).

     At the present time a major disadvantage  to microwave spectroscopy
is  the  lack of adequate commercially-avail able equipment.   However,  the
recent  advances  discussed in  the  previous  paragraph auger  well  for  the
wider application of this technique.

     A  spectrometer covering the two bands  69.5  ±3.5 GHz  and  23 ±  1.5
GHz could detect  several  important gas components  (Table 4.10)  based on
the  occurrence  of  spectral  lines  in  these regions  and could be con-
structed with technology and components.  Such components would include a
single  locked rf  source with improved noise characteristics and a high-Q
Fabry Perot microwave  resonator with improved multispectral coupling and
filter  characteristics.  This  instrument  could, for example, measure S02
concentrations down  to  1  ppm using a time constant  of 1 sec.  Table 4.10
lists a number  of pertinent molecules  and  their  transition frequencies,
which would be covered by such an instrument.
                                    230

-------
TABLE 4.10.  VARIOUS GASES DETECTABLE BY A DUAL-WAVELENGTH MILLIMETER
             WAVE SPECTROMETER
                                                      MHz
     Molecule                                     Frequency
     C12 H3 S32 H                                 21,973.2
     H N14 C12 O16                                21,981.7
        H2016                                     22,235.1
        N14 H3                                    23,098.8
         SO                                       66,039.9
     O16 O18 O18                                  22,527.0
     S32 0216                                     22,482.5
     N14 0216                                     70,589.7
     H2 C12 O16                                   72,409.1
     O16 C12 S32                                  72,976.8
                                   231

-------
4.5.3  Mass Spectrometry

     Although mass  spectrometry  is not an in situ  technique,  as are the
other  methods  discussed in  this section, it  is included  at  this point
because  of its ability  to determine multiple species in  a complex mix-
ture.  Because  the mass  spectrometer operates  under  vacuum,  the sample
stream not only must  be  diluted  but typically must  also be  dried and
scrubbed of corrosive gases in order to prevent damage to the high-vacuum
components.

     Advantages  of  mass  spectrometry  include  the  ability to  monitor
several  species  with a  high degree of specificity and  flexibility on a
continuous, real-time  basis by  monitoring a corresponding  set of mass-
to-charge  ratios.   The  list of monitored gases  can be easily  changed by
programming a  different  set  of peaks.   Of course,  this simple-minded
approach assumes the availability  of interference-free peaks,  an assump-
tion that has  not yet been verified for retort gas.

     One of the  principal  disadvantages of mass spectrometry is the lack
of  commercially available,  inexpensive,  and  sufficiently rugged  mass
spectrometers  for  field use.  Laboratory mass  spectrometers are usually
expensive  and  require  considerable skilled  maintenance.   High-vacuum
materials  are normally  incompatible with  corrosive gases and high levels
of  water  vapor.   The  availability  of  an inexpensive,  low maintenance,
rugged mass spectrometer would thus appear essential  before the applica-
tion of mass spectrometry to process streams  could become routine.
                                    232

-------
4.5.4  UV and IR Absorption

     UV and IR absorption is obviously widely used in both laboratory and
field  conditions  for  fn situ  analyses,  and  several  of the  stack gas
monitors shown  in Table 4.4 operate on the principal of UV or IR absorp-
tion.   However,  such  instruments  are typically  optimized for  a  single
species,  such as S02,  and cannot  serve as multiple-species  detectors.
Among  the  instruments  listed  in  Table 4.4,  only the WiIks  Scientific
instrument—which is actually designed for industrial hygiene purposes—
is tunable to different gases,  and it is  not  clear whether this instru-
ment can be used without modification for retort gas analysis.

     In my experience,  UV and IR absorption methods  have not been thor-
oughly tested  for the  in-line analysis of retort  gas,  and several ques-
tions  must  be  addressed prior  to their  widespread use.  The existing
monitors for criteria gases are primarily designed and tested for ambient
air or thoroughly combusted stack gases.   What additional  spectral  inter-
ferences occur  in retort gas?   To what degree  can absorptive  methods be
extended to  additional  species  such  as NH3,  CHN, mercaptans,  COS,  and
other  gases  of  potential  interest?   Although absorptive  techniques  in
theory operate  in  situ,  in practice  sample streams  are  often  dried and
cooled in  order to  protect optical components.   Does  such pretreatment
alter the composition of retort gas, or can instrument optics be modified
to tolerate untreated retort gas?
                                    233

-------
4.6  Mercury

     This section  treats  the measurement of trace metals (especially Hg)
in the  gas  phase while" trace metals in solids and liquids are treated in
Chapter 2.   Since  any number of techniques are available once the sample
is collected,  the  emphasis here is  on  the sampling technique unless the
monitor is  used  to detect Hg on-line in the gas phase.

4.6.1  Methods

      Table  4.11 summarizes  methods  which are available for  sampling  and
 analyzing mercury  vapor.   As can be seen, most of the methods are design-
 ed primarily for ambient air or the industrial workplace.   However,  these
 methods may often be adapted to waste streams through  the judicious use
 of  sorbents  or  combustors  and  scrubbers   upstream  from  the  mercury
 collector.

      The  sampling  techniques  in Table  4.11 can be  roughly categorized
 into groups:

        (1)   Aqueous  solutions  of stabilizing  reagents,  the most
             popular  being IC1,  HN03, and  acidic KMn04

        (2)   Solid  adsorbents, such  as  Mn02 and activated  carbon

        (3)   Au or Ag wool amalgamation cartridges


  In addition, the on-line Zeeman atomic absorption (ZAA) analyzer requires
  no  sample collection  but rather  detects  Hg vapor  directly in the gas
  phase.

        Driscoll  (1977) briefly  reviews  the major advantages and disadvan-
   tages of  the  common methods  of  sampling stationary sources for Hg,  and
   the reader  is referred to that  source for a more  detailed descnption.
                                       234

-------
         TABLE 4.11.   METHODS  FOR COLLECTING  AND MEASURING MERCURY VAPOR
 General  Methods         •

 Collection  on  Au/quartz wool;  hydrocarbon
  :streams and  air

 Collection  with  KMn04/H2S04  solution; air

 Collection  on  carbon-loaded  filter paper;
   industrial atmospheres

 Collection  on  Mn02;  industrial
   atmospheres

 Piezo-electric Hg monitor for
   industrial hygiene

 Collection  on  hopcalite; industrial
   atmospheres

 Review of sampling methods for
   stationary sources:  Absorption,
   amalgamation, and solid sorbents

 Collection  on  Ag wool; industrial
  environment

 Zeeman atomic  absorption; on line measurement
  for retort gas

 Molecular speciation through selective
  sorption

Application to Retort gas
 Ohkawa and Kondo  (1977)


 Gardner (1976); Kara  (1975)

 Janssen et al.  (1976)


 Janssen et al.  (1977)


 Scheide and Warnar  (1978)


 McCullen and Michaud  (1978)


 Driscoll (1977)



 Kneip  (1975)


 Girvin  et  al.  (1979)


 Braman  and  Johnson  (1978)
Hg in product gas from simulated in situ
  retort; comparison of Zeeman atomic
  absorption and Id absorption solutions

Hg speciation through selective sorption
  at the Paraho retort; total Hg

Hg evolution during Fisher assay
Fox et al. (1978)
Fox et al. (1977)
Fruchter et al. (1979)
Donnell and Shaw (1977)
Schendrikar & Faudel (1978)
                                     235

-------
     The  methods  which appear most commonly in the literature are the Au
amalgamation  technique and  ZAA  spectrometry,  and  these deserve further
discussion.   While  the former is widely practiced and commercial instru-
ments  are available  for  monitoring ambient air, the  latter is still in
the  prototype  stage  and  field-worthy models  are currently  being per-
fected.   Id  absorption solutions have also been evaluated with respect
to retort gas.

     The  principal advantages  of  the amalgamation  techniques  are  the
general ease  of use and the pre-concentration factor obtained by passing
a  large volume of gas through a  small  mass of metal.  However, Ag as an
amalgamation  metal  is probably contraindicated because  the ready forma-
tion of AgS  with H2S and  other S gases,  poisons  the surface and reduces
the  collection efficiency (Driscoll, 1977).   The principal advantage of
Au  is  thus  its  lower reactivity  and associated  greater  freedom  from
interferences.

     Most applications  of the Au amalgamation technique  have  been com-
pleted on thoroughly  combusted  gas streams, and adaptation to retort gas
may  require  a combustor  in-line prior  to the  amalgamator.   While  Au
should be inert with  regard to reactions with retort gases, condensation
of hydrocarbons or water may readily degrade  the collection efficiency.

     A  particular  advantage of  the  Au  amalgamation  technique is  its
excellent  detection  limits.   For  example,  suppose a nominal  1  CFM
(typical  of  EPA method  5  sampling train)  is  passed  through the amalga-
mator for one  hour (Kalb,  1975).   A typical detection limit for Hg using
the standard   flameless  atomic  absorption technique is conservatively 20
ng, thereby yielding  a detection  limit of  approximately 10 ng/m3 in the
original gas stream.  This figure compares with measurements of Hg in raw
and burned retort gas  which vary from 4 to 75 ug/m3.

     In addition to good  detectability,  other advantages of the Au amal-
gamation  system  include   freedom from   liquid  solutions  and  relative
                                    236

-------
freedom from  Interferences.   Ohkawa and Kendo (1977) report that samples
collected on Au wool are stable for 1 week at room temperature.                        t

     Disadvantages  of "amalgamation techniques  include the  possibility              .
that the collection efficiency may be degraded by unsuspected condensates
or  adsorbents,  and the  limitation of measuring only  elemental  Hg.   (An
upstream combustor  would also permit the measurement of total Hg.)  Real
time measurements are  not possible, and time resolution much better than
an  hour seems  unlikely.   In addition, if the sample is injected directly
from the Au  amalgamator  into the AA spectrometers, off-scale samples are              i
irretrievably lost  unless  sample recovery or splitting is built into the              •,
system.                                                                                j

     The design  and operation  of the ZAA  spectrometer is  explained  by              \
Girwin et al. (1979), Hadeishi (1972), and Miller and Koizumi (1979).  As              j
the  name  implies,  ZAA  spectrometry  is  unique  in using  Zeeman shifted
lines  to   account  for  background  correction.   Because  a single  light
source, optical path,  and detector are employed,  long  term stability  is
superior to  other AA spectrometers.  Low frequency noise  and changes  in
background absorption  are accounted  for  by rapidly  alternating between
shifted and non-shifted lines.                                                         j
                  '          '                                -                          !
                                   '    .                    .          -                  !
     The ZAA spectrometer  described by Girwin et al.  is being especially              j
designed and optimized  for  the  on-site monitoring  of retort  gas.   In              i
practice the sample is mixed with oxygen and burned prior to introduction              j
into the optical  absorption  cell in order to  minimize  interferences and              |
smoke  formation.  This technique therefore  records total mercury content              ;
                                                       "                                i
although gaseous  elemental  Hg might be tractable with  an  unheated cell.              |
When operated  in the  field,  the ZAA  spectrometer  exhibited  a detection              I
limit of approximately 20 ug/m3 in the absence of background and approxi-              j
mately 100 ug/m3  in the  presence of 85% broadband extinction (Girwin and              '
Hodson, 1979).
                                    237

-------
      The  unique advantage of the  ZAA  technique  is the ability to monitor
 Hg  on  a  real-time  continuous  basis.   This  is  particularly important
 because  Hg emission from an  in  situ retort may vary markedly during the
 course  of a burn.  Also  important is  the ability of the ZAA spectrometer
 to  achieve background  correction  at  levels up  to 98% extinction, which
 permits  the  determination  of  free Hg  in  smoky  and  organic-laden  gas
 streams.   In  addition,  the  ZAA  avoids  any question  regarding sampling
 efficiency since the entire  sample  stream  passes  through the absorption
 cell.

      One  current disadvantage of ZAA  spectrometry is the requirement for
 expensive,  custom  made  instrumentation.   Laboratory  ZAA  spectrometers
 have  been available commercially only sporadically, and a change in this
 situation  would  increase  the  attractiveness   of  ZAA  spectrometry.   A
 field-worthy  ZAA spectrometer is  currently  under  development  at  LBL but
 is  not  yet commercially  available.   Although the  ZAA spectrometer does
 achieve   superior  correction  for  broadband,  molecular absorption  and
 scattering,  suppression  and  enhancement due  to  the  presence of other
 chemicals  remain a problem.   Finally,  the detectability of the ZAA tech-
 nique is poorer  than the Au amalgamation technique.

 4.6.2  Applications

     The  attempts  to collect Hg  in retort  gas  using  aqueous  solutions
 illustrate the  difficulty of this approach.  For example, Shendrikar and
 Faudel (1978)  attempted  to  measure the Hg mass  balance  during a Fischer
 assay  using a  HN03  solution to  scrub  Hg  from the retort gas.   Their
 results varied  from  58% to 196% recovery, which may be explained in part
 by  the  loss of  Hg to the  condenser or  other cooled  parts  (Donnell  and
 Shaw, 1977).   Fox et  al.  (1978)  used a series of  impingers  containing
 IC1,  NaOH, H202,  and HN03  to  collect  Hg  vapor  in  retort gas  from  a
 simulated  in  situ  retort.   Dark brown deposits  were formed  during samp-
 ling  on the walls  of the impingers, which had to be removed with  an MIBK
wash.   Including the  various  bubbler solutions  and washings, a total  of
                                    238

-------
seven media had to be analyzed for each sampling period.   The majority of
the Hg deposited  in  the brown precipitates rather than in the solutions,
further  complicating the  procedure.   In  spite  of this  somewhat  ardous
approach,  excellent  agreement  was seen with  ZAA measurements  (0.74 vs
0.71 ng/min) when the two methods were operated simultaneously.

     The  applications  of  the Au  amalgamation technique  is  well  illus-
trated by Fruchter et al., (1977) and Kalb (1975).  Fruchter measured.Hg
vapor  in the off gas from  the chemical oxidizer and in the recycle gas at
the  Paraho retort.  By measuring  Hg  in  all  other major  process streams,
he completed a  reasonable Hg mass balance (84%)  for  the retort, a  fact
which  speaks well  for the Au  amalgamation technique.   Fruchter et  al.
 also attempted to speciate both  Hg  and  As through the application of the
 selective sorbents described  in their paper.

      Kolb (1975)  measured Hg  mass balances  in a  coal-fired utility.   He
 recovered on the average 20% more Hg than expected in the flue gas,  which
 is  not  significant considering  the  difficulties  in  representative sam-
 pling and analytical errors.

      Fox et al.  (1978) describe  the  application  of  an on-line ZAA  spec-
 trometer for the measurement  of Hg  in product  gas  from a simulated  ir.
 situ  retort at  Laramie.   They were  able to  detect  Hg  vapor  on  a  con-
 tinuous basis,  and observed  Hg  eluting as  a front as  the burn  neared
 completion.   By  measuring  Hg in the  raw and spent shale,  retort  water,
 oil,  and retort gas,  they completed  a mass balance of 97%!
                                       239

-------
                               REFERENCES
 :                             - 4.0 Gases -


Adams, D.F.   1979.   Sulfur  Gas Emissions  from  Flue Gas Desulfurization
     Sludge Ponds.  Air Poll. Cont. Assoc.  29:963.

Adams, D.F.   1975.   Hydrogen  Sulfide in  Air  Analytical Method.  Health
     Lab. Sci.  12:3628.

Adams, J.,  K.  Menzies  and P,  Lewis.   1977.   Selection and Evaluation of
     Sorbent  Resins  for the Collection of Organic Compounds.  EPA-600/7-
     77044.

American  Chemical Society.  1978.   Trace Analysis  of Nitrogen, Sulfur
     and  Halogen Compounds  in the  Atmosphere.   Preprints  of the 176th
     National  ACS Meeting, Sept.  10-15.   Division of  Environmental Chem.

American  Industrial  Hygiene  Association.   Analytical Abstracts.  A
     Looseleaf, Updateable Collection/

Aneia,  V.P.,  E.P.  Stahel, H.H.  Robers,  A.M.  Witherspoon and W.W.^Heck.
     1978.   Calibration  and Performance  of  a Thermal  Converter  in the
     Continuous Atmospheric  Monitoring of Ammonia.   Anal.  Chem.  50:17058.

APCA.  1976.   Air Pollution  Measurement Accuracy as  it Relates to
      Regulation Compliance.  Pittsburgh.

Artemov,  V.M., E.M. Artemov,  N.D.  Balyansyi,  A.V. Gorelik,  I.M. Nararov,
      Sh.  D.  Fridman, L.I.  Solov'eva, and V.I.  Chernen'kii.   1977.
      Determination  of  the Ammonia Concentration Along  an Extended  Route
      According to Resonance Absorption of Carbon Dioxide Laser Radiation.
      Meteorol. Gidrol.   7:1038.

 ASTM.    1974.  Instrumentation  for Monitoring  Quality.   Philadelphia.

 Austin,   R.R.  1977.   Determination of Hydrogen  Sulfide  and Total  Sulfur
      by  Titration  Methods.   Proc.  Int.  Sen.  Hydrocarbon Meas.   52:249.

 Barras,   R.C.  1974.    Instrumental  for  Monitoring  Air   Quality  ASTM
      Special Technical Publication 555.  Philadelphia.

 Beuerwan,  D.R.   and  C.E.  Meloan.    1962.    Determination   of  Sulfur  in
      Organic  Compounds  by  Gas   Chromatography.   Anal.  Chem.   34:319.
                                    240

-------
                         4.0 REFERENCES (cont.)


Blurton,  K.F.   and   J.R.   Stetter.   1978.   Sensitive  Electrochemical
     Detector for Gas Chromatography.  J. Chromatogr.  155:35-45.

Bostrom,  C.A.,   P.  Grennfelt,  and  K.  Peterson.   1978.   Field  Test of
     Some  Hydrogen  Sulfide-Registering  Instruments  Suitable  for  Work
 i    Environment Monitoring.  Inst.  VattenLuftvardsforsk. B. 42.

Braman,  R.S.,  J.M.  Ammons,  and  J.L. BHcker.   1978.  Preconcentration
     and  Determination  of  Hydrogen Sulfide in Air  by Flame  Photometric
     Detection.   Anal. Chem.  50:9926.

Braman,  R.S.  and  D.L.  Johnson.  1974.   Selective  Absorption Tubes and
     Emission Techniques  for Determination  of Ambient  Forms  of Mercury
     in Air.  Environ. Sci. Techno!.  8:996.

Brerichley, D.L., C.D. Turley,  and  R,F. Yarnac,  1974.  Industrial Source
     Sampling.   Ann Arbor Science.

Bruner,  F.,  P.   Ciccioli,   and  G.  Bertoni.   1976.   Analysis  of Sulfur
     Compounds  in  Environmental  Samples  with  Specific  Detection and
     Selective Columns.   J. Chromatogr.  120:2002.

Buonicore,  A.J.  and L. Theodore.   1975.   Industrial Control  Equipment
     for Gaseous Pollutants—Volume.1.  CRC Press, Cleveland.

Butcher,  J.M.S.  and  G.F.   Kirkbright.   1978.   Direct  Determination of
     Ammonium-Nitrogen  by  Flame  Emission  Spectrometry  in a  Hydrogen-
     Nitrogen Diffusion Flame.  Analyst.  103:1104.

Burnett, C.H., D.F.  Adams,  and S.O.  Farwell.   1977.   Potential Error in
     Linearized  RPD  Responses  for.Sulfur.   J. Chromatogr.  Sci. 15:230.

Cahill, J.E.  1979.  Derivative Spectroscopy.  Amer.  Lab.  79pp.

Carlson, EJ?.',  M.V. Schneider and R.F. McMaster,   October 1978.  Subhar-
     monically Pumped Mi 11imeter-Wave Mixers.   IEEE Trans. MTT
     26:10.   706-715 pp.

Chamberland, A.M.  and J.M.  Gauthier.   1977.   Sensing Sulfur  Oxides and
     Other  Sulfur  Bearing  Pollutants  with  Solid  Electrolyte Pellets.
     I.  Gas Concentration Cells.   Atmos. Environ.  11:257.

Compton, L.E.  1978.  Removing Hydrogen Sulfide from  a Gas.   U.S. Patent
     4,121,663.   Chemical Abstract  90:74264U.
                                    241

-------
                         4.0 REFERENCES (cont.)


Cotter, J.E.,  C.H.  Prien,  J.J.  Schmidt-Collerus,  D.J.  Powell,  R. Sung,
     C. Habenicht and  R.E.  Pressey.   1978.   Sampling and  Analysis  Re-
     search Program at the Paraho Shale Oil Demonstration.
     EPA-600/7-78-065.

Cresser,  M.S.    1977.   Nitrate  Determination  by  Reduction  to Ammonia
     Gas-Phase  Ultraviolet Absorption  Spectrometry.    Analyst  (London).
     102:99.

de Nevens,  N.  et al.   1979.  Analysis of the Environmental Control Tech-
     nology for Oil Shale Development.  DOE Report COO/4043-1.

de Souga, T.L.C.  and  S.P.  Bhatia.   1976.  Development of Calibration and
     Monitoring Systems to  Measure TRS and SO in the PPB Range  Presented
     at the 1976 Pittsburgh Conference.

Deering, R.F.  1978.   Oil  Shale Retorting Process with Desulfurization of
     Flue Gas.   U.S.  Patent 4,069,132.

Derge,  K.   1963.   Furnace  Atmosphere Analysis  by  Means of  Gas Chroma-
     tography.   Geisserliproxes.   21:433.

Dimitriades, B.   1975.   Ammonia  in Air  Analytical Method.   Health Lab.
     Sci.   12:369.

Doehler, W., L.  Huber,  and P. Seifried.   1977.   The  Use of Flame Photo-
     metric  Detectors in  the  Analysis of  Sulfur-Containing Substances.
     GIT Fachz. Lab.   21:937.

Dressall,  J.N.   1977.   Sampling  and Analytical Techniques for Mercury in
     Stationary Sources.  Methods of Air Sampling and Analysis.  Washing-
     ton,  DC.  APHA.

Edrich, J.   December 1979.   Josephson Mixer Receiver  for  400 GHz.  Dig.
     4th Int.  Conf.  on  Infrared and Near-Millimeter Waves. IEEE Cat.  No.
     79 CH 1384-7 MTT.  152-153 pp.

Edrich  J. ,  W.L.  Shuter.  W.H.  McCutcheon and P.C. Gregory.  January 1978.
     Spectral Line Search and Observations Around 100 GHz Using  a Compact
     Cryogenic  Mixer  Receiver  Without  LO  Filter.    Dig.  URSI  Meet.
     Boulder.  1-6 pp.

Edrich, J.   September 1979.  Microwave Techniques in Diagnosis and Treat-
     ment of Cancer.   (Inv. Pap.)  Proc. 9th Europ.  Microw. Conf. Microw.
     Exhib. & Publ.  LTD.   Kent England.  25-33 pp.

Edrich, J.   April  1980.   Development of Multifrequency Upconverter and a
     Subharmonically  Pumped  Mixer.   Prop.   E8026-31  to  W.F.  Kolbe,
     LBL/DOE.
                                   242

-------
                         4.0 REFERENCES (cont.)


Eguchi, H.  and H.  Hamada.   1979.   Determination  of Ammonia  in Various
     Types of Gases.  CA 90:47951.

Farwell, S.W., S.J. GLuck, W.L. Bamesberger, T.M. Schutte and D.F. Adams.
     1979.   Determination of  Sulfur-Containing Gases  by  a Deactivated
     Cryogenic Enrichment and Capillary Gas Chromatographic System. Anal.
     Chem.  51:609.

Fogan, M. L.  and  L.  Dubois.   1974.   The Determination of Ammonium Ion in
     Airborne  Particulates  with  Selective  Ion  Electrodes.   Anal.  Chem.
     Acta.  70:157.

Fox,  J.P.,   R.D.  Mclaughlin,  J.C.   Barket, J.J.  Duvall, K.K.  Mason and
     R.F.  Paulson.  1978.  Mercury Emissions from a Simulated in Situ Oil
     Shale Retort,  llth Annual Oil  Shale Symposium, Golden, CO.

Fraclkin,  B.I.  and B.A.  Petrov.  1978.  Gas Chromatographic Determination
     of Sulfur-Containing Gases in  the Air of Industrial  Premises.  Gig.
     Tr. Prof. Zabol.  6:53.

Gardner,  D.   1976.   A  Rapid Method for the Determination  of Mercury in
     Air  by  Flameless  Atomic Absorption Spectrometry.  Anal. Chim. Acta.
     82:321.

Girvin, D.C.,  T.  Hadeishi   and  J.P. Fox.  1979.   Use of  Zeeman Atomic
     Absorption Spectroscopy for the Measurement of Mercury in Oil Shale
     Gases.   DRI  Symposium  on Oil Shale  Sampling,  Analysis, and Quality
     Assurance, March 26-28, Denver.

Girvin, D.C. and A.T. Hodgson.  April-June 1979.  Distribution of As, Cd,
     Hg,  Pb,  Sb  and Se  During Simulated  in  Situ Oil  Shale Retorting.
     Quarterly Progress Report.

Godin, J.  J.,  L.  Cluet and  C.  Boudene.   1979.  Flame Photometric Deter-
     mination of Carbon Disulfide in Air After Specific Preconcentration.
     Anal. Chem.  51:2100.

Goodfellow,  L.  and  M.T.  Atwood.  1974.  Fischer Assay  of Oil Shale Pro-
     cedures  of the Oil  Shale Corporation.   Proc. of  the  7th Oil Shale
     Symp., Golden, Colorado, Colorado School of Mines.  69:205.

Gordy, W.  and R.L. Cook.  1970.  Microwave Molecular Spectra.  Wiley, New
     York.

Gunther,  F.A.,  J.H.  Barkely, M.J. Kolbezen, R.C.  Blunn and E.A. Staggs.
     1956.   Quantitative Microdetermination  of Gaseous  Ammonia  by its
     Absorption at 204.3 NM.  Anal.  Chem.  28:1985.
                                    243

-------
                         4.0 REFERENCES (cont.)


Hadeishi, T.  1972.  Isotope Shift Zeeman Effect for Trace Element Detec-
     tion.  Appl. Phys..Lett.  21:432.

Hager,  Jr.,  R.N.   1973.  Derivative  Spectroscopy  with Emphasis on Trace
     Gas Analysis.  Anal. Chem.  45.-1131A.

Hager,  Jr.,  R.N.  and  R.C.  Anderson.  1970.   Theory of Derivative Spec-
     trometry. J. Opt. Soc. Arm.  60:1444.

HaVI, R.C.   December  1978.   The Nitrogen Detector in Gas Chromatography.
     CRC Critical Reviews in Analytical Chemistry.  323 p.

Hara, N.   1975.   Capture of Mercury  Vapor  in Air With Potassium Perman-
     ganate Solution.  Industrial Health.  13:243.

Hegyessy, G.  1978.   Determination of Hydrogen Sulfide  in  the Air Using
     the Ion Selective Electrode.  Munkavadelem.  24:21.

Horton,  A.D.,  et al.   1973.   Determination  of N, S, P, and C in Solid
     Biological  Materials Via  Hydrogenation and Element Selection Detec-
     tion.  Anal. Chem. 7:449.

Holb,  G.W.   1975.   Total   Mercury Mass  Balance  at  a  Coal-Fired  Power
     Plant. Trace Elements  in  Fuel.   Advances  in Chemistry.  Washington,
     DC. Series 141.

Hrubesh,  L.W.   March 1973.  Microwave  Rotational  Spectroscopy:   A Tech-
     nique  for  Specific  Pollutant  Monitoring.   Radio  Science.    8:3
     167-175 pp.

Hrukesh, L.W.  June 13, 1977.  Instruction Manual for a Microwave Ammonia
     Monitor.  Lawrence Livermore Laboratory Report UCID17503.

Hurose,  Y.  S.  Kochiwa  and H.  Taniguchi.   1979.   Conversion of Ammonia
     into  Nitric  Oxides by  Using a Heated Catalyst  for  Analysis of Gas
     Mixtures.  CA 91:116860P.

Inn,  E.C.Y.,  J.F.  Vedder,  B.J.  Tyson and D. O'Hara.   1979.   Carbonyl
     Sulfide in the Stratosphere.  Geophys. Res. Lett.  6:191.

Janssen, J.H., J.E.  Van den Enk, R. Bult and D.C. De Groot.    1976. Deter-
     mination of  Total Mercury  in Workroom Air by  Atomic  Absorption or
     X-Ray  Fluorescence Spectrometry  After Collection  on  Carbon-Loaded
     Paper.  Anal, Chim. Acta.   84:319.

Jones,  Jr.,  J.B.   1976.   The  Paraho  Retort.   9th Oil  Shale Symposium.
     Colorado  School   of  Mines,  Golden,  Colorado.   School of  Mines.
     71:39.
                                    244

-------
                        4.0 REFERENCES (cont.)
    Pap. B-l.  2 p.
                 •
     Absorption  Spectroscope   Health Lab.  Sci.   12:158
     Spectrosc.  Radiat.  Transfer.   18:47.

Kothnv   E  L   1974.   Tentative Method  of  Analysis  for Ammonia  in the
^Atmosphere (nitrate method).   Health Lab. Sen.  11.151.

      •    A    H M   T^nunas   1976    Determination  of Hydrogen  Sulfide
K°VatS(HS)  Ai'n aEnndvironTmSe°nt9abSy  Indire^ Atomic  Absorption  Spectroscopy.
     Bull.  Environ. Contain. Toxi col.  14:412.


                ,,SK,
      2524726.
      Acta.   94:213.

 Lawrence Berkeley  Laboratory.   1976    Instrumentation for Environmental
      Monitoring.  University of California.

 Leithe, W.  1971.  The Analysis of Air  Pollutants.  Ann Arbor.

                                 October 1978.   Detection  and  Measurements
       LDL  Kepui l-  /-"-•«  «"~  --i	
       Science Symp.   Washington, DL.

           n D   anrt  J W  Paliuq    1973.   Monitoring  Sulfur  Compounds  by

  ^"Fla^^ottry"-  A^pec. Tech.  Publ.   74:555.
       Chromatogr. Sci.  15:560.
                                       245

-------
                         4.0 REFERENCES (cont.)


Martel,  R.A.  and  A.E.  Harak.   July  1977.   Preliminary  Results  from
     Retorting Michigan Autrium Shale.  LERC/TPR-77/1.

Martin,  R.L. and  J.A.  Grant.   1965.   Determination  of Sulfur Compound
 ;    Distribution  in  Petroleum  Samples   by  Gas  Chromatography  with a
 4    Coulometric Detector.  Anal. Chem.  37:644.

Maier, C.G.  and J.E. Drapeau.  1924.  The  Effects of Various Gases  on the
     Recovery of  Ammonia  from Oil Shale.   Bulletin  from State School of
     Mines.  Univ. of Utah.  14:62.

Mazur,  J.F.,  R.L.   Bamberger,  G.E.  Podolak  and  G.G.  Esposito.   1978.
     Development  and Evaluation  of an Ammonia Dosimeter.  J.  Am.  Ind.
     Hyg. Assdc.  39:749.

McCullen, R.E. and M.T. Michaud.  1978.  Stability Effect  of Hopcalite on
     the Graphite Furnace Determination of Mercury in  Air.  Am. Ind.  Hyd.
     Assoc.  J.  39:684.

Miller, J.D.  and  H.  Koizuwi.   November 1979.  Analytical  Applications of
     Polarized Zeeman AA.   American Laboratory.  35 p.

Nalendycak,  L.  and  M.  Mionski.   1972.    Electrochemical  Generation  and
     Chromerourimetric  Determination of  Hydrogen  Sulfide  in  Air. Chem.
     Anal.   22:55.

Natusch, D.F.S.  and T.M. Thorpe.  1973.  Elemental Selective Detectors in
     Gas Chromatography.  Anal. Chem.  45:1184.

Nelson, G.O.  1972.  Controlled Test Atmospheres.  Ann Arbor.

Nichols, M.  1979.  Electrochemical Vapor  Detector.  CA  96456h.

NIOSH.  Manual of Analytical Methods.  DHEW/NIOSH Pub. No. 77-157-A.  1-5.

Oil Shale  Work Group.  July 1979.   Pollution Control  Guidance  for  Oil
     Shale Developments.  EPA.

Ohkawa, T.  and  M.  Kondo.   1979.  Dry Method for Collection of Mercury in
     Ambient Air  Using  Goldquartz Wool Collector with Thermal Decomposi-
     tion Method.   Eisei Kaguku.  23:191.

Okita, T.  and S.  Kavamori.   1979.  Determination of Trace Concentrations
     of Ammonia in  the  Atmosphere Using the Pyridene  Pyragolene Reagent.
     Atmos.  Environ.  5:621.

Owen,  Q.T.E.  and  F.R.  McDonald.  Preprint.   Preliminary Report on Stack
     Gas  Pollutant  Levels During  the Sideburn at Rock  Springs Site  12.


                                    246

-------
                         4.0 REFERENCES (cont.)


Pearson, C.D. and W.J.  Mines.   1970.  Determination of Hydrogen Sulfide,
     CarbonyT Sulfide, "Carbon  Disulfide,  and Sulfur Dioxide in Gases and
     Hydrocarbon Streams  by Gas  Chromatography/Flame  Photometric Detec-
     tion.   Anal. Chem.   49:123.

PEDCO Environmental,  Inc.   Standards  of  Performance  for  New Stationary
     Sources.  EPA 340/1-77-015.

Petersen, 0.  and H.D.  Schmidt.   1976.   Electrochemical  Cell  for Deter-
     mination of Hydrogen  Sulfide in a Gas Mixture.  Patent No. 2657570.

Pinigina, I.A.,  V.V. Zykova and A.N. Gorchakova.  1976.  Determination of
     Carbon Disulfide in Air.   Gig. Sanit.  12:74.

Popp, P.,  H.J.  Grosse  and G.  Oppermann.   1978.   The  Aerosol  lonization
     Detector:   A  New Detector  for Gas  Chromatography.   J.  Chromatogr.
     147:47.

Reid, A.,  J.  Schewchun,  B.K.  Garside and E.A.  Ballick.   1978.   Point
     Monitoring  of  Ambient Concentrations of  Atmospheric Gases  Using
     Tunable  Lasers.  Opt. Eng.  17:56.

Reid, J., J.  Shewchun,  B.K. Garsied and E.A. Ballik.  1978.  High Sensi-
     tivity  Pollution Detection  Employing  Tunable Diode  Lasers.  Appl.
     Opt.  17:300.

Sandal!s, F.J. and S.A.  Penkett.  1977.  Measurements of Carbonyl Sulfide
     and Carbon Disulfide  in  the Atmosphere.   Atmos.  Environ.  11:197.

Scheide, E.P. and  R.B.J.  Warner.  1978.  A Piezoelectric-Crystal Mercury
     Monitor.  J. Am. Ind. Hyd. Assoc.  39:745.

Schiller, R.G. and R.B.  Bronsky.  1977.  Gas Chromatographic Analysis for
     Hydrogen Sulfide,  Organic  Sulfides, Mercaptans,  and Carbon Dioxide
     in Hydrocarbon Matrices Using an Electrolytic  Conductivity Detector.
     J.  Chromatogr. Sci.  15:541.

Sedlak,   J.M.,  K.F.  BTurton and  R.B.  Cromer.   1976.  Performance Charac-
     teristics  of   an Electrochemical  Hydrogen  Sulfide  Analyser.  Anal.
     Instrum.  14:7.

Simpson, W.R. and  G.  Nickless.   1977.  Rapid Versatile Method for Deter-
     mining Mercury at Subnanogram  Levels by Cold-Vapor Atomic-Absorption
     Spectroscopy.   Analyst.  102:86.

Sims, E.W.   1974.   Determination of Ammonia in Dilute Aqueous Alkali and
     in   Nitrogen-Ethylene  Environment   by   Gas   Chromatography.    J.
     Chromatog.  12:172.


                                     247

-------
                         4.0 REFERENCES (cont.)


Snrirnova, D.M.  and E.G.  Chupeev.  1976.  Colorimetric  Method for Deter-
     mining the Content of Hydrogen Sulfide in Gases.  Tr. -Volgogr. Gos.
     Nauchno-Issled. Prbektn. Inst. Neft. Prom-st. 28:57.

Smith, D.B.  and L.A.  Krause.   1978.  Analysis  of  Charcoal  Tube Samples
:    for  Carbon Disulfide  Using  a  Photonionization Detector.  Am.  Ind.
;    Hyg. Assoc. J.  39:939.

Smith, R.G.  1975.   Tentative Method for Gas Chromatographic Analysis of
     Oxygen,  Nitrogen,  Carbon  Monoxide,  Carbon  Dioxide,  and Methane.
     Health Lab. Sci.  12:173.

Steczkowski, J.  B.  Gronowski, J. Szota  and J.  Bialas.   1975.  Apparatus
     for  Measuring  the   Ammonia Content  in  Gases.   Patent  No.  96409.

Steinle,  K.   1962.  Bestandteile  der Wachenroderschen  Fluessigkeit und
     Fifren  Bildungmechanismus.   Ph.D.   Thesis.   LudwigMaximilians Univ.
     Muenchen.

Stetter, J.R.,  J.M.  Sedlak and K.F.  Blurton.   1977.  Electrochemical Gas
     Chromatographic Detection of Hydrogen Sulfide at PPM  and  PPB  Levels.
     J. Chromatogr. Sci.   15:124.

Stevens, R.K.  and  W.F. Herget. 1974.  Analytical  Methods Applied to Air
     Pollution Measurements.  Ann Arbor  Publishers.

Stevens, R.K.  and  A.E. O'Keefe.   1978.  Modern  Aspects of Air Pollution
     Monitoring.  Anal. Chem.  143A.

Tanaka,  S,   Y.  Hashimoto  and K.  Nakamura.   1977.   Fluorometric Method
     Using the Fluorescein Mercuric Acetate Reagent  for  the Determination
     of  Hydrogen  Sulfide in  the  Atmosphere.   Bunseki  Kagaku.   26:241.

Thompson, B.  Preprint.   Determination of Sulfur Compounds by  Gas  Chroma-
     tography.  Van"an Associates.

Thrun,  K.E.,   J.C. Harris  and  K.   Beltis.    April  1979.    Gas  Sample
     Storage. EPA  600/7-79-095.

TRACOR  Instruments.   1978.   Selective Detection  in Gas  Chromatography.
     Application Note  78-4.

Tsunashima,  S.  T.  Toyona  and  S.  Sato.  1973.   Emission  Processes  in
     Cadmium Photosensitization.  Bull. Chem.  Soc.  Jap.  46(9):  2654-9.

Von  Lhemden, and J. Darryl.   1978.  Suppression Effect of Carbon  Dioxide
     on  FPD  Total  Sulfur  Air Analyzers and Recommended  Corrective  Action.
     Jt. Conf.  Sens. Environ.  Pollut.  4:360.
                                     248

-------
                         4.0 REFERENCES (cont. )


Walker,  D.S.   1978.   Gas-Chromatpgraphic Determination  of  Some Sulfur
     Gases  at  the Volume  Per  Million  Level  in Air  Using  Texax-GC.
     Analyst.  103:397.

Wallace   LD   etal.  1970.   Comparison of Oxi dative  and Radioactive
     Methods for the Microcoulometric Determinations of Six Hydrocarbons.
     Anal. Chem.  42:387.
Willi
           D  T    et al   1973.  Evaluation of Second Derivative Spectros-
           for Monitoring Toxic Air Pollutants.  AD/A-000-949.

Wolff,  G.   1976.  New Measuring  Processes  for Determination of  Hydrogen
      Sulfide.   Draegerheft.   303:15.

Zoellner  W.D., W.F.  Kolbe and  B.  Leskovar.   October 1979.  Noise  Con-
      siSerations  in Millimeter-Wave  Spectrometers.   LBL  Report  7294
      UC-37.   December  1978  and  Dig.  IEEE  Nuclear  Science  Symp.   San
      Francisco.
                                      249

-------
5.0  QUALITY ASSURANCE

     Quality assurance has  been  discussed extensively in the literature^
including  technical,  legal, and  management aspects.  Although  emphasis
and technique may  vary,  quality  assurance can be described and practiced
in terms of the following six categories:

     Personnel.   Persons  carrying out  chemical  analyses,  sampling,  and
     data handling should be adequately educated and experienced.

     Instrumentation.   The instruments should be appropriate for the type
     of  measurement  intended  and  should  be  carefully calibrated  and
     maintained.    Records  of  calibration should  be readily  available.

     Methods.   The  methods of sampling,  preservation, and chemical  analy-
     sis  should  be  thoroughly  tested.   For  this  reason  standardized
     methods are often preferred  when appropriate.   The  exact  analytical
     method should  be  recorded for later reference.

     Quality Control.   Quality control—as  opposed  to  quality  assur-
     ance—refers  to   the  application  of statistical  methods  in  the
     laboratory.   The  precision of each method should be determined as  a
     function of concentration by repeating  the analysis of a  sufficient
     number  of   real  samples.  The   accuracy  should be established by
     analysis of  standard reference materials,  by participation in  "round
     robin"  studies,  and  by  measurement  of the  recovery  of analyte
     material  which has been added to real samples.   In  routine assays,
     control  charts  may  be  used for  detecting  anomalous situations.
                                    250

-------
     Record Keeping.  Accurate  records should  be  kept for  all  samples,
     including  site,  sampling  techniques,  preservation  and  shipment
     methods,  date of  receipt  by the  laboratory,  the exact  analytical
     method, and the raw data that was used to calculate the final value.
     Record keeping generally also includes maintaining permanent labora-
     tory notebooks  and archiving exact  descriptions of  the  analytical
\    methods used.

     Management Structure.    Quality   assurance  programs  provide for  an
     independent  review process.   In  a  research  effort,  reviews  are
     typically undertaken by peers.  In a routine testing laboratory or a
     forensic laboratory, technical reviews may be initiated by a quality
     assurance  manager  who  reports  independently  of  the  laboratory
     manager.  Materials  of known composition  may also  be  submitted by
     the quality assurance laboratory.

     The details  of  such quality assurance programs have been treated by
a number of organizations,  including the Environmental Protection Agency
(EPA), the  American Society for Testing and  Materials (ASTM),  and  the
American  Public  Health  Association  (APHA).   Mills  (1979)  provides  an
extensive  list  of  references  dealing  with  quality assurance.   Books
dealing  more specifically  with oil  shale  include Analytical Chemistry
Pertaining to Oil Shale and Shale Oil,  by   Siggia   and   Uden   (1974),
Science and Technology of Oil Shale,  by T.F. Yen  (1976):  and proceedings
of   the   recent   Oil Shale Symposium:  Sampling, Analysis, and Quality
Assurance,  sponsored  by the Denver Research Institute  (1979).   In  addi-
tion,  standard  setting  agencies  and associations,  including  the  EPA,
ASTM,  APHA,  National  Institute  for  Occupational  Safety  and  Health
(NIOSH), the Association of Official Analytical Chemists  (AOAC) publish
compendia of standardized analytical  procedures.

     The  content of  the previous  chapters  indicates  clearly  that  the
standard methods of chemical analysis and sampling, which are principally
                                    251

-------
 designed for  routine  testing laboratories,  have not always been  suffi-
 cient to assure the correct  chemical  analysis  of oil  shale wastes.   This
 insufficiency arises primarily because the standardized  methods  of  chemi-
 cal  analysis and sampling were not designed for oil  shale wastes and must
 often be  improved  or  entirely  changed.    Further,  realistic  standard
 materials  must be developed for the analysis of oil  shale wastes.

      Analytical  methods are  discussed in the remaining  sections in this
 report.   In  the next section  I discuss the availability  and importance  of
 standard reference materials, as  well  as methods of sampling and  sample
 preservation.   In addition, because of the still  undeveloped character  of
 analytical  methods  for  oil  shale  waste,   and  because   of the  reported
 difficulties  in  analyzing  such  materials,  further quality  assurance               j
 operations are  recommended  until  procedures  have  been more  thoroughly
 tested.  These recommendations are described in the  following sections  of
 this chapter.

 5.1   Standard Reference Materials                                      .      "

      The ability  of an individual  laboratory  to reproducibly  analyze  a
 complex  sample often produces  a false sense of confidence  which is dis-
 rupted  only  when  several  laboratories  analyze the  same,  homogenized
 sample.   In  such  cases it is  not unusual to find  intra-laboratory pre-
 cision on the order of a few percent while  inter-laboratory differences
 may  be larger than a  factor  of ten.   This situation is  amply illustrated
 by von Lehmden et al.  (1974) for  coal,  fly ash,  fuel, and gasoline.  The
 availability of realistic  samples  of  known composition   is  essential for
 revealing  otherwise  unsuspected  analytical  problems and  for  providing
 accurate standardization.                                                               ;
                                                                                        i
;                                                                      :                  i
      Table  5.1  provides  a  list  of  environmental  standards  which are
 currently  used  to  evaluate  the performance of an analytical method for
 environmental  samples.   Because only  two  of the standards  are directly
 related  to oil  shale  retorting, additional standards would  be  valuable.

                                     252

-------
The  gas  standards are  simple  mixtures in pure air  or  nitrogen,  and are
far  removed from the  hot, moist, hydrocarbon-based  gases  that are pro-
duced  in oil shale  retorting.   A more  realistic standard  material  for
retort gas  would therefore  be an important  contribution.   Although the
Laramie  Energy  Technology Center maintains  a  storage facility  with  a
variety  of  product waters from oil shale retorting, to my knowledge only
one  has  been characterized sufficiently to serve as a standard material.
The  water  quality   standards  available  from the  EPA  represent  a  far
simpler  matrix  than  would be encountered with realistic retort waters or
leachates.  Nevertheless,  if a laboratory is unable to determine volatile
organic  compounds in pure water,  it is unlikely  they could do so in the
unstable brines produced by oil shale retorting.

     The standards materials in Table 5.1 represent a substantial  advance
in the ability  of the analytical  chemist to handle complex environmental
samples.   Nevertheless,  for  the  purpose  of  oil  shale retorting three
types of standards are notably missing:

      o    Gas  standards  which simulate  the  wet,  hot,  hydro-
           carbon-based stream produced by oil shale retorting.
           Although   standards for  regulated gases such as  S02,
           CO,  and NO can be  purchased (albeit  in  an  unreal-
           istically  pure  matrix),  standards  for trace metals
           in gases are entirely unavailable.

      o    Standards   with  known  concentrations   of  organic
           materials.

      o    Solid samples for leaching studies.  The  main inter-
           est in solid waste  is  in  the leachates which may be
           produced,    and  standardized   leaching   protocols
           clearly require the  availability of a wide  variety
           of solid  waste  with  known  leaching characteristics.
                                    253

-------
CO
LU
a.
to
I-
3E
'Z.


§
I-H
 ai
 u
 £Z
 CD

 ai
<*-
 co
ce.

 to
 o

 3
 O
                        in
                        en  <•—\
                        rH  en
                             en
                             rH
en        O)

en   eo   3
rH        J3

     o   co
co   -a   co
CQ   c   s-
z   oo
cr>
r»»
en
CO
ea
en
r«.
en
co
00
en
r^
en
co
CO
                                                o

                                                "re
                                                                                  en
                                                                                  en   a>
                                                                                  t-i   cr
                                                                         CO   C
                                                                         CO   O
                                                           O
                                                           U
LU
O
CO
I-H
CO
•>-
._]
;«
•Z
•=t
IJLJ
1—
O
U_
co
1— 1
1—
•a: to
SE _CO

UJ (J

                                                      c
                                                      a>

                                                      0}
                                                     i —
                                                      a>

                                                      QJ
                                                      (J
                                                      tO
                                                      s-
                                            4J

                                            CO


                                            "o
                                            (A   (A

                                            r—   re
                                            re  -P
                                            •i-   CO
                                            (A

                                            3
                                                                                        OJ
                                                                                        O
                                                                                        re
                                                                                   c   s-
                                            JD

                                            3
o
Q£
•a:
P
«c

co
UJ
._i
QQ
 co

"o.


co

<»-
 o


 CL
               a>
              4->
               re

          5   3
          re  T-
                       .
                     ~   re
                   •—   a.
                         re
                   i—  X>
                    O)   S-
                    03
                         O)
                   i—   CO
                   M-   U)

                     »   s-   ai
                   r—   CO  *>
                    re   >  +J
                    o  •!-   re
                   o   s-   E
                                       re
                    o
                    Q.
                          cs
                    re  z
                   •r-
                    t-   i-
                         -
                                                          >   S-
                                                          re   co
                                                          co   >
                                                          •o

                                                          re
                                                                         co
                         O)   O   >
                         o    s-   o
                        i—    O   JD
                         O   v^
                        *»—
                        OQ
                                                                         (A
                                                                         CO
                                                                        .a
                              c
                              o
                             •r-
                             •P
                              U   **~*
                              CO   C
                             r—   re
                             i—   -P
                              O   (A
                              O   O

                             r—   O
                              re   c_>
                              o   s->
                              u
                              i-   (A
                              re   r—
                                  t-
                                  co
                                            re
                                                     254