United States       Air and Energy Environmental    EPA/600/9-90/005d
          Environmental Protection   Research Laboratory       January 1990
                     Research Triangle Park NC 2771 1
          Research and Development
v>EPA     The 1990 International
          Symposium on Radon
          and Radon Reduction
          Technology:
          Volume IV. Preprints

          Session VII: Radon
           Reduction Methods
          Session D-VII: Radon
           Reduction Methods—
           POSTERS
          Session D-IX: Radon in
           Schools and Large
           Buildings—POSTERS
          February 19-23,1990
          Stouffer Waverly Hotel
          Atlanta, Georgia

-------
       Session VII:
Radon Reduction Methods

-------
                                                                        VII-1
        EVALUATION OF SUB-SLAB VENTILATION FOR INDOOR RADON REDUCTION
                            IN SLAB-ON-GRADE HOUSES

                    by:  D.  Bruce Henschel
                        Air and Energy Engineering Research Laboratory
                        U.  S. Environmental Protection Agency
                        Research Triangle Park,  NC  27711

                        Arthur G. Scott
                        Andrew Robertson
                        AMERICAN ATCON, INC.
                        Wilmington, DE  19899

                        William 0. Findlay
                        Acres International Corp.
                        Amberst, NY  14228
                                   ABSTRACT

      Sub-slab ventilation  (SSV)  radon reduction systems were  tested in nine
slab-on-grade houses around Dayton, Ohio.  The testing addressed the following
house design/construction variables:   slab size; foundation wall  material of
construction; presence  or  absence  of sub-slab  forced-air  supply  ducts;  and
presence of sunken living rooms as  an obstruction and soil  gas  entry route in
the slab interior.  SSV design/operating variables addressed  include:  placement
of the ventilation pipes inside or outside the house;  location and number of vent
pipes;  SSV  fan  capacity;  and  operation of the  SSV  system  in  suction vs.  in
pressure. The results suggest that large  slabs, block foundation  walls, and sub-
slab  ducts  can  sometimes necessitate  additional care in SSV design (number,
location of vent points), but that, in general, radon levels can be reduced to
2 pCi/L or less  with one or  two points if there is a good aggregate layer under
the  slab.   SSV  from inside and  outside the  house give  generally comparable
performance, although  in some cases  interior might  be preferable  for  large
houses.   Increasing the number of suction pipes from one to  two,  and increasing
fan  capacity  from  half to  full  speed,  generally  (but not always)  appear to
improve SSV performance.  Only in the case of one large house  with sub-slab ducts
were  two  pipes  required to  reduce  levels below 4 pCi/L.   Low fan speed was
generally sufficient to reduce levels below 4 pCi/L.   Operation of SSV systems
in pressure never gave better reductions  than did operation in suction.  Results
from  testing of slab-on-grade houses  in Florida indicate that  SSV systems
generally require more suction pipes than were found necessary in Ohio, when the
sub-slab material has poor permeability.

      This paper has been reviewed  in accordance with the U.  S. Environmental
Protection  Agency's peer  and administrative  review  policies,  and  has  been
approved for presentation and publication.

-------
                                 INTRODUCTION

      Much of the previous  testing of residential radon reduction techniques has
focused on houses having basements,  including basements with adjoining slab-on-
grade or  crawl-space wings.  Testing  has  been more  limited  in  houses having
solely a slab-on-grade substructure.  Slab-on-grade houses offer various design
and construction features  which  could impact  the installation and performance
of traditional sub-slab depressurization systems, and which, in some cases, might
offer potential for application of alternative radon reduction techniques beyond
sub-slab systems.

      Previous reports  (1,2)  presented results  from  initial  (Phase  1) testing
of indoor radon reduction methods in four slab-on-grade houses in Dayton, Ohio.
This paper expands  those previous results  to  include further  testing and data
analysis for  those  four  Phase 1  houses,  plus  testing on five  additional slab-
on-grade houses  as  part of a Phase  2  effort  in the  Dayton area.   To provide
additional perspective on this  substructure  type,   recent  results  are  also
summarized from  testing  in 1  slab-on-grade house in  Maryland,  and in 14 slab-
on-grade houses  in Florida  (3).

      A number of house design and construction  variables can impact the design
and performance  of radon reduction systems in slab-on-grade houses.   The house
design/construction variables addressed in the Ohio study include:  slab size;
nature of the foundation stem wall (hollow-block vs. poured concrete); presence
or absence of forced air heating  supply ducts underneath the slab; the presence
or absence of interior slab openings/soil gas  entry  routes,  such as plumbing
openings under bathtubs;  and sub-slab obstructions, such as sunken living rooms.
Among  other   house  variables  that  could be   significant  but  that are  not
specifically addressed by the study in Ohio  are:  absence of aggregate underneath
the slab  (all of the Ohio  houses had a good aggregate layer); nature  of the
interface between the floor slab and the foundation; sub-slab obstructions other
than sunken  living  rooms (e.g.,  interior grade beams); and the  impact  if the
forced-air ducts under the slab are return ducts (at negative pressure) rather
than supply ducts (positive pressure).   Some of  these  other variables are being
addressed in testing underway or  planned in slab-on-grade houses in Florida and
New Mexico.

      The characteristics  of  the native  soil underlying the house can also be
important, especially  if there is no aggregate  underneath  the  slab.   All nine
of the  slab-on-grade  houses  in Ohio  (and the one  house  in Maryland)  were
underlain by  clay soil, with a layer  of  aggregate between the clay and the bottom
of the  slab.   The Florida houses  were on top of  fine,  moist  sand with no
aggregate, providing relatively poor sub-slab communication; as discussed later,
these characteristics greatly  influence the design of  sub-slab ventilation (SSV)
systems for the Florida houses. Future  testing in New Mexico will address houses
with no aggregate over expansive clays and coarse, dry sands.

      SSV was the radon reduction method emphasized in the Ohio testing, and is
the technique  discussed in this paper.   Mitigation  design variables studied
during this testing included:  placement of the ventilation pipes vertically down
through the slab from inside the house ("interior" SSV) vs.  horizontally through

-------
the foundation wall from outdoors  ("exterior" SSV); location of the vent pipes;
number of vent pipes;  SSV fan capacity;  and operation of the SSV system to draw
suction on  the sub-slab region vs.  to  pressurize the sub-slab.   Other radon
reduction approaches which were  tested in addition to SSV  -- but which were less
effective and which are not covered in this paper -- were:  continuous operation
of the central furnace  fan in houses having  sub-slab  forced-air supply ducts,
in an effort to pressurize the sub-slab region; closure of the slab opening where
the bathtub  plumbing comes up through the  slab;  and "site ventilation."  Site
ventilation involves suction  on  a pipe embedded in the  ground outside the house,
in an effort to draw soil gas from the entire lot.

      Table  1 lists the nine  slab-on-grade study houses in Ohio, including the
results of pre-mitigation alpha-track measurements in each house over a 3-month
period during the winter.  Various pertinent house variables are also summarized.
Soil gas is the predominant source of the radon in all  of  the houses; well water
and building materials were determined not to be significant radon sources.
                              MEASUREMENT METHODS

      For short-term measurements  of mitigation system performance, radon gas
was measured using a Pylon Model AB-5 continuous radon monitor equipped with a
285 mL  (17.4 in.3)  Lucas scintillation  cell.   The Pylons  were  programmed to
measure  radon  hourly.    Because  of the  significant day-to-day variability in
indoor radon concentrations  observed in the Ohio  houses,  the short-term radon
measurements consisted  of at least 48 hours (and  commonly  up  to 96 hours) of
Pylon readings both before and after the mitigation  system was activated.  System
on/off measurements were made  back-to-back,  to the extent possible, to reduce
temporal  variations.   Measurements  were made  in  the main  living  area under
closed-house conditions.  Essentially all of the monitoring was  completed during
the heating season (between  December  and March), except for a few of the Phase
1 measurements.

      Diagnostic  testing was  generally limited primarily  to  measurements of
suctions, flow  rates,  and radon  concentrations in the piping of operating SSV
systems.   Sub-slab communication  measurements were performed in some  of the
houses,  but  were  generally not conducted in houses with  sub-slab ducts.   The
ducts  commonly radiated out  from a  centrally located  furnace,   effectively
subdividing  the  sub-slab  region  into  multiple  zones.    Initial  sub-slab
communication measurements in houses  with ducts confirmed that suction did not
extend between  the sub-slab  zones  created by the  ducts, presumably because of
air leakage through the ducts; however, the suction  field did extend well through
the aggregate within a single zone.

      Three-month alpha-track detector measurements were made in the houses prior
to mitigation, as reported in Table 1.  To measure  long-term  system performance,
alpha-track measurements will be repeated for  at least 1 year with  the systems
operating.  For quality  assurance, these detectors are deployed  in  clusters of
two or three;  in addition, unexposed  detectors, and detectors exposed to known
radon levels in a chamber, are sent blind to the analytical  laboratory.

-------
                    SUB-SLAB SYSTEM DESIGN AND INSTALLATION

      In houses  having poured concrete foundation walls,  an exterior suction
point was created by coring a 4-in. (10-cm) diameter hole horizontally through
the wall below slab level, to expose the sub-slab aggregate.  A limited amount
of aggregate  in  the  immediate vicinity of the hole was removed; otherwise, no
effort was made  to  excavate an opening under the slab.   A fan was mounted on
the foundation wall over the hole, exhausting horizontally away from the house
at grade level.  In the final installations,  this  fan was  mounted on a piece of
marine-grade  plywood  that  was  sealed onto the  foundation wall over the hole.
No piping extended from the fan through the hole into the sub-slab region.  In
houses having hollow-block foundation walls, exterior SSV points involved a 5-
in. (13-cm) hole drilled through the foundation wall;  a short segment of 4-in.
FVC pipe was  inserted through the hole, and expanding foam  was injected into the
annular gap between the pipe and the hole in the wall.  This step was taken to
reduce the amount of  air drawn  into  the fan through the block cores (i.e., to
reduce the block wall ventilation component of this system), since significant
leakage of air through the block  cores might be  expected.   The  fan  was then
mounted on the outdoor end of the pipe segment, exhausting away from the house
at grade.

      For interior SSV systems, vertical PVC pipes penetrated the slab, generally
in a furnace room, utility room,  or bedroom closet.  Again,  aggregate immediately
under the hole was removed, but no effort was made to  excavate a pit under the
slab.  Each pipe extended up through the ceiling into the  attic.  If there were
more than one suction pipe, the multiple pipes would be manifolded together in
the attic,  and connected to a single fan.  The fan  was mounted in the attic, and
exhausted through the roof at a  convenient location.  In House 58, to determine
whether interior SSV could be simulated from outside the house, a PVC pipe was
inserted horizontally through the foundation wall  from outdoors, for a distance
of 6 to  7  ft  (about 2 m) under the slab,  simulating  an interior pipe through
the slab at that location.

      Both the "exterior" and "interior" approaches were tested in four of the
houses,  to permit direct  comparison of the two approaches  in the  same house.
The interior systems in three of the houses  included two suction pipes, located
in different parts of  the  house, permitting testing of either pipe by itself,
or both  simultaneously.   This multiple-pipe  installation  was to enable  an
assessment of  the effect  of the number and location  of the vent  pipes.   The
exterior system in the largest house (House 77) had two vent points, which could
be operated one at a time, or both simultaneously (each with its own fan).  In
several houses where only exterior  SSV  was  tested, the single fan was tested at
two locations.  In most houses,  different fan speeds were  tested.

      In all  cases,  Kanalflakt T2 plastic-bodied  fans were used,  capable of
drawing 270 cfm  (127  L/s)  at  zero  static  pressure,  and capable of 1.4 in. H,0
(350 Pa) suction at zero flow.

-------
                                    RESULTS

SSV SYSTEMS IN SUCTION

      The best performance of sub-slab suction systems in each of the slab-on-
grade houses is shown  in Table  2.   Where best performance was obtained during
operation of two suction points  simultaneously,  these cases  are marked in the
table; in all other cases, only one suction pipe was operating.   In houses where
both exterior and interior SSV were tested, the results of the best installation
of each are shown in the table  for  comparison.   In houses where single exterior
SSV suction points were tested at two locations, the performance  is reported for
each location  (designated Site  1 and  Site 2).   In all cases, best performance
was observed with the fan operating at high speed, although good performance was
also attained at reduced fan speed.   In general, these best results were obtained
with the slab opening under the bathtub having been foamed closed.  The pre- and
post-mitigation radon  levels presented  here  are short-term  (48-  to 96-hour)
Pylon measurements,  generally made  immediately before  and  immediately after
activation of the SSV fan, for back-to-back system on/off comparison.  The first
5  hours of  data  taken immediately  after the  fan was  turned  on or  off are
excluded, in order to reduce the effect of the transient on the reported results.

      Several  conclusions are  apparent  from  Table  2.   Substantial reductions
--  generally to a residual  radon  level of less than 1 to 2 pCi/L  (37 to 74
Bq/m3)*  -- were achieved in all  houses, with no more than two suction points (and
generally only one).   The generally good performance was probably aided by the
very low air-flow resistance of the good layer of aggregate under the slab.  As
indicated in the  table, the suctions in  the  SSV system piping were generally
above 0.5 in. H20  (125  Pa), and the flows were generally below 140 cfm (65 L/s) ,
which is half  fan capacity.  These suctions/flows generally represent good SSV
operation, and confirm that large amounts of  air were  not short-circuiting into
the systems via sub-slab ducts or block  foundation walls.

      The apparent effects of the house variables and SSV system design/operating
variables are  discussed below.

Effect  of Sub-Slab Forced-Air Supply  Ducts--

      The initial concern had been  that houses with sub-slab heating ducts would
not achieve high radon reductions with a reasonable number of  SSV pipes, because
the ducts would prevent effective distribution of the  suction field underneath
the slab.  However,  the results  in Table  2 indicate that  -- with the aggregate
present under these houses  -- any effect of the ducts is limited.  The  two houses
without sub-slab heating ducts  (Houses 70 and  72) did  not  consistently perform
better  than did comparable  houses  with  ducts.  On one  hand,  House  70 (large
house, block foundation, no ducts)  appeared to give better results with two-pipe
interior SSV (0.2 pCi/L, 99% reduction)  than did House 77,  which was comparable
in characteristics except that it had ducts (1.7 pCi/L, 89% reduction).  But on
*  Radon  in Bq/m3 - 37 x value in pCi/L.

-------
the other hand,  House  72 (moderate size, block foundation,  no  ducts)  did not
seem to do as well with one-pipe interior SSV  (1.6 pCi/L, 93% reduction) as did
House 58, which was comparable but had ducts (0.8 pCi/L, 97% reduction).  These
radon levels  are sufficiently  low that --  given temporal  and house-to-house
variability, and measurement  uncertainties --  it  is  not possible to quantify,
from these data, any limited effect that the sub-slab ducts were having on SSV
performance.  The results from House 70 vs. House 77 are  further compared later.

Effect of Interior vs. Exterior SSV--

      From the standpoint of aesthetics, the choice between interior and exterior
SSV for a given house will depend on the layout of the house and the preferences
of the homeowner.  It  was desired to determine whether technical considerations
might suggest  one over  the other  under certain circumstances.   For example,
exterior SSV might be  aiding in good performance in houses with radial sub-slab
ducts, because  the suction is  drawn at  the  perimeter  and  thus  might extend
through the loose fill around the perimeter slab/foundation wall junction.  In
houses with block foundation walls  below the slab,  exterior SSV might be helpful
or be a disadvantage, depending  on  the effects  of the  wall ventilation component
that is likely to result.  Large slabs with potential interior entry routes (in
addition to the  perimeter wall/floor  joint) might be less effectively treated
with exterior SSV.

      In the four houses in Table  2 where both interior and exterior SSV were
tested, the two SSV approaches appeared to perform about equally well in three
of the houses (Houses  31, 58,  and 77).  These  three houses cover the range from
a  small  house with a  poured foundation  (31)  to a  large  house  with  a block
foundation  (77);  all had sub-slab ducts.  In  the fourth  house,  House 70, the
interior SSV system appears distinctly superior; however, it must be noted that
the interior SSV  system  utilized two  suction  pipes,  while the exterior system
drew on only one point.   The large size of this house, and its block foundation,
might also have contributed to this result.  It should also be noted that in one
of the houses where interior  and exterior SSV appeared  comparable (House 77),
a two-point interior system connected  to a single fan  is being compared to a two-
point exterior system where each exterior point has its own fan;  it is not known
whether  the two-point exterior system would  have   performed  as well  as the
interior system if both  exterior points had been connected to a single fan.

Effect of Slab Size--

      Re gar ding the effect  of house size,  the largest house (House  77, with a
slab of 2,600 ft2,  or  240 m2) required two suction pipes to reduce radon levels
below 4 pCi/L;  one pipe was  insufficient.  The  other  large house  (House 70, with
2,350 ft2,  or  220 m2)  is also  shown  in Table 2  as  having  achieved  its best
performance with  two interior  suction pipes;  however, operation of  any one of
the pipes was sufficient to  reduce  levels below 2 to  3 pCi/L.  The  likely reason
for the better performance with one point in  House  70 is  that it did not have
sub-slab ducts; in other  respects, Houses 70 and 77  are similar (large houses,
block foundations), although House 77  did have the added complexity of a sunken
living room.  For all of  the other houses (ranging in size from 1,100 to 1,700
ft2, or 90 to 160 m2), one suction point was sufficient to reduce levels below

-------
2 pCi/L.  The effects  of multiple  suction points  are discussed further later.
As  indicated previously,  it  appears  that  interior SSV  might sometimes  be
preferred in houses with large slabs; but  exterior and interior SSV appeared to
perform equally well in houses with smaller slabs.

Effect of Block vs. Poured Concrete Foundation--

      Regarding the effect of foundation  material of construction,  the houses
having  poured concrete  foundations would  appear to  have most  consistently
achieved the best  reductions.  All four of the houses with poured foundations
(Houses 15, 31,  47, and 58) were  reduced to below  1 pCi/L (97 to 99% reduction)
with each SSV approach  (interior or exterior)  tested in that house -- even though
all of  these houses had sub-slab ducts.   By comparison, only  one  of the five
houses  with block  foundations was reduced  below 1  pCi/L,  with  a  two-point
interior SSV  system (House 70, having no sub-slab ducts).  The  other block-
foundation  houses  were  reduced to  levels of 1.6  to  3.7 pCi/L  (81  to  93%
reduction).   In part,  the  poorer reductions  in the  block  houses might  be
explained by the fact  that the block houses were  generally larger;  except for
House 85, all of the block houses were 1,700 ft2 or larger, whereas all of the
poured-foundation houses were  1,400 ft2 or smaller.  But size might not be the
only explanation:   House 85,  with  a block foundation,  was  only 1,100 ft2,  but
it was not reduced  below 1.6  -  2.5 pCi/L.   Perhaps  air leakage through the block
foundation walls is reducing  the effectiveness of suction distribution under the
slab.

Effect of Number and Location  of Suction Points--

      In  three  of  the houses, two-point  SSV  systems  were  tested.    Table  3
compares the results obtained in these  three houses when the system is operated
with each of the suction points by themselves,  and when the system is operated
with  both points.    For Houses  70  and  77, the  results  reported  previously
in  Table  2 reflected  operation with both points.   For all cases  reported in
Table 3, the mitigation  fan is being operated  at high speed.

      Several observations can be  made from Table 3.   In  House 31  --  a small
house with  a  poured foundation and sub-slab ducts --  the  best performance is
achieved  when only  the one  centrally located pipe in  the  furnace  room  is
operated; connecting the second pipe, from the bath,  causes a deterioration in
performance compared  to the  furnace room pipe alone.   And,  not surprisingly,
operation  of the  bath pipe   by  itself gives  the  poorest performance  of  all
(although still reducing levels to below 3 pCi/L).  The reason for this effect
is not clear from the  flow/suction  measurements in the piping.  The suction pipe
in  the bath maintains  reasonable suction (0.4 to 0.8  in. H20, or 100 to 200 Pa)
and reasonable flows (29 to 49 cfm, or 14 to 23  L/s),  with  the lower end of each
range reflecting operation of this pipe  in  combination with  the  furnace room
pipe, and the upper end reflecting  operation of this pipe by itself.  From these
measurements, the  bath  pipe  does  not appear  to  be degrading  performance  by
providing a short-circuit for air leakage into the system.  The furnace room pipe
appears to be best  intercepting the radon  source,  and operation  of this pipe by
itself -- resulting in maximum suction  (0.6  in. HjO, or 150  Pa) and maximum flow
(69 cfm, or 32  L/s) in  this  pipe  -- appears clearly desirable.  The apparent

-------
conclusion is that more suction pipes are not always better.

      In Houses 70 and 77, operation of two suction points is always better than
operating only one.  However, in House 70, operation of either point by itself
is sufficient  to  reduce  the house below 4 pCi/L.  By  comparison,  in House 77,
operation of any one point is only sufficient  to  reduce levels to between 6 and
8 pCi/L;  operation of both points is required to achieve sufficiently low levels.
Both of these houses are  large with block  foundations; the probable explanation
for the better performance of single points in House 70 is that it does not have
sub-slab heating ducts, and House 77  does  have  sub-slab ducts.  The presence of
a  sunken living  room  in House  77 could also be  contributing to  the  poorer
performance of the single points in that house.  It is noted that slightly better
single-point performance with the interior SSV in House 70  was  achieved with
Point #2, which was located in the right front quadrant of the house, rather than
with Point #1, which was  centrally located.  Thus,  central location of a single
suction pipe is not necessarily required,  even  in a large house,  if there is good
communication without  obstructions such  as sub-slab ducts.   It  is  also noted
that, in House 77, the individual exterior SSV  suction points seemed to perform
slightly better than did  the individual interior pipes.  This  result could be
suggesting that -- where  there are insufficient suction points in houses having
sub-slab ducts -- suction at the perimeter of  the slab might be helpful because
it  "taps into" the  permeable  region  where excavation for the  footings  had
occurred during construction, a  region which extends around the entire perimeter
and which is less interrupted by the  sub-slab ducts.  With the block foundation
in House 77,  this result could also be  suggesting that the wall ventilation
component  that might  exist  with exterior SSV  systems could be  important,
especially when there is an insufficient number  of points.  (As noted previously,
steps were  taken  in most  of the  block-foundation houses  to reduce  the wall
ventilation component of exterior SSV systems.)

      In the  two  houses where exterior SSV was tested at different locations,
there appeared to be no effect of location in one (House 15), and some possible
effect in the second (House  65).  Both  of  these houses are small, and have sub-
slab ducts; House  15 has a poured foundation,  and House 85  a block foundation.

Effect of the Wall Ventilation Component in Block-Foundation Houses--

      An effort was made  In  House 1 to  investigate  the significance of the wall
ventilation component of exterior SSV in block-foundation houses.  The original
Phase 1  exterior  SSV system  in  this house  had consisted  of a T2  fan mounted
directly over the  hole through  the foundation  wall; there was  no pipe through
the wall,  nor were the  block cores surrounding  the  hole  mortared  or  foamed
closed.  Under these circumstances, it would be expected that there would be a
significant  wall  ventilation   component  to this  sub-slab-plus-block-wall
ventilation system.   As  discussed in  Reference  1, this Phase  1 installation
resulted  in  indoor levels  of  1.3  pCi/L (during  mild  weather),  the highest
residual radon level  among  the  four Phase 1 slab-on-grade  houses.   The level
increased to 10.6 pCi/L when  the central furnace  fan was operated continuously
(to pressurize  the sub-slab, counteracting the  efforts of  the  SSV  system to
depressurize  the  sub-slab);  continuous  fan operation had  not caused  such  a
problem  in the other three Phase 1 houses.

-------
      The first step  taken  at  House 1,  as part of the Phase  2  effort,  was to
remove the original fan, and to  insert a 4-in. (10-cm) diameter pipe horizontally
through the foundation wall  into  the sub-slab  aggregate.   The T2 fan was then
re-mounted on this pipe.  Expanding foam was injected to close the block voids
around where the pipe penetrated the foundation.  These steps should have greatly
reduced the wall ventilation component of  the  system.   Following these system
modifications, radon levels  in the house averaged 7.8 pCi/L over a 96-hour period
in December, higher than had been observed with the original system.  The pipe
and foam were then removed,  and the fan mounted on the foundation wall over the
hole, as with the original installation.  With  this original configuration, the
indoor levels averaged 2.8 pCi/L over 94 hours  in January  (somewhat higher than
the 1.3 pCi/L reported earlier,  perhaps because  the earlier measurements had been
made in April).   The conclusion would appear  to be that the  wall ventilation
component can be important in some cases.   Taken together with the results from
House 77, discussed above,  it would appear that the wall ventilation component
might be most important where the SSV system is marginal  (e.g.,  in large houses
having only one suction point).  Since interior SSV performed well in the block-
foundation houses (70, 72, and 77) when sufficient  suction points were provided,
it would appear that a major direct wall ventilation component  is not required
when the system is adequately designed.

Effect of Fan Capacity--

      In many of the houses, the SSV system was operated with  the Kanalflakt T2
fan both at maximum suction/flow and at about half suction, to  investigate the
effect of fan  capacity.   The performances  at the  two different  fan capacities
are compared in Table 4.  The ability to achieve good performance at  reduced fan
capacity is important.  Some percentage (probably between 30 and  70%) of the gas
exhausted by the fan is air withdrawn from the  house;  thus,  if fan capacity can
be reduced without greatly reducing the radon reduction performance,  the heating
and cooling penalty of  the  SSV  system will also be reduced.

      As  apparent  from Table 4, operation of  the  fan at high capacity almost
always  resulted in  lower  indoor  radon  concentrations  than  did operation at
reduced capacity.  The exceptions are Houses 15, 31 (with interior pipe #1), and
85 (exterior site 1), where  performance was the same at both capacities.  In the
majority  of  cases,  operation at reduced  capacity  was  sufficient to reduce the
house below  4 pCi/L, although  operation at  high  capacity reduced levels even
further (generally below 2 pCi/L).  The most dramatic improvements from operating
at high capacity occur  in  those  cases  where  high-capacity  operation  is not
sufficient to reduce  levels  below  about 2 pCi/L -- Houses  1,  31  (interior pipe
#2),  70 (exterior,  and interior  pipe #1),  and 77  (all  single-suction-point
cases).  For these cases, the average radon level with the fan at high capacity
was 4.0 pCi/L,  while the average  at reduced capacity was 9.2 pCi/L.   For the
other cases, where  high-capacity operation achieved 2 pCi/L or  less, reducing
the  fan to  about half capacity increased  radon levels  from an average of 1.1
pCi/L (at high capacity) to an  average of 2.1 pCi/L (at reduced capacity).  Only
in  the  case of  House 77  (both interior points)  did high-capacity operation
achieve less than 2 pCi/L while  reduced-capacity operation did not reduce levels
below 4 pCi/L.  In summary,  it would appear  that houses which are  "easy" to treat

-------
with SSV can be reduced below 4 pCi/L with fans of this size operating at reduced
capacity; however, even with  "easy" houses,  there  is usually (but not always)
some marginal additional radon  reduction that  can  be achieved by operating at
high capacity.

      The long-term objective defined by the  Indoor Radon  Abatement Act of 1988
is to reduce  indoor  radon concentrations to the levels which exist outdoors.
To approach this objective, operation of the  fans at  high  capacity in all cases
would be preferred,  since  radon is generally reduced to  lower levels at high
capacity.  Comparing the system flows  indicated in  Table 4, flows average about
35 cfm  (16  L/s)  less when  the fan is  at low capacity.  Electrical consumption
at low  capacity is roughly one-third that  at high  capacity.   Depending upon
specific assumptions  regarding the fraction of the 35 cfm which is withdrawn from
the house,  the  local climate,  and energy costs,  the  increased heating penalty
and electricity  cost resulting from operating  the  fan at  high capacity rather
than low capacity would be roughly $55 per year.

SSV SYSTEMS IN PRESSURE

      In four of the Phase 2  slab-on-grade  houses, the  SSV fans were reversed
so that the system was pressurizing  the sub-slab region rather  than drawing
suction.  In almost all cases,  these sub-slab pressurization tests were conducted
with the exterior SSV systems, since there  is easy  access  to  the external fans;
only in House 70 was pressurization tested with an interior SSV system having
the fan mounted  in the attic.  The results are shown in Table  5.  Table 5 also
presents the SSV pressurization results reported previously for the houses tested
during  Phase 1.   In all cases, the suction vs. pressure comparisons shown in the
table  are  with  the  suction  and pressurization  systems  being operated  at
comparable conditions:  the same ventilation points,  and the  same fan capacity.

      These results,  now with seven houses, confirm  the earlier conclusions drawn
from the three houses tested in the Phase 1  effort.  In all cases, operation of
the SSV system  in suction gives greater radon reductions  (88-99%)  than does
operation in pressure (43-90%).  In several  cases  (Houses 31,  47, and 70), the
difference between suction and pressure is dramatic.

      For  sub-slab pressurization to be  effective,  the  system probably must
create  sufficient flows of fresh air  under the slab  to dilute  the radon in the
sub-slab gas.   Because the  sub-slab  is under pressure, any  radon in the sub-slab
gas will be forced up into  the house.  Unlike sub-slab suction,  where the system
works by reversing the flows of sub-slab gas,  establishing  a pressure field under
the slab may not be sufficient by itself with pressurization.   With the geology
present in Dayton, flows in the sub-slab systems were apparently not sufficiently
great to  provide the sub-slab dilution necessary  to make pressurization work
effectively.

COMPARISON WITH  RESULTS FROM OTHER STUDIES

      The one slab-on-grade house tested by Infiltec in an  EPA-sponsored project
in Maryland was a relatively  large house  (2,700 ft2, or  250  ma)  with a block
foundation  stem  wall and with forced-air supply ducts under the slab, similar

-------
to House 77 in Ohio.  The geology in Maryland was also similar to Ohio -- clay
soil, with a good layer of aggregate underneath the slab.  However, the results
with  interior  SSV  in  the Maryland house were  distinctly better  than  those
observed in House 77.   A single  interior SSV pipe, located approximately in the
middle of the slab  and operating with  a Kanalflakt  T2 fan in suction, reduced
radon concentrations in the Maryland house from a pre-mitigation level of 6 to
21 pCi/L, down to 0.7  pCi/L,  a reduction 90% or greater.  In House 77, a single
interior pipe had been sufficient only to reduce levels from 16 to about 7 pCi/L
(see Table 3).  It is not clear why the performance is so different between these
two houses.

      Fourteen  slab-on-grade houses  tested in  Florida  (eight  of  which  are
reported in Reference 3) had no aggregate underneath the slabs.  The slabs rested
directly on fill material consisting of fine, compacted sand and clay which often
had  limited permeability to  soil  gas   flow,  especially when moist.    The
permeability of  this  fill may have been  poorer  than that of the  native soil
underlying the aggregate under the Ohio and Maryland houses,  since the fill is
specifically selected and compacted to support the elevated slabs characteristic
of Florida slab-on-grade houses.  The Florida houses varied in  size between 1,500
and 2,570 ft2 (140 and 240 ma),  had  either block or  poured concrete foundation
walls, and did not have forced-air ducts underneath the slab.  Several had sunken
living rooms.  Performance of interior and exterior SSV in these houses has been
far less effective than it was in the Ohio and Maryland houses.  In the Florida
houses,  two  or  three  suction  pipes  have  been  necessary   to  reduce  radon
concentrations to post-mitigation levels of 3 to 9 pCi/L (generally 70 to 90+%
reduction), compared to  the  81  to 99%  reduction  achieved with only one or two
pipes at the other sites.  The two or three suction pipes were  needed  in Florida
despite efforts to improve performance by excavating pits underneath the slabs
where the pipes penetrated, and  by using high-suction fans  (capable of suctions
up  to 33 in.  HaO,  or 8,100  Pa).   In general,  exterior SSV  gave  comparable
reductions to interior SSV in Florida,  although,  in  one house,  exterior SSV was
significantly  less effective.   These  interior  vs.  exterior results  appear
generally comparable to those from Ohio (where the two approaches also generally
compared well except in one house).  However, in Florida, performance of exterior
systems was reduced by  high air leakage rates into  the system,  a problem far less
pronounced in Ohio.  In the Florida houses, the exterior suction pipes were often
inserted horizontally some distance toward the  interior of the slab in order to
reduce leakage, making  these exterior systems more similar to  interior SSV.

      It would appear that major obstructions such as sub-slab supply ducts (in
the Ohio and Maryland houses) can fairly readily be handled by SSV systems, as
long  as  there is  a  good aggregate layer.   But  where there is no aggregate and
where the soil (or fill) under the slab is not reasonably permeable, design of
the SSV system can be made more  difficult,  even when there  are  no apparent sub-
slab  obstructions, as  indicated by the experience in Florida.

-------
                               ACKNOWLEDGEMENTS

      The authors wish to express their appreciation to:  the homeowners who made
their homes available for this testing;  the Dayton Regional Air Pollution Control
Agency, which provided substantial  support  during the selection of the houses
for testing; the  Ohio Department  of Health,  which provided support throughout
the project; and Air Chek, Inc., which contacted candidate homeowners prior to
the house selection process.
                                  REFERENCES

1.    Henschel, D. B.,  A.  G.  Scott, W. 0. Findlay, and A. Robertson, Testing of
      indoor radon  reduction methods in 16  houses  around Dayton, Ohio.   In:
      Proceedings:  The 1988  Symposium on Radon and Radon Reduction Technology.
      Vol. 1.  Symposium Oral Papers.  EPA-600/9-89-006a, U.  S.  Environmental
      Protection Agency, Research Triangle Park, NC, and Washington, D. C., March
      1989.  746 pp.  (NTIS No. PB89-167480)

2.    Findlay,  W. 0., A.  Robertson,  and A. G. Scott, Testing of indoor radon
      reduction techniques in central Ohio  houses:  Phase 1 (Winter 1987-1988).
      EPA-600/8-89-071, U. S. Environmental Protection Agency,  Research Triangle
      Park, NC. July 1989.  315 pp.   (NTIS No. PB89-219984)

3.    Fowler,  C.  S.,  A.  D.  Williamson, B. E. Pyle, F.  E.  Belzer III,  D.  C.
      Sanchez,  and T.  Brennan, Sub-slab depressurization demonstration in Polk
      County,  Florida,  slab-on-grade houses.    In:   Proceedings:   The  1988
      Symposium on Radon and Radon Reduction Technology.  Vol.  1. Symposium Oral
      Papers.  EPA-600/9-89-006a, U. S. Environmental Protection Agency, Research
      Triangle  Park, NC, and  Washington, D. C., March 1989.  746 pp.  (NTIS No.
      PB89-167480)

-------
                   TABLE  1.  SUMMARY OF  SLAB-ON-GRADE HOUSES TESTED DURING PHASES  1 AND 2
                            OF THE RADON REDUCTION FIELD  PROJECT  IN DAYTON, OHIO
House
ID Ho.
 15°
 31"

 47"

 58
 70
 72
 77
 85
Pre-mitigation
Radon (pCi/Ll*
                   Slab size
17.5
20.9
16.3
12.6
14.0
22.9
24.1
21.6
15.4
17.9
16.0
10.3
(kitchen)
(bedroom)
(bedroom)
(kitchen)
(bedroom)
(kitchen)
(bedroom)
(liv.  area)
(liv.  area)
(liv.  area)
(liv.  area)
(liv.  area)
1,700

  950
  950

1,100

1,400
2,350
1,700
2,600
1,100
Foundation
 Material

   Block

  Poured
  Poured

  Poured

  Poured
   Block
   Block
   Block
   Block
Sub-slab
 Ducts?

 Yes

 Yes
 Yes

 Yes

 Yes
  No
  No
 Yes
 Yes
 Duct
Config.

 Radial

 Radial
 Radial

 Radial

 Radial
 Radial
 Radial
Sunken
 LR?

   No

   No
   No

   No

   No
   No
   No
  Yes
   No
      Pre-mitigation results are from  alpha-track detectors  exposed during the period December 19B7  -
      March 1988 (Phase 1 houses)  or November  1988  -  January  1989  (Phase 2 houses), prior  to completion
      of mitigation systems.  Detectors exposed in clusters of two (Phase  2)  or  three  (Phase 1); figures
      shown here are averages.

      Houses initially tested during Phase 1.
Conversion factors:  Bq/m1  = 37  x value  in  pCi/L.  Area in ma = 0.093 x value in ft2.

-------
                   TABLE 2.  BEST PERFORMANCE OF SUB-SLAB SUCTION IN SLAB-ON-GRADE HOUSES
                                                 IN DAYTON
Interior/
House
_MJ.
1
15'

31

47a
58

70

72
77

85

Exterior
SSV
Exterior
Exterior
Exterior
Exterior*
Interior
Exterior
Exterior
Interior
Exterior
Interior
Interior
Exterior
Interior
Exterior
Exterior
Radon Concentrations (oCi/L)
System Off


(Site
(Site

(#D£




(#1 &

(#1 &
(#1 &
(Site
(Site


Db
2)






2)"

2}'
2)«
1)
2)
Range
14-32
6-42
If
7-27
n
20-39
19-43
rr
12-A4
n
12-30
9-23
n
3-34
n
Mean
20
17
11
14
II
29
27
tl
23
II
22
15
II
12
II
.2
.9

.7

.6
.5

.5

.1
.7

.9

System
Range
2-
0-
0-
0-
0-
0-
0-
0-
2-
0-0
1-
1-
1-
1-
0-
6
1
1
1
1
2
1
2
5
.5
3
4
3
5
5
On

Mean
2.
0.
0.
0.
0.
0.
0.
0.
3.
0.
1.
1.
1.
2.
1.
8
4
5
1
3
9
5
8
7
2
6
7
7
5
5
% Reduction
in Mean
86
98
97
99
98
97
98
97
84
99
93
89
89
81
88
Suction
fin. H,0}
N/Ae
1.4
1.2
0.85
0.6
1.2
0.55
0.7
N/A
0.5-0.85
1.2
0.25-0.3
0.65-0.8
1.2
1.0
Flow
(cfm)
100
N/A
43
64
69
69
140
63
111
104
12
114
148
45
42
        Testing completed during  Phase  1;  these results have been reported  in References 1 and 2.
        Sites  1 and 2 refer to different positions around the house where the one exterior suction point
        was located.
        #1 and #2  refer  to individual suction points  in homes having  two-point systems installed.
        Two suction points operating simultaneously to achieve these best results.  In all other cases,
        only one point is operating.
        N/A -  not  available.
        Fan operating at high speed  in  all cases.
Conversion factors:
        Bq/m3 - 37 x value in pCi/L.
Pa = 248 x value in in. H,0.
L/s •= 0.47 x value in cfm.

-------
    TABLE 3.   EFFECT OF ONE VS.  TWO SUCTION POINTS IN  SSV  SYSTEMS  IN  SLAB-ON-GRADE HOUSES


        Interior/       Mean      	Mean  Performance After	
House   Exterior        pCi/L       Point #1 Only      Point #2 Only          Both Points
 No.      SSV          Before     pCi/L   % Red.      pCi/L   % Red.       pCi/L     % Red.

 31     Interior        14.7        0.3      98        2.7      82          1.0         93
 70     Interior        23.5        3.0      87        1.8      92          0.2         99
 77     Interior        15.7        6.7      57        7.7      51          1.7         89
        Exterior        15.7        6.0      62        5.7      64          1.7         89


Locations of suction points:
  House 31: #1  in furnace  room (center of house);  #2 in bath (center  rear of house).
  House 70: #1  in furnace  room (center); #2 in bedroom closet (in right front quadrant).
  House 77: interior  -  #1 in laundry room  (center of  left half of house); #2 in
           bedroom closet (center of right half).
           exterior  -  #1  on rear wall,  center of left half of house; #2  on  rear
           wall,  center of  right half.

    Fan at high  speed  in all cases  reported  here (best results).   Both suction points
    connect to a single fan, except for exterior  system in House 77, where each  exterior
    point has its own fan.

Conversion factor:  Bq/mJ  -  37 x value in pCi/L

-------
                     TABLE ft.   EFFECT OF FAN CAPACITY ON THE PERFORMANCE OF SSV SYTEMS
                                           (SLAB-ON-GRADE HOUSES)

House
No.
1
151

31



47'
58

70



72
77





85



Interior/
Exterior
SSV
Exterior
Exterior
(Site 1)
Exterior'
Interior
Interior
Interior
Exterior
Exterior
Interior
Exterior
Interior
Interior
Interior
Interior
Exterior
Exterior
Exterior
Interior
Interior
Interior
Exterior
(Site 1)
Exterior
(Site 2)
Suction
Point
Number
0

b
D
#1
#2
#1 & 2
0
B
»
b
#1
#2
#1 & 2
b
#1
#2
#1 & 2
#1
#2
#1 & 2

b

&
Mean
pCi/L
Before
20.2

17.9
14.7
tl
it
11
29.6
27.5
"
23.5
II
II
11
22.1
15.7
n
11
11
n
11

12.9

n
Mean Performance After
Hi eh
pCi/L
2.8

0.4
0.1
0.3
2.7
1.0
0.9
0.5
0.8
3.7
3.0
1.8
0.2
1.6
6.0
5.7
1.7
6.7
7.7
1.7

2.5

1.5
Capacity
% Red.
86

98
99
98
82
93
97
98
97
84
87
92
99
93
62
64
89
57
51
89

81

88
Low Cacacitv
oCi/L
10.1

0.2
1.4
0.3
13.7

2.9
1.0
2.2
6.0
6.5
2.0
1.7
3.2


2.1

9.7
5.5

2.5

2.8
% Red.
50

99
90
98
7

90
96
92
74
72
91
93
86


87

38
65

81

78
Suction
Highc
N/A

1.4
0.85
0.6
0.8
0.4-0.55
1.2
0.55
0.7
N/A
0.75
0.95
0.5-0.85
1.2
N/A
N/A
0.65-0.8
0.4
0.35
0.25-0.3

1.2

1.0
fin. H..O)
Lovc
N/A"

0.8
0.5
0.3
0.1

0.45
0.3
N/A
N/A
N/A
0.35
—
0.75


0.4-0.8

0.06
N/A

0.35

0.5
Flow
Hlghc
100

N/A
64
69
49
80
69
140
63
111
39
74
104
12
98
125
148
96
83
114

45

42
rcfm)
Lowc
• J-—
37

N/A
N/A
33
14

25
82
23
64
N/A
32
	
9


132

24
N/A

16

27
*  Testing completed  during Phase  1;  these  results have  been reported in References 1 and 2.
"  Only one suction point  was  included  in these  systems.
c  Suctions,  flows  in system piping shown for both high  and low fan capacities.
"  N/A = not  available.
Conversion factors:   Bq/mJ = 37 x  value  in  pCi/L. Pa  = 248 x value in in. H,0.        L/s = 0.47 x value in cfm.

-------
                TABLE 5.  EFFECT OF PRESSURIZATION VS.  SUCTION IN SSV SYSTEMS
                                    (SLAB-ON-GRADE HOUSES)

                      Average Radon Concentration fpCi/L)
                                                                    % Reduction
Post-Mitigation
Pre-Mitigation
17.9
14.7
29.6
27.5
23.5
15.7
12.9
Suction
0.4
0.4
2.9
0.5
0.2
1.7
1.6
Pressure
1.8
4.2
16.0
5.7
13.4
3.1
2.6
                                                                 Suction Pressure

     15'                17.9                  0.4     1.8            98      90
     31"                14.7                  0.4     4.2            97      71
                                                                    90      46
     58                27.5                  0.5     5.7            98      79
     70"                23.5                  0.2    13.4            99      43
     77*                15.7                  1.7     3.1            89      80
     85"                12.9                  1.6     2.6            88      80


 '   Testing completed during Phase 1;  these results  reported in References  1 and 2.

 "   Performance shown is  for a two-point interior  SSV system.   All  other results reported
    here  are for exterior SSV systems.
 c   Results obtained with the two-point exterior system.
 d   Results obtained with one-point exterior system,  fan located at Site 2.

Conversion factor:  Bq/m1 - 37  x value  in pCi/L.

-------
                                                                        VII-2
                    RADON MITIGATION EXPERIENCE IN HOUSES
                  WITH BASEMENTS AND ADJOINING CRAWL SPACES
                  by:    Marc Messing
                        INFILTEC
                        Falls Church,  VA 22041

                        D.  Bruce Henschel
                        Air and Energy Engineering Research Laboratory
                        U.S. Environmental Protection Agency
                        Research Triangle Park,  NC 27711
                                  ABSTRACT

      Active soil depressurization systems were installed in four basement houses
with adjoining crawl spaces in Maryland.  In addition, existing soil depressur-
ization systems were modified in two additional basement-plus-crawl-space houses.
These six  houses were  selected  to include both  good and  poor  communication
beneath the basement slab, and different degrees of importance of the crawl space
as a source of the indoor radon.  The radon reduction effectiveness was compared
for:  depressurization only under the basement slab; depressurization only under
a  polyethylene  liner  over  the  unpaved crawl-space  floor; and  simultaneous
depressurization under  both  the basement slab and the crawl-space liner.   The
objective of this testing was to identify under what conditions treatment of the
basement alone might  provide sufficient radon reductions  in houses  of  this
substructure,  and what  incremental benefits might be achieved by also treating
the crawl space.

      The results suggest  that,  when there  is  excellent communication beneath
the basement slab, basement-plus-crawl-space houses can sometimes be treated by
basement sub-slab depressurization  (SSD) alone, even when radon levels in the
crawl space indicate that the crawl space could  be  a source.   With excellent
communication, the  incremental benefit  of including sub-liner depressurization
in  the  crawl  space, in addition  to basement  SSD,  can  sometimes  be limited
(although the  two in conjunction may still be needed  if the goal is to achieve
"ambient" levels indoors).  However, when communication beneath the basement slab
is  poor,  and if  the  crawl  space  is  an  important  source,   then sub-liner
depressurization can provide potentially significant additional radon reductions
compared to basement SSD alone.

      This  paper has been reviewed in accordance  with the U. S. Environmental
Protection  Agency's  peer  and  administrative  review  policies,  and  has  been
approved for  presentation  and publication.

-------
                                 INTRODUCTION

      Twelve existing houses in Maryland were selected to study cost-effective
radon mitigation methods in three types of houses:   five houses with basements
and adjoining  slabs  on grade; six houses  with basements and  adjoining crawl
spaces; and one house  which was  slab on grade with heating  ducts  beneath the
slab.    All of these  houses  had previously reported radon levels between 10 and
70 pCi/L (370 and 2600 Bq/m3)*.

      Active soil depressurization systems were installed in all of the houses.
For the basement houses with adjoining wings, one sub-slab depressurization (BSD)
pipe was  installed beneath the basement slab, and a second suction  pipe was
installed either beneath  the adjoining slab or  through  polyethylene  sheeting
placed over the floor of the adjoining crawl space.  Radon concentrations were
measured in the basement and on the first floor with:  none of the suction pipes
operating; SSD applied to the basement slab only; depressurization beneath the
crawl-space liner or  beneath the adjoining slab on grade only; and depressuriza-
tion beneath both the lower slab and the crawl-space liner or upper slab.  The
objective of this  testing was to identify under what conditions  treatment of the
basement alone might  provide sufficient radon reductions in such basement-plus-
adjoining-wing houses,  and what additional reductions might be achieved by also
treating the adjoining wing.   Diagnostic testing  (including sub-slab or sub-
liner radon concentrations,  and sub-slab suction field extensions and air flows)
was conducted to help understand the conditions influencing the performance of
these systems.

      This paper  addresses the  testing on the  six  basement-plus-crawl-space
houses.  These six houses  included four which had had no prior radon mitigation
system, and two which had existing systems which  were not fully effective.
                             MEASUREMENT METHODS

INDOOR RADON MEASUREMENTS

      For short-term measurements of  radon mitigation system performance, radon
gas was measured using a Pylon Model AB-5 continuous radon monitor equipped with
either a 17.4 in.3 (285 mL) Lucas scintillation cell or a  PRD-1  passive radon
detector.  The Pylons were programmed  to measure  radon hourly.   One Pylon was
placed  in  the basement,  and  one in  the  living  area  above the  crawl  space,
generally for about 2 weeks; during  this time,  the mitigation  system would be
cycled through its  range of operating conditions (system off, basement treatment
only, crawl-space treatment only, treatment of both wings), with  the system being
left at a given condition for at least 48  hours.

      For  a  long-term  measurement  of SSD  performance,  3-month  alpha-track
detector measurements are being  repeated for at least  1 year with the systems
operating.   These long-term results are not covered in this paper.
      Radon in Bq/m3 - 37  x value  in pCi/L.

-------
DIAGNOSTIC TESTING

      In selecting houses for this study, it was desired to cover a range of the
key variables,  including the communication under  the basement slab,  and the
relative  importance  of the  basement  vs.  the crawl  space as a  radon source.
Knowledge of these parameters is also important for understanding the conditions
under which basement-only treatment is sufficient,  and what additional benefit
can be derived by also treating the crawl space.

Sub-slab Pressure and Flow Measurements

      To characterize communication under the basement slab, sub-slab pressure
and flow measurements were  taken with a "Blower Floor,"  consisting of a high-
suction vacuum cleaner, a calibrated  flow meter, and a digital micromanometer.

      Where a sump existed,  the  sump cover was removed (if feasible) and a plate
with a  1.5-in.  (4-cm)  opening was secured over  the  sump  with  duct tape.   The
vacuum cleaner nozzle was  sealed  into the 1.5-in.  opening to depressurize the
sump.  Measurements of induced air flow in the vacuum were taken as the suction
in the sump was measured using  the micromanometer.  The suction underneath the
basement  slab  was also measured through  several 0.25-in. (0.6-cm)  test holes
drilled through the slab at  selected locations, usually in the basement corners.

      In  houses  with no  sump,  a 0.5-in.  (1.25-cm)  suction hole  was drilled
through the  slab  (at a  location which appeared promising for the ultimate SSD
suction point).   Suction was drawn  on the sub-slab through this hole with the
vacuum cleaner.  Flow in the vacuum cleaner was measured, as was  the differential
pressure across the slab through the 0.25-in. (0.6-cm) test holes.  One test hole
was  located  4  in. (10 cm)  from the suction hole,  and several  test holes were
usually located in the  basement corners.

Radon Measurements Under  Basement Slab

      To  assess   the  source  strength  under  the  basement,  sub-slab   radon
measurements were taken through 0.25-in.  (0.6 cm) holes drilled in each corner
of the basement slab area.  Sub-slab gas was pumped through a 17.4 in.3 (285 mL)
Lucas cell at approximately  0.03 cfm  (1 L/min),  and photomultiplier counts were
recorded  at  1-minute intervals.  Although  this "sniffing" technique does not
allow  the radon to reach equilibrium with its decay products, the sensitivity
of this non-equilibrium measurement system  was  determined experimentally  to  be
about  1.4 times the sensitivity of an equilibrium situation;  this sensitivity
was  used  to  estimate sub-slab values.

Radon Measurements in Crawl-Space Soil

      As  a rough estimate  of  the  source strength  in  the crawl space,  radon
measurements were made near the  surface of  the crawl-space soil.    Where  the
crawl-space  floor was covered with a  preexisting vapor barrier, a 0.25-in. (0.6
cm)  hole was made through the vapor barrier  (generally within arm's reach of the
access  door to the  crawl  space), and sub-barrier  samples were taken with  the
sniffing  technique described above.   Where  there was no  existing  vapor barrier

-------
in the crawl space, the air sampling tube was inserted approximately 6 in. into
the soil,  or  into the aggregate on  top of the soil, to collect  the  soil gas
sample. Generally, no measurements were taken of radon levels in the breathable
area of the crawl  space.

Post-Mitigation Diagnostics

      To define the operation of the SSD system  after  installation, measurements
were made  of  the  suction  and flows  in the  system  piping,  and the extension of
the suction field  created under the basement slab by the system.
                          DESCRIPTION OF STUDY HOUSES

      The six houses  included in this testing are summarized  in  Table  1.   In
addition to certain house  construction characteristics and the pre-mitigation
radon  concentrations,  the  table summarizes  results  from  the pre-mitigation
diagnostic testing.

      In all cases, the crawl-space floor consisted of bare soil (often with an
existing vapor barrier) or of aggregate on top of soil  {with no vapor barrier).
Only  one  house  (House 208)  was a  split level,  with a half-depth  basement
adjoining a relatively low crawl space; the remainder of the houses were ranches.
The basements were somewhat larger in floor area than were the  crawl spaces, by
factors ranging from 1.05 to 2.2.

      Two of the houses  (Houses  208  and 277)  already  had operating mitigation
systems which had reduced radon levels to the  range of  4  - 10 pCi/L at the time
this work began.  House 208 had a one-pipe basement SSD system.  House 277 had
a one-pipe basement SSD system tied into a sub-liner depressurization system in
the crawl space, connected to a single fan.

      House 208 is a  split  level  house.   The  lower half is approximately 4 ft
below grade in the  rear.  The upper level is above  a crawl space which is at the
same level as the basement slab.   The crawl space and upper level are separated
from the lower level by the block foundation wall and an access door.  There is
no sump.  The  total  floor area of the basement slab and the  crawl space (the
"house  footprint")  is 1375 ftz (130 m2) ,  with a basement/crawl area  ratio of
1.05.

      House 277 is a  ranch-style  house built  over an  L-shaped basement and an
interlocking L-shaped crawl space.  The basement is unfinished  and has no sump.
The crawl space  is at the same grade  as  the  basement  floor.   The crawl-space
floor is uneven and has been covered (as  part  of the earlier mitigation effort)
with a sealed plastic barrier.  The foundation walls are hollow block.

      No data are available  on pre-mitigation sub-slab  radon levels in these two
houses.  As  indicated in Table 1, both had poor airflow beneath  the basement
slab.  As part  of  this project,  additional basement SSD pipes were installed;
these pipes improved the distribution of the suction field beneath the basement

-------
slab and resulted in effective mitigation, as discussed later, greatly improving
the performance of the original installations.

      House 582  has  a basement which  is  divided into two parts:   a finished
family room; and  an unfinished utility room with washer/dryer,  furnace, and sump.
The crawl space  is open to both parts of the basement.  The house has a garage
on grade which is adjacent to the  crawl space, and there are living spaces over
the basement and crawl space, but not over the garage.  This house has excellent
sub-slab air  flow.   Of particular interest,  higher radon concentrations were
measured in the crawl-space soil than were found beneath the basement slab, as
shown in Table 1, suggesting that the crawl space might be a major contributor
to the indoor radon.   The total footprint is 825 ft2 (77 m2),  and the basement/
crawl area ratio is 1.8.

      Houses 1165 and 1248 were both characterized by relatively high sub-slab
radon levels  for this  area  (>1,000  pCi/L)  and  excellent sub-slab  flow.   By
comparison, radon levels in the crawl-space soil  were  only moderate, suggesting
that the basement may be the primary source. House 1165 has a basement, of which
approximately 80% is finished living space and the remainder is a utility/work
room.  There  is no sump.  There is  also  a large crawl space  open  to the basement,
and a garage on grade  adjacent  to the crawl space.  There are living areas above
both the basement and the crawl space,  although  there is no living space above
the garage.   There  are both duct and  plumbing  penetrations  between the crawl
space and  the living  area above it,  and there is a full bath in the basement.
The footprint is 1375 ft2 (130 m2), and  the area  ratio 1.7.

      House 1248 has a relatively  small basement  with  poured concrete walls and
an adjoining crawl space which is open to the basement. There are living spaces
above both the crawl space and the basement, with forced air ducts in both the
basement and  the crawl  space.   The basement is  a deep basement (rather than a
walkout),  and there  is a sump on the  side of the  basement opposite the crawl
space. There are no adjacent  slabs on grade.  With a footprint of 900  ft2 (84 m2)
and an area ratio of 1.8, House 1248 is approximately 65% of the size of House
1165 (both in the area of the basement  slab and  in the area of the crawl space
surface).

      House 1357  is a brick and block colonial,  with  a garage on grade, a deep
basement, and a crawl  space between the  garage and the  basement.  The  foundation
walls are hollow cinder block.  The crawl space has an uneven surface of packed
soil and is open to the basement; there is no sump.   There is forced air heat,
and there  are several plumbing penetrations  through  the  basement  slab (for a
washing machine  and sink).   House 1357 was characterized by high radon levels
under the basement slab, and poor  sub-slab airflow;  in this instance, there was
no measurable sub-slab airflow.  Radon  levels in the  crawl-space soil were not
as high  as at most of the other houses; however, concentrations in  the crawl-
space air  were  once  measured at 25 pCi/L at a time when the levels  in the
basement were only  5  pCi/L,  suggesting that  the crawl  space  could still be a
contributor.  The house footprint is  875 ft2  (81  m2), and the area ratio is 2.2.

-------
                      DESCRIPTION OF REMEDIATION SYSTEMS

      In the four houses not having sumps,  one  or more SSD pipes were installed
through the basement slab.  Houses 1165 and 1357 each received one-pipe systems;
House 208 received an additional pipe and fan to supplement the previous system,
thus winding up with a two-pipe system,  each pipe having its own fan; House 277
received an additional two pipes  tied  into the previous system,  to  create a
three-pipe  system.  Typically, about 0.5 ft3 (15 L) of aggregate was excavated
from under  the  slab where  the pipes penetrated.   In the two  houses which had
sumps (Houses 582  and 1248),  the sump was  fitted  with a permanent radon sump
cover and a suction pipe.   In all  houses,  the  suction  pipe  was thin-walled 6-
in. (10-cm) PVC  pipe.   Except  for House 208 (which had two independent SSD pipe/
stack/fan systems and no crawl-space sub-liner system),  the SSD pipe(s) were T'd
together with the sub-liner suction pipe from the crawl space, and connected to
a single in-line fan mounted outside the house  or in the garage attic, with the
exhaust released above  the roof  in all  houses.   The fans had a rated capacity
of 180 cfm  (85 L/s) at  0.5 in. H20 (125 Pa).  All basement cracks were caulked
along accessible  floor/wall  joints (i.e.,  not behind finished walls  or other
inaccessible locations).

      Sub-liner depressurization systems were  installed in each  house  except
House 208.   Except in House 277, the liner  material was 8-mil  (0.2-mm) high-
durability polyethylene sheeting, laid to form  a barrier over the entire crawl-
space floor.  In  the  final  installations,  the  liner was  sealed by overlapping
the individual sheets (by 6 to 12 in. ,  or 0.15 to 0.3 m), and sealing these seams
with urethane caulk; the junctions of the sheeting with the perimeter foundation
wall and with interior piers were sealed in  the same manner.  A  single 4-in. (10-
cm) PVC suction pipe was inserted through the  plastic liner at one location, with
a plywood "flange" around the pipe  underneath the liner to support the pipe and
to lift the  polyethylene off  the soil  at that  location to  aid in distribution
of the suction field,   wherever possible, the measurements of mitigation system
performance  included  measurements with  and without the  sheeting  sealed around
the  perimeter wall,  to  evaluate   the  benefits of this  sometimes-difficult
installation step.

      In House 277, where the sub-liner  system had  been installed by an earlier
mitigator as part of the original mitigation effort, the liner was 20-mil (0.5-
mm) plastic, and a loop of perforated piping was  located around  the perimeter
of the crawl space under the plastic.   Suction was drawn on this  loop.   The
plastic was  already sealed at seams and around the perimeter,  so  that testing
without perimeter  sealing was not possible during this  project.
                                    RESULTS

      The  radon reduction  results  from  the testing  in the  six houses  are
summarized in Table 2.  These results represent at least 48 hours of measurements
at each test condition, using the continuous Pylon monitors.

-------
 EFFECTS OF  TREATING THE  BASEMENT  SLAB ONLY

      Applying a vacuum beneath only the basement slab reduced the radon levels
 both in the basement and in the living area above the crawl  space in  all of the
 houses except House 1357.  With the exception of  that house,  reductions in the
 basement were consistently high  (82 to 98%), reducing concentrations to below
 1 pCi/L in  these  short-term  measurements.   The one house having less than 90%
 reduction in the basement (House 582) had the relatively low percentage reduction
 of  82%, at  least  in part because  the pre-mitigation value  was below 4 pCi/L.
 (In all of the  houses, pre-mitigation  values  at  the  time of  these  Pylon
 measurements were much lower  than they had been when  the earlier, cold-weather
 charcoal detector measurements had been made, even though the Pylon measurements
 were largely completed during the month of February.)  In the houses  other than
 1357 for which Pylon measurements  were successfully completed in the living area
 above the crawl space, radon  reductions were about 80% in the  living area with
 basement-only treatment, reducing concentrations  to  1.2 pCi/L and less.

      In House  582  --  where  the  crawl space appears to  be  an important radon
 source but  where  sub-slab  communication is  excellent, as discussed  earlier --
 basement-only treatment did about as well in reducing levels upstairs as it did
 in  House 1165,  where  the communication is also  excellent but where  the crawl
 space seems less of a source.   Basement-only treatment in House 582 appeared only
 slightly poorer in reducing levels in the basement, compared to Houses 1165 and
 1248 (also good communication, crawl space less of a  source).   Based  upon these
 fairly good results in House  582,  it would appear that -- when communication is
 very good -- basement-only  treatment may be sufficient even when the crawl space
 is  an important radon source.

      In House 1357, SSD under the basement slab  reduced basement levels about
 40%, much less  than in the other  houses.   Concentrations  in the  living space
 above the crawl  actually  appeared to increase from 2.5 to 3.5  pCi/L when basement
 SSD was applied (although, given  the variability of radon measurements, this 1
 pCi/L differential is so small that it is not certain that any real  change had
 occurred at all).   This relatively poor  performance in House  1357 with basement
 SSD only might intuitively  be  attributed to the  very poor  sub-slab communication
 (and possibly the relative importance of the crawl space as a  radon  source, as
 suggested by the 25  pCi/L once measured in the crawl-space air). House 208 also
 had poor sub-slab communication (although not as  poor as  House  1357,  which gave
 no  airflow  during the  pre-mitigation  diagnostic  testing); House 208  had still
 had elevated radon  levels  with its original  one-pipe SSD system,  and achieved
 the high removals shown  in Table 2 only because of the additional suction pipe
 and fan added as part of this project.

 EFFECTS OF TREATING CRAWL SPACE ONLY

 Sub-Liner Depressurizatton with the Liner Sealed

      As shown  in Table  2,  sub-liner depressurization  in   the crawl  spaces,
without any treatment of the  basement, effectively reduced radon levels in both
 the basements and  the living areas of all five test houses  in  which this approach
was tested.   (House  208 did not receive crawl-space treatment;  House 277 was not

-------
tested in the crawl-space-only mode.)  The  results  shown in the table are for
the  case where  the  liner was  sealed  at  seams  and  around  the  perimeter.
Reductions ranged  from 38  to 73% in  the  basement,  and from 15  to  88%  in the
living area  above  the crawl space.   Except for House  1357,  reductions  in the
basement from crawl-space-only  treatment were not as  great  as  the reductions
achieved by basement-only treatment, as might be expected.  For the two houses
other than 1357 where measurements were successfully completed in the  living area
with both mitigation approaches, crawl-space-only  treatment  gave living-area
radon reductions which were comparable to basement-only treatment in one house
(House 582,  88 vs.  84%), and substantially poorer in the other house (House 1165,
15 vs. 82%).  The  fact that crawl-space-only  treatment is so much poorer than
basement-only treatment in House 1165  suggests that,  in this house, either:  a)
most of  the  radon in  the  living area was probably  flowing upstairs  from the
basement; or b) the basement SSD pipe was being very effective in  achieving "site
ventilation" (i.e.,  in extending its  suction  field underneath the crawl-space
soil).

      Crawl-space-only treatment in  House 1357 gave significantly  greater radon
reductions in the basement  than did  basement-only  treatment  (73 vs. 41%).  This
result would appear to be consistent with the  expectations that  the crawl space
is  an important source of  the  radon in the house,  and that  poor  sub-slab
communication would  reduce  the effectiveness  of a one-point SSD system in the
basement.  The levels  in the living area above the crawl space were reduced to
1.3  pCi/L  with  the  crawl-space-only  system,  a 48%  reduction  from  the  pre-
mitigation level of  2.5 pCi/L;  the  basement-only  system had given no apparent
reduction in the living area.

Sub-Liner Depressurization without Sealing the Liner

      The results  discussed above are for the final sub-liner depressurization
installation in each  house,  which involved sealing the liner  at seams and around
the perimeter.   Prior  to that complete sealing, the sub-liner depressurization
system was tested without the liner's being sealed around the perimeter (although
it was still sealed  between the  sheets of polyethylene).   Testing without the
liner sealed was conducted only for the crawl-space-only case, not in conjunction
with basement SSD.  Comparison of the  radon reductions  achieved with the crawl-
space-only systems  before and after liner sealing indicates that sealing improved
reductions in all four houses where before/after measurements could be made, both
in the basement and upstairs.   Basement reductions increased from  a pre-sealing
range of 0 to 60%,  to  the post-sealing range of 38 to 73%; upstairs reductions
increased from 0 - 56% to 15 - 88%.

      It is  emphasized  that  --  although the differences  in  the  percentage
reductions may appear  significant --  these  differences often  represent only a
few pCi/L.   Given  the  variability in  indoor radon levels with time, it is not
clear if the apparent differences reflect a real effect of liner sealing.

EFFECTS OF TREATING BOTH THE BASEMENT AND THE CRAWL SPACE

      As shown  in Table 2, simultaneous  SSD in  the  basement  and sub-liner
depressurization in  the crawl  space  was clearly  superior to basement-only

-------
treatment in House 1357.  Radon reductions both in the basement and in the living
area above the crawl space are significantly higher with the combined treatment
compared  to  basement-only treatment  (90 vs. 41% in  the  basement, 84  vs.  0%
upstairs).   Again,  this  result is  consistent with  the  diagnostic  results
indicating that  sub-slab  communication is poor,  and suggesting that the crawl
space might  be  an important source in that  house.   In one other house (House
582), which had excellent sub-slab communication, simultaneous  treatment appears
somewhat superior to basement-only treatment (92 vs.  82% in the basement, 96 vs.
84% above the crawl space).  However,  the differential  between the radon levels
for simultaneous vs. basement-only treatment is so small in House 582 (fractions
of a pCi/L),  that measurement uncertainty and temporal variation in radon levels
could easily explain the differences;  thus, it is not clear that there is truly
a significant difference in the performance of the two approaches  in  this house.
In the other two houses for which complete data are available  (Houses 1165 and
1248) -- both of which had excellent  communication  --  basement-only treatment
appears at least as effective  as simultaneous treatment in reducing the radon
levels both  in the basement and upstairs.

      In summary, in the one house where sub-slab communication is poor and the
crawl space  might  be  an  important  source,  treatment of both  wings  appears
necessary. (Additional  suction pipes  in the basement might also  have improved
the  performance of  the basement-only  system;  post-mitigation  suction field
extension measurements in this house confirm that the suction  was not extending
underneath the  entire  slab.)  In the three houses having excellent sub-slab
communication, treatment of  the  crawl space appears  either unnecessary, or of
reduced incremental value.

      Simultaneous  treatment of  the  basement  and  the  crawl  space  was  always
superior to crawl-space-only treatment, providing greater reductions  both in the
basement and upstairs.  Reductions of 38 to 73% in the basement  with crawl-space-
only treatment  increased  to  90 to 96% with simultaneous treatment;  reductions
of 15 to 88% in  the living area  above the crawl space  increased  to 84 to 96%.
Therefore --  although crawl-space-only treatment was  sufficient to reduce levels
in these houses  below 4 pCi/L on both floors, due to the low pre-mitigation
values --it  appears always beneficial  to supplement a sub-liner depressurization
with SSD  in the basement,  regardless  of  the  sub-slab communication  or  the
importance of the crawl space as a source.

      In House 277  --  which  has  poor  communication  --  the three-pipe basement
SSD  system,  combined  with  the   sub-liner depressurization  system having  a
perforated piping  loop under the liner, is  the most extensive of the systems
tested under  this project.  Yet this system gave the lowest percentage reductions
(82% in the basement, 84% upstairs) and the highest residual radon levels (1.7
pCi/L in the basement,  0.7 pCi/L upstairs) of any of the  basement-plus-crawl-
space systems.   Even though the  residual levels are less  than 2  pCi/L, these
results suggest  that,  with poor-communication houses, more extensive systems
(e.g., multi-pipe basement  SSD installations, crawl-space sub-liner systems with
perforated piping under the liner) will sometimes be needed.  The system at House
277 was not  tested  in  the  basement-only or the crawl-space-only  mode,  so that
it is not  possible to assess  the relative contributions   of  the  basement  and
crawl-space components of the system.

-------
                                  CONCLUSIONS

      The four conclusions listed below are based  on  the  results from the six
basement-plus-crawl-space houses tested in this study.  These conclusions must
be  qualified,  recognizing that:   the  number of  houses  tested was  limited,
considering the  range of house variables  to  be tested  (e.g.,  communication,
importance of crawl space as  a source); and the pre-mitigation radon concentra-
tions existing at the time of testing  (February  1989  in  most cases)  were low,
so that changes in the radon concentrations caused by the mitigation systems were
small  (in  terms  of  pCi/L)  and within  the  background variability created  by
temporal radon variations and measurement uncertainty.

1)    When communication  under  the  basement slab  is  excellent  (as defined  in
      Table 1),  and  when the crawl  space is not a primary  contributor to the
      indoor radon levels:   basement-plus-crawl-space houses can sometimes  be
      treated by basement SSD alone.  Under these circumstances, the incremental
      benefit of including sub-liner depressurization in the crawl  space,  in
      addition to basement SSD, may  sometimes be limited.   This result can  be
      true even where diagnostic measurements indicate radon  levels in the crawl
      space several  times higher  than  those in  the living areas of  the house.
      However, even in houses with excellent communication, basement SSD may have
      to be supplemented with sub-liner treatment in order to achieve "ambient"
      radon concentrations indoors (i.e.,  a fraction of a pCi/L).

2)    When communication  under the basement slab is poor  (as defined in Table
      1), and when the  crawl space  might  be  a primary contributor  to indoor
      radon:  sub-liner  depressurization in the crawl space,  in conjunction with
      basement  SSD,   can  provide potentially  significant  additional  radon
      reductions compared to basement SSD alone.  This study did not definitively
      address  the potential  for  achieving these  increased radon  reductions
      through an improved design of  the basement SSD system  (e.g., an increased
      number of suction pipes), rather than by adding sub-liner depressurization.

3)    Combined treatment  of  the  basement and the  crawl  space  always provided
      greater radon  reductions  in this  study than did  crawl-space  treatment
      alone.  Thus,  it  would not usually be desirable to install a  sub-liner
      depressurization system in the crawl  space without also tying in basement
      SSD pipes.   In  most  cases, crawl-space treatment  alone  is not as effective
      as is basement  treatment alone; only in one house, which had poor sub-slab
      communication  in the basement and high radon concentrations in the crawl
      space,  did sub-liner  depressurization  in  the  crawl  space prove  more
      effective  in reducing  indoor radon levels than did SSD  applied to the
      basement slab.

4)    When applying  sub-liner depressurization in a crawl space, with a single
      suction pipe penetrating  the liner  at  a central location, sealing the
      polyethylene sheeting  to the crawl space perimeter  walls  (as well as  at
      seams between  sheets)  appeared to provide  roughly  a 10 to 40  percentage
      point improvement over the reductions  achieved with the liner sealed only
      at seams between sheets.

-------
      If these conclusions from the six study houses can be extrapolated to the
general housing population,  they  are  relevant  to  several  different aspects of
commercial radon remediation.  First,  they clearly suggest that simple sub-slab
vacuums applied to  the lowest levels of basement houses  with adjoining crawl
space are likely to reduce the radon levels substantially, especially in houses
with good sub-slab flow beneath the basement slab.  Therefore  -- where the goal
of remediation is to reduce  radon levels below 4  pCi/L or to achieve the most
cost-effective remediation (as has commonly been the case) -- then SSD beneath
the basement slab is likely to be an appropriate approach.

      However, where the  goal of remediation  is to achieve  ambient  levels (a
fraction of a pCi/L) indoors, these findings suggest that, even under the best
of conditions, it  may sometimes be necessary to supplement  SSD  systems with sub-
liner depressurization of exposed crawl space soil.  Since installation of sub-
liner systems  is  generally difficult and labor intensive,  adding crawl-space
treatment to  a basement  SSD  system may be expected  to increase  the  cost of
remediation by more than  100% in many instances.

      It  is  important  to note that  the  experience  with House  1357  clearly
illustrates that,  in some instances, the crawl  space may be a primary source of
indoor radon.  In such cases  -- and when,  in addition, communication under the
basement  slab  is  poor --  the likelihood  is  increased that a sub-slab vacuum
beneath  the  basement  slab  will  have   to   be   supplemented  by  sub-liner
depressurization,  even to achieve 4 pCi/L.

      For  the professional   diagnostician  and mitigator,   therefore,  it  is
important to utilize appropriate  diagnostic tests  to  assess  when basement SSD
by itself will potentially not be sufficient.   The  diagnostic tests used in this
study suggest that sub-slab flow beneath the basement slab, together with radon
measurements in the  crawl  space air or soil gas, may be  useful predictors of when
crawl-space treatment will also be desirable.

      It  is  re-emphasized that additional  testing -- on a  larger  number of
basement-plus-crawl-space houses,  with more adjustments to  the mitigation system
design/operating variables, and with higher pre-mitigation radon levels -- would
be necessary to confirm these conclusions.
                               ACKNOWLEDGEMENTS

      The authors wish to express their appreciation to:  the homeowners who made
their homes  available  for  this  testing;  and  the  Maryland Department  of the
Environment,  which provided substantial support during the selection of houses
for testing.

-------
   TABLE 1.  SUMMARY OF BASEMENT-PLUS-CRAWL-SPACE HOUSES TESTED IN MARYLAND
House
 No.

 208
 277
 582
1165
1248
1357
Foundation

  Block
  Block
 Poured
  Block
 Poured
  Block
     Pre-mit.
     Indoor Rn
      (pCi/D*

     10.2-23.2
      7.6-27.8
      9.8-17.3
     13.2-13.8
     25.3-41.8
     15.1-18.4
        Source Concentration (pCi/L)
           Sub-slab         Soil
          (Basement)       (Crawl)
            117- 225
              N/A
            138- 305
            193-1570
           1725-3674
            600-1800
                      N/A
                      N/A
                      888
                      144
                      374
                       67
                       Basement
                       Sub-slab
                       Airflow**

                         Poor
                         Poor
                       Excellent
                       Excellent
                       Excellent
                         Poor
*  Pre-mitigation  indoor  radon levels  as  determined by charcoal  detector
   measurements  in the basement during cold weather.   Measurements  made  by
   homeowner prior to EPA study, and by EPA during house selection process.
** "Poor" airflow  is  defined here  as less  than  30 cfm (14 L/s)  drawn by the
   diagnostic vacuum cleaner through a 0.5-in. (1.25-cm) suction hole when 1.5
   in. H20 (370  Pa)  suction is being maintained under the slab at a test hole 4
   in. (10 cm) away (when no sump present).  "Excellent" airflow is defined as
   greater than 50  cfm  (24 L/s) under the same conditions.  When sump is present,
   suction is drawn through a 1.5-in. (4-cm) hole through sump cover, -0.1 in.
   H20 in sump; sumps connecting to drain tiles usually result in excellent flow.
      TABLE 2. SUMMARY OF RADON REDUCTIONS ACHIEVED (PYLON MEASUREMENTS)
House
 No.

 208
 277
 582
1165
1248
1357
   System Off*
   Avg. pCi/L**
   Bsmt   L.A.
        Bsmt SSD Only
         Avg. pCi/L
         Bsmt   L.A.
    8.9
    9.7
    3.8
    4.8
    5.1
   10.5
N/A
4.4
2.5
3.4
N/A
2.5
0.5
N/A
0.7
0.1
0.4
6.2
1.2
N/A
0.4
0.6
N/A
3.5
               Crawl Sub-Liner Only
                    Avg. pCi/L
                 Bsmt
***
N/A
1.4
3.0
2.8
2.8
***
N/A
0.3
2.9
N/A
1.3
                                Bsmt + Crawl
                                 Avg.  PCi/L
                                Bsmt    LA.
***
0.7
0.1
0.4
N/A
0.4
1.7
0.3
0.2
0.5
1.1
*  " System-off values are Pylon measurements made back-to-back with the various
   system-on tests reported in the  table.   These Pylon values differ from the
   pre-mitigation charcoal detector levels reported in Table 1 probably due to
   temporal effects.
** Radon levels are the average generally of at least 48 hours of hourly Pylon
   measurements in the basement ("Bsmt") and in the living area ("L.A.") above
   the crawl space.  Upstairs Pylon malfunctioned in House 1248.
***House 208 did not include sub-liner suction in crawl space.

N/A - Not available.
Conversion factors:  Bq/m3 - 37 x value in pCi/L;   cm -  2.54  x value  in in. ;
                     L/s - 0.47 x value in cfm;  Pa - 248 x value  in in. H20.

-------
                                                                   VII-3
          ENGINEERING DESIGN CRITERIA FOR SUB-SLAB DEPRESSURIZATION
                      SYSTEMS IN LOW PERMEABILITY SOILS

                 by: Charles S. Fowler, Ashley D. Williamson, Bobby
                     Fyle, Frank E. Belzer III, Raymond N. Coker
                     Southern Research Institute
                     Birmingham, AL 55305

                     David C. Sanchez
                     U.S. Environmental Protection Agency, AEERL
                     Research Triangle Park, NC 27711

                     Terry Brennan
                     Camroden Associates
                     Oriskany, NY 13424
                                   ABSTRACT

      Radon mitigation by sub-slab depressurization (SSD) in houses built over
soils with low permeability (less than 10"5 cm2) requires different techniques
than the more common case of slabs laid over permeable gravel beds.  Over the
past 2 years 14 slab-on-grade houses in South Central Florida with compacted
soil fills and initial indoor radon concentrations between 400 and 4000 Bq/m3
(10-100 pCi/L) were mitigated using SSD systems. Studies of these installations
indicate that the completion and analysis of certain diagnostic tests can yield
parameters useful to the mitigator for the design and installation of
successful mitigation systems.   Results from the study houses have been
combined into tables and graphs that can be used to help determine recommended
numbers and placement criteria of the suction holes.  Fan and pipe size
selection is assisted by other tabulated or derived information.  Installation
techniques are suggested to enhance the system operation and effectiveness.

    This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.

                                 INTRODUCTION

     Sub-slab depressurization (SSD) is generally the most common and most
effective radon mitigation strategy employed in basement and slab-on-grade
houses.  In many areas of the country, the standard building practice is to
place a layer of coarse gravel directly beneath a vapor barrier before pouring
the slab.  When this has been done, an SSD system is usually quite effective

-------
because of the good permeability and communication afforded by the gravel
layer.  However, many older houses were built before using gravel became a
common practice and in some areas of the country gravel is not readily
available.  In these houses the slabs are poured over either the native soil or
a fill soil that has been compacted to some degree to prevent settling away
from the slab once the concrete has hardened.  Host of the time such a soil
fill has much lower permeability to air flow (usually less than 10"5 cm2).  In
such instances an SSD system will not operate as effectively as it would over a
coarse aggregate bed.  Since much of the literature (1-4) about SSD systems
addresses slabs poured over gravel, guidance in the installation of SSD systems
over low permeability soils has generally been lacking.  Ericson et al. (5) in
Sweden and other researchers (6) have reported cases of low permeability
beneath the slabs and either have made some generic observations about the
average slab area affected by given suction holes or have offered unique
remedies found to work in specific houses.  However, no uniform guidance
uniquely addressing design and installation strategies for solving this problem
seems to exist.

     The South Central Florida (Folk County) area is one area in the U.S. where
coarse gravel is not readily available.  The customary building practice is to
prepare a base of fill soil, overlay it with a vapor barrier, and then pour the
slab.  This practice usually produces somewhat low, but fairly uniform,
communication in sub-slab gas flow.  From December 1987 to September 1989, 14
single-story slab-on-grade houses with living areas of about 120-240 m2 (1300-
2600 ft2) and initial indoor radon concentrations of 400-4,000 Bq/m3 (10-100
pCi/L) have been mitigated with (SSD) systems.   The systems have ranged from
central- and perimeter-located single suction hole systems to up to four
central and/or five perimeter suction holes, with a variety of combinations.
Suction pits ranged from no pits to up to 0.05-0.06 m3 (12-15 gal) in size.
Different sizes of fans and pipes have been installed.  Suction holes were
drilled through the slab and through stem walls under the slab.  Fans have been
located in attics and outside the houses.

     There is continuing research being conducted relevant to design criteria
for similar sub-slab mitigation systems in the same and other areas of Florida,
across the U.S., and in other parts of the world.  The University of Florida,
in particular, is contributing much complementary research to houses in a
different part of the state.  The purpose of this report is to outline some
aids in design and installation of SSD radon mitigation systems for use by
mitigators working on houses where slabs are laid over uniform low-permeability
soils.

                            BACKGROUND INFORMATION

PROBLEM ASSESSMENT

     Before a mitigator or homeowner starts to design a radon mitigation
system,  it should of course be established that there is an indoor radon
problem.  It is reasonable and ethical for a mitigator to communicate to the
homeowner that EPA has published guidance (7) for making reproducible
measurements of radon concentrations in residences, including recommendations

-------
for using the results to make well-informed decisions about the need for
additional measurements or remedial action.  Other complimentary publications
that give more detail and updated information on the specific use of
measurement techniques are also available (8,9).

     Once it is determined that the house in fact does have elevated radon
concentrations, before any other action is taken, certain basic house
information needs to be obtained.  Some of the most crucial elements to note
include the area of the slab(s), the sub-slab media or aggregate, the floor and
ceiling covering, the existence and characteristics of any interior load-
bearing walls, and the existence of interior footings.  Any information that
can be determined about the slab/wall interface is important, as is the
existence of any slab cracks and utility penetrations including plumbing lines.
The type of heating and air conditioning and the location of the duct work and
returns are also helpful to know.  Some of this information can be obtained
from the homeowner, from either existing knowledge or plans, documents, or
pictures taken during construction or renovation.  The rest may be visually
noted or measured during a visit to the house.

     A visit and visual inspection provides an excellent opportunity to check
for potential radon entry points into the building shell.  The cracks and
utility penetrations noted above are likely candidates. One technique for
detecting radon gas almost instantaneously is called the radon "sniff."  It
consists of pulling the gas sample through a scintillation cell while counting
the alpha disintegrations detected each minute.  This type of investigation is
strictly a diagnostic tool and has no set EPA protocol.  Such a device and
procedure tests the candidate entry points for higher radon concentrations than
the ambient room air.

SUB-SLAB COMMUNICATION AND PERMEABILITY

     All of the information described above in the problem assessment process
is useful regardless of the type of mitigation plan to be employed.  Before
attempting to design an SSD system, one other diagnostic test needs to be run.
The diagnostic sub-slab communications and permeability measurement involves
drilling at least one 30-40 mm  (1.25-1.5 in.) hole just penetrating the slab in
the corner of some closet or other space designated by the homeowner near the
center of the slab and drilling several 10-13 mm (0.375-0.5 in.) pressure and
velocity sample holes (6-10) at various distances (1-5 m or 3-15 ft) in several
directions from the suction hole.  A variable speed/suction vacuum cleaner is
used to depressurize the volume beneath the slab at the suction hole.
Instruments capable of measuring pressures in the 500-5000 Pa (2-20 in. we)
range and low flows (0.5-20 L/s or 1-40 cfm) are needed to make the sub-slab
permeability measurements, and measurements down to 0.2 Pa (0.001 in. we) are
needed for the pressure field extension (communication) test.  Figure 1 is a
floor plan of a house in which one suction hole was drilled in a back bedroom
closet and nine test holes were drilled in available corners of rooms and
closets.  The resulting approximate pressure contours have been drawn.

-------
                          SUCTION HOLE DETERMINATION
DETERMINING THE NUMBER OF SUCTION HOLES

     Once, the decision has been made to install an SSD system for radon
mitigation purposes, the most critical questions to answer are how many suction
holes will be needed to remedy the problem and where to put them.  If the house
has more than one slab then, for determining the number of suction holes, each
slab is treated separately.  The single most useful diagnostic tool to use as
input in this determination is the sub-slab pressure field extension
measurement.  From it, the mitigator should have a reasonable feel for what
types of communications are present under the slab.  The recommended procedure
calls for a small test hole to be placed about 300 mm (12 in.) from the vacuum
cleaner suction hole.  With the vacuum cleaner set to produce about a 375-500
Pa (1.5-2 in. we) pressure differential at that test hole, which is about what
a standard mitigation fan would produce, the pressure field measurements should
be taken at the various pressure and velocity sample holes.  At most of the
nearby holes some differential pressure may be measured, but at some of the
more remote ones, more than likely no consistent reading will be possible.  It
is important to remember that in low-permeability soils sufficient time must be
allowed for the pressure field to be established (3-5 minutes for nearby holes
and successively longer times for the more remote ones).  The greatest distance
from the suction hole at which a definite pressure is recorded before the first
zero pressure is read should be taken as the effective radius of extension, r,
of the pressure field from a suction hole in that location.  The analysis to
follow assumes that the sub-slab medium is fairly uniform.  If there seems to
be reason to doubt the validity of this assumption, an additional suction
hole(s) may be recommended.

    Once the effective radius of extension from the suction hole is determined,
the next input required is the approximate area of the slab being considered.
Figure 2 is a graph in which the effective radius of extension is plotted on
the x-axis (from right to left) and the area of the slab is plotted on the y-
axis.  The diagonal lines divide the regions of the effective coverage area of
the indicated number of suction holes.  Find the effective radius, of extension
r, that was determined, go straight up parallel with the y-axis until you find
the area of the slab.  The region between the diagonals where the radius and
area intersect indicates the approximate minimum number of suction holes
required by that slab.  This number may need to be increased if features such
as interior footings, sunken slab areas, sub-slab obstructions, or geometrical
shapes of the slab, seem to limit sub-slab communication.  Erratic or
discontinuous results of the communication test will indicate the possibility
of such a condition.  If the diagnostic test was made when the sub-slab soil
was unusually dry, then the soil permeability and the pressure field extension
determined will most probably be greater than those that would have been
measured during a wetter season.  Generally in low-permeability soils, there is
little likelihood in producing too great a flow for the depressurizing fan, so
when in doubt, an extra hole is a better option than not having enough.

-------
DETERMINING THE SUCTION HOLE PLACEMENT

     If the mitigation system is being installed in an unfinished space such as
a basement wbere there are no restrictions on the placement of the suction
holes, then the mitigator has a great amount of flexibility in making his
decision.  A floor plan drawn to scale, perhaps one on which the sub-slab
communication is plotted, is very useful at this point.  Sketching in the
effective areas of pressure field extension from various suction hole
placements will give an idea of the optimum configuration to ensure the best
coverage of the slab.  Geometry suggests that holes located about 1 effective
radius, r, from the nearest exterior wall(s) will give the widest coverage.
However, in practice, sometimes the soil near the edge of a slab has not been
compacted as well as that near the center, producing either a possible settling
space between the top of the soil and the bottom of the slab or else just a
more permeable trench near the perimeter of the slab.  If the diagnostic
communication test indicated that a greater pressure field extension resulted
from a near-perimeter suction hole than from a central suction hole without a
much larger flow, then the placement of the suction holes nearer the perimeter
is recommended.  If, however, the communication test showed much greater flows
from perimeter holes without much greater pressure field extension, then slab
cracks or other leakage is probably limiting the pressure field extension, and
perimeter suction holes should be avoided.

     When the slab being mitigated is predominantly under finished space, such
as a finished basement or a slab-on-grade house, practical locations are
usually far more restricted.  In such a circumstance a floor or house plan is
very helpful.  The finished basement scenario is probably the more difficult
system to design.  Usually the best locations from the homeowners' viewpoint
are corners of closets because there the installations will be less noticeable
and obtrusive.  However, quite often closets will not be spaced to give full or
adequate pressure field coverage.  If that is the case, one may consider
placing the suction hole in the corner of a room and then perhaps "boxing off"
that corner if the homeowner does not want the pipe to show.  Boxing off can be
used for more central locations as well.  The added difficulty with finished
basement installations involves finding a place or places for the pipes to
penetrate the basement ceiling which will line up with an acceptable first
story floor penetration.

     Slab-on-grade houses usually also have most, if not all, of the area to be
mitigated as finished space.  Closets may be spaced more advantageously than
are often found in finished basements.  Moreover, there may be a pantry or
other location where a suction hole may be concealed.  There may still be large
areas that cannot be affected by near-closet suction holes.  These are most
typically open living room/dining room/kitchen/den areas.  Quite likely there
would be more resistance from the homeowner to placing any interior piping,
even concealed, in such spaces.  One possibility to pursue in such a occurrence
would be an exterior suction hole penetrating horizontally through a stem wall
beneath the slab rather than vertically through the slab in an interior space.
What is required for such an exterior penetration to succeed is that the stem
wall must be accessible from outside the house (no porches, patios, or concrete
or paving directly adjacent to the outside wall where the penetration is

-------
proposed) and  that the penetration can be Installed without losing the pressure
field  to slab  cracks and stem wall leakage as was mentioned earlier with near-
perimeter placement in slab-on-grade houses.  If the footing is on expansive
soils  or there seem to be foundation or structural weaknesses near the stem
wall in question, a suction hole should not be placed there.

     One other possible suction point location in some slab-on-grade houses is
through an attached garage area.  Some garages actually have a portion of the
house  slab exposed at one end of the space.  Even if not, other garages are a
few steps down from the house floor level.  In such an instance, the house stem
wall may form  the lower course or two of the interior wall of the garage.  Then
a horizontal penetration through the stem wall beneath the slab could be a good
suction point.  Even if the garage is just a small step down from the house
slab,  it may be possible to penetrate the garage slab and extend the system
depressurization under the house.  A potential problem with using a garage
penetration is that often the garage slab has settled and/or cracked, leaving
possible by-passes where garage air may be drawn into the system, reducing the
effective suction head and limiting the effectiveness of the system.

                                FAN SELECTION

COMPARING SUB-SLAB FLOW CURVES WITH VARIOUS CURVES

     While the pressure field extension measurements of the sub-slab communica-
tion diagnostic give a good approximation of an effective dspressurization
radius, the pressure and flow measurements are indicators of the sub-slab
permeability.  Specifically, recommended procedures call for the simultaneous
measurement of the suction at the scaling baseline hole and flow from the 30-40
mm (1.25-1.5 in.) suction hole at suctions of at least 0.5, 2.0, and 5.0 kPa
(2, 8, and 20  in. we) produced by a vacuum cleaner.  When these measured values
are plotted on an x-y axis, such as in Figure 3 for the highest permeability
(10"5  cm2) and one of the more typical (10'7 cm2) encountered in the Polk
County, Florida, study houses, one obtains a flow curve for the sub-slab fill
material.

     Also plotted in figure 3 are fan performance curves taken from the EPA
Training Course Manual (10) and from other published fan company figures.  The
RDS and R-150/K-6 are inline centrifugal fans designed with radon mitigation in
mind.  The radial and vortex blowers are higher suction instruments that may be
adapted for use in mitigation systems.  On such a simultaneous plotting, the
intersections  of the soil curves with the fan curves indicate about where the
system will operate.  Figure 3 suggests that, for both soils, but especially
the one with the lower permeability,  the system will tend to operate near the
high suction/low flow end of the fan curve for the RDS,  R-150/K6, or the radial
blower.  The fan curve data for the vortex blower did not extend further than
the 1.5 kPa (6 in. we) suction in the plot, but it obviously intersects both
soil curves at higher suctions and higher flows.

FAN CHOICE CONSIDERING OTHER FACTORS

     Because the mitigation field experience in low-permeability soils is still
in an early phase, it is not clear what the durability of a fan will be when it

-------
is operated at low flows and relatively high suctions.  Some indications
suggest that fan failure nay occur sooner in such an operating environment. The
fans are often placed in attics which will be quite hot during the cooling
season.  High heat with low flows through the fans may lower the durability of
the fans.  Research is currently underway to determine if the system
deteriorates with time or if it is maintained fairly constant until some type
of failure occurs abruptly.  Harrje et al. (11) at Princeton University have
developed a diagnostic which investigates the durability of the fan as well as
the mitigation system as a whole.

     The inline centrifugal fans, since they have been designed for radon
mitigation situations, have been kept fairly lightweight and affordable.
The blowers that produce the higher suctions are generally built for industrial
applications and therefore are somewhat heavier and more costly.  The power
requirements to operate these various fans may differ quite widely.  The
lightweight inline fans are designed to perform in the 100-200 W power range.
The heavier, higher suction blowers will sometimes have several times this
power requirement.  Therefore, the operating costs may vary greatly with the
choice of mitigating fan.  Since research data have not been collected for a
long enough time in this area, it is not clear how to predict the long-term
costs of these various systems.  If the inline fans have too short a lifetime,
replacement costs may make this system more expensive.  If their durability is
long enough, then their lower initial cost and operating costs may make them
the more cost-effective system.

     The inline centrifugal fans are designed to run very quietly (less than 6
sones) and, according to most reports, receive very little, if any, criticism
from homeowners in this regard.  However, the larger, more powerful blowers,
especially if designed for industrial applications, characteristically produce
quite a bit more noise, often a steady, high-pitched whine.  This noise factor
usually is dealt with by installing the fan as far from the living space as
possible and Including soundproofing when the system is first installed.  Both
of these options may increase the initial installation costs, and a remote fan
placement will require longer piping runs which may reduce the system
effectiveness.  Even with the additional precautions to limit the noise output,
some people sensitive to noise may still object to the larger fans.

     Decisions on interior versus exterior suction holes and piping definitely
have a bearing on fan selection.  If the exhaust pipe from suction holes in a
basement is routed through a rim joist to the outdoors, or if a suction hole in
a slab-on-grade house is through an exterior stem wall, then the fan will
probably be placed somewhere outside the house.  Such a fan must be rated for
exterior applications.  In some model lines these fans are more expensive than
Interior fans.  The ease of handling and weight of the units with the supports
required are other factors to consider in the fan selection process.

                                PIPE SELECTION

     Generally most mitigators use PVC pipe when installing SSD systems.  It is
lightweight, easy to cut and handle, convenient for fittings and accessories,
strong in its glueing characteristics, noncorrosive,  and smooth so as to offer

-------
low resistance to air movement.  For permeable sub-slab environments conducive
to high volumes of airflow, 100 mm (4-in.)  or larger FVC pipes are generally
used.  For the low flows resulting from the low permeability soils addressed in
this document, 100 mm (4-in.) or smaller PVC pipes are usually adequate.  The
smaller piping has the added advantages of being lighter and easier to handle,
less obtrusive to the homeowner, easier to conceal if desired, and usually less
expensive for the pipe, fittings, and accessories.  Therefore, an important
determination is what size of pipe is the best to use for the given mitigation
project.

AIRFLOW VERSUS APPLIED SUCTION

     The choice of pipe size is most directly governed by the volume rate of
flow (or velocity) expected to move through the pipe.  The larger the volume
moving, the greater the resistance if the cross sectional area is kept
constant.  However, increasing the pipe size will lower the frictional loss for
a given volume rate of movement.  Therefore, the best inputs for estimating the
optimum pipe size for a mitigation system again come from the sub-slab
communication diagnostic pressure/flow measurements.  If the fan has been
selected that is expected to be used in the mitigation system, then the point
of intersection of the fan curve with the sub-slab flow curve will give a good
approximation of the airflow that can be expected in the system.

     From the airflow estimate, one may use a chart such as Figure 4 to
estimate the friction loss in various sizes of pipes or ducts.  This chart,
like most of the available documentation on airflow through pipes or duct work
(12), is calculated for "average" pipe, which is usually some type of iron pipe
with a given smoothness and joints estimated to be present at some regular
frequency.  PVC pipe is quite a bit less resistive to air movement because of
its greater smoothness.  Therefore, these approximations usually overestimate
the friction loss that would actually be found in FVC pipes.  If the fan
selected is one in which the sub-slab flow curve intersection with the fan
curve is in the 375-500 Pa (1.5-2 in. we) range, then one would probably want
to keep the friction loss to 0.8-1.5 Pa/m of pipe.  If the fan curve intersects
the sub-slab curve at something greater than 1 kPa (4 in. we), then a friction
loss of 3-5 Pa/m of pipe could be tolerated.

     To use a chart such as Figure 4, find on the y (vertical) axis the airflow
determined from the sub-slab/fan curve intersection.  Go across horizontally
until you are in the friction loss range (x-axis) you determined as above.  The
closest pipe size diagonal (those rising from left to right) would be approxi-
mately the best pipe size to achieve your goal.  It is advantageous from the
perspective of friction loss to go with the larger pipe, but if other factors
such as expense, ease of handling, or homeowner preference indicate otherwise,
the smaller pipe would probably still be a safe choice, especially in light of
the lower friction of PVC pipe discussed previously.

     The friction loss in straight pipes is only part of the loss of suction
head that is experienced in a system.  Usually the next most significant
features contributing to friction loss are the bends or tees in the system.  A

-------
90" elbow or tee in a pipe usually contributes the greatest pressure drop
potential of any of these features.  A 45* elbow has slightly over half the
friction loss of a 90" elbow, and a 30° elbow has less than half that of a 90°
one.  Table 1 lists the approximate length of pipe that produces the same
friction loss as each of several of the more commonly used connectors.

      TABLE 1.  APPROXIMATE FRICTION EQUIVALENCIES FOR VARIOUS FITTINGS
                                     Equivalent Run of Pipe (m)

Type of Fitting
Tee
90* Elbow
45' Elbow
30s Elbow
Pipe Diameter (mm)
37.5
0.5
0.3
0.25
0.15
50
0.6
0.5
0.3
0.25
75
1
0.6
0.5
0.3
100
1.5
1
0.6
0.5
     To determine the friction loss in Pascals (Fa) for a system, determine the
total length of pipe and the number and kinds of fittings for each pipe size.
Multiply the number of fittings for a pipe size by the equivalency from Table 1
for that fitting and pipe.  Add the total equivalent meters so determined to
the actual length of pipe to be used to get the adjusted total length of pipe.
Then use the friction loss factor determined from Figure 4 to multiply by that
adjusted total.  Repeat the calculation for each pipe size and add the total
together for the whole system.

APPLICABILITY AND AVAILABILITY

     If the above calculation indicates a larger pipe size than is feasible or
desired by the homeowner, then perhaps a fan that can draw a larger suction at
lower flows is called for.  If, however, a certain pipe and fitting size is
determined that is acceptable, then local supply stores should be investigated
to ensure that enough pipe, fittings, and accessories are easily available.
PVC pipe comes in a variety of thicknesses (sometimes called schedules).  The
thicker walls are for high-pressure applications; consequently that PVC is
heavier and more expensive.  The applications described here require no extra
thickening, so the thinnest-walled PVC pipe is usually adequate and preferred
for its weight, ease of cutting, and cost.  However, some of the fittings and
couplings for one schedule will not fit properly or tightly on the same size
pipe of a different schedule.  So a crucial part of the pipe selection process
is that there be an adequate supply of fittings and accessories for the size
and schedule of the PVC selected.  Other couplings, reducers, bushings, etc.,
should be investigated at this phase of the process to ensure complete
compatibility and availability for the system.  These are used chiefly at the
various interfaces--pipe to slab, pipe to fan, and fan to exhaust.

-------
                                  REFERENCES
1.   Henschel, D.  B.  Radon Reduction Techniques for Detached Houses:  Technical
     Guidance (second edition).   EPA-625/5-87-019,  U.  S.  Environmental
     Protection Agency, Cincinnati,  Ohio,  1988.  245 pp.

2.   Osborae, M. C.,  T. Brennan, and L.  D.  Michaels.  Radon Mitigation in 10
     Clinton, New Jersey, Houses: A Case History.   EPA-600/D-87-164 (NTIS PB87-
     191847), Presented at Second APCA International Conference on Indoor
     Radon, Cherry Hill, NJ, April 6-10, 1987.   12  pp.

3.   Michaels, L.  D., T. Brennan, A. S.  Viner,  A. Mattes, and W. Turner.
     Development and Demonstration of Indoor Radon  Reduction Measures for 10
     Homes in Clinton, New Jersey.  EPA-600/8-87-027 (NTIS PB87-215356),  July
     1987.  166 pp.

4.   Findlay, V. 0.,  A. Robertson, and A.  G. Scott.  Testing of Indoor Radon
     Reduction Techniques in Central Ohio Houses:   Phase 1 (Winter 1987-1988).
     EPA-600/8-89-071 (NTIS PB89-219984),  July 1989.  301 pp.

S.   Ericson, S. 0.,  H. Schmied, and B.  Clavensjo.   Modified Technology in
     New Constructions and Cost Effective Remedial  Action in Existing
     Structures, to Prevent Infiltration of Soil Gas Carrying Radon.
     Radiation Protection Dosimetry 7:224-225,  1984.

6.   Scott, A. G., A. Robertson, and V.  0.  Findlay.  Installation and Testing
     of Indoor Radon Reduction Techniques in 40 Eastern Pennsylvania Houses.
     EPA-600/8-88-002 (NTIS PB88-156617),  January 1988.  388 pp.

7.   Ronca-Battista,  M., P. Magno, and P.  Nyberg.   Interim Protocols for
     Screening and Follow-up Radon and Radon Decay  Product Measurements.
     EPA-520/1-86-014, U.S. Environmental Protection Agency, Washington,  D.C.,
     1987. 22 pp.

8.   U.S. Environmental Protection Agency.   Indoor  Radon and Radon Decay
     Product Measurement Protocols.   EPA-520-1/89-009, Washington, D.C.,  1989.
     102 pp.

9.   U.S. Environmental Protection Agency.   A Citizen's Guide Co Radon.
     OPA-86-004, Washington, D.C., 1986.  13pp.

10.  U.S. Environmental Protection Agency.   Reducing Radon in Structures (2nd
     Edition).  Washington, D.C., 1989.  356pp.

11.  Harrje, D. T.,  K. J. Gadsby, and D. C. Sanchez.  Long Tern Durability and
     Performance of Radon Mitigation Subslab Depressurization Systems.
     Presented at 1990 International Symposium on Radon and Radon Reduction
     Technology, Atlanta, GA.  February 1990.  14 pp.

12.  American Society of Heating, Refrigerating and Air-Conditioning Engineers,
     Inc.  ASHRAE Handbook 1981 Fundamentals.  Atlanta, GA. Chapters 33 and 34.

-------
                                                                   1 m
Figure 1. Approximate pressure contours from a suction hole in a representative
         house plan (5 kPa suction at suction hole).

-------
     100
CM


 CO
 
-------
                                                    High-permeability soil

                                                    Low-permeability soil

                                                    RDSian

                                                    Radial blower

                                                    Vortex blower

                                                    R-150/K-6fan
 20
                           50
                           0
1000
1500
                                   Suction (Pa)
Figure 3.  Sub-slab curves for two permeabilities plotted with fan curves for four
          different kinds of fans.

-------
100
.£•
•i-*




6
 10
    0.1
                                         Friction Loss (Pa/m)
                                                                    10
100
                                Rgure 4.  Friction chart for average pipes.

-------
                                                                 VII-4
              RADON MITIGATION TECHNIQUES FOR BASEMENT HOUSES
                     WITH POOR SUB-SLAB COMMUNICATION

               by:  Bobby E. Pyle
                    Southern Research Institute
                    Birmingham, AL  35255

                    Michael C. Osborne
                    AEERL, U.S. Environmental Protection Agency
                    Research Triangle Park, NC  27711

                                 ABSTRACT

     While the sub-slab depressurization (SSD) technique is widely used to
mitigate radon levels in basement houses, there are instances for which
this method is not a viable option.   For example,  in some houses the slab
is poured directly on the soil resulting in poor-to-nonexistent
communication under the slab.  To apply SSD methods requires elaborate
plumbing and multiple suction holes in the slab.  In an effort to develop
viable alternatives, EPA has funded research to explore other radon
mitigation options.  Case studies will be presented that include:  basement
pressurization with and without heat extraction, and filtration using
charcoal.  In the first, air from the upper portion of the house was forced
into the basement producing a pressure barrier at the slab/soil interface.
In one case heat was extracted from the upstairs air using a heat pump to
supply hot water for the occupants.  The air filtration method used a bed
of charcoal to remove the radon gas.  The charcoal bed was flushed with
outdoor air to extract the radon before it decayed completely to radon
daughters.

     This paper has been reviewed in accordance with the United States
Environmental Protection Agency's peer and administrative review policies
and approved  for presentation and publication.

-------
                                INTRODUCTION

     Phase I of this EPA Radon Mitigation Demonstration,  in Nashville,
Tennessee, identified many basement houses which had poor sub-slab
communication.  In attempting to implement a sub-slab depressurization
(SSD) system in houses such as these, special attention had to be taken to
ensure an adequate means to prevent radon entry into the basement.  In an
effort to determine if less expensive techniques were available that might
have a higher probability of success, the mitigation methods described
below were investigated.  These techniques may be applicable in other areas
such as Central Florida where the lack of communication under the slab
increases the complexity of SSD systems.
                           DESCRIPTION OF HOUSES

     Nine houses were selected for the Phase II demonstration in Nashville,
TN.  Of these, six are basement houses and three are crawl space houses.
Of the six basement houses only four have been mitigated as of the date of
this paper.  A brief description of these four houses follows.  A final'EPA
report describing all of the mitigation efforts in Nashville should be
available by the summer of 1990.

     Dtt 01 is a one story frame house with brick veneer upstairs.  The
basement is a daylight walkout at the back of the house with walls of
concrete block.  The basement area is approximately 70% finished with a
bedroom and den in the front (below-grade) part of the basement and an
unfinished laundry room and workshop at the rear.  The heating system for
the house includes a central heat pump which is seldom used and two wood
stoves, one upstairs and one in the finished basement.  There are no other
combustion appliances in the house.  Charcoal measurements made in the
winter of 1987-88 in the basement by the homeowner showed the radon level
to be 21.4 pCi/L.  Results of blower door infiltration measurements at 4 Pa
showed 2 ACH (air changes per hour) for the basement and 1.4 ACH for the
whole house.

     DW 35 is a two story frame house with brick veneer on the exterior.
The basement is almost totally below grade with walls of cut stone.  The
basement is semi-finished but unused because of water problems. A central
gas fired furnace in the basement has numerous leaky joints around
attempted seals with asbestos tape.  Also located in the basement are a gas
fired clothes dryer and water heater.  Visual inspection showed numerous
thermal bypasses from the basement to upstairs.  A seldom-to-never used
fireplace is located in one room of the basement.  The flue has been closed
loosely with sheet metal but still shows air exfiltration.  A previous
radon measurement in the basement was 18 pCi/L.  Infiltration measurements
at 4 Pa showed that the basement was very leaky:  6.3 ACH for the basement,
and 4.4 ACH for the whole house.

-------
     DW 38 is a one story brick veneer house with a walk-out  daylight
basement.  The walls of the basement are concrete block with  three of  the
walls about half below-grade by about 3-4 ft (0.92 - 1.22 m).  The basement
is 50% finished with a den in one part and a laundry room and work shop in
the other.  The heating and air-conditioning (HAG) system is  a central unit
in the basement (heat pump) with very leaky ductwork (both supply and
return).  There were no combustion appliances in the basement.  Previous
measurements in this house were 11.1 pCi/L in the basement.  Infiltration
measurements at 4 Pa resulted in values of 2.8 ACH for the basement and 1.6
ACH for the whole house.

     DW 47 is a single story frame/brick veneer house with a  walk-out
basement.  The homeowner was in the process of finishing the basement when
mitigation was undertaken.  The walls of the basement are concrete block
with three sides below grade.  The house is totally electric with the air
handler located upstairs.  The basement contained a roughed-out opening for
a fireplace with a large amount of bypass area to the upstairs flue box.
Also located in the basement is a heat pump water heater using basement
heat.  Previous radon measurements by the homeowner were 13.1 and 16.9
pCi/L upstairs and 21.1 pCi/L in the basement.   Infiltration measurements
at 4 Pa gave values of 2.4 ACH for the basement  and 1.3 ACH  for the whole
house.
                             MITIGATION SYSTEMS

     In three of the houses  (DW 01, DW 38, and DW 47), radon levels were
mitigated by pressurizing the basement with air withdrawn  from upstairs.
This was  done  in two of the  houses  (DW 01 and DW 38) by using an  in-line
blower with maximum flow capability of approximately 400 cfm at 0 in. H20
 (11.35 m3/min at 0 Pa)  (T3A Turbo 8, R.B. Kanalflakt,  Inc., Sarasota, FL).
In DW 01  the air inlet was located  in the central hallway  of the  house  and
in DW 38  it was located in the  floor  of  the kitchen.

     In the third house (DW  47) a heat pump water heater (HPV-80,  Therma-
Vent™,  Therma-Stor Products Group, DEC International, Inc., Madison, WI)
was used  to withdraw  air from the upstairs area, down  through the heat
pump, and then exhaust the air  into the  basement.   The heat extracted from
 the air was used  to provide  hot water for the homeowner.   In DW 47 the  air
 inlet was located  in  the central hallway.

     Several parameters of each of  these three houses  were monitored and
 recorded  using a  datalogging device.   The parameters  were: air temperatures
upstairs, in the basement, and outdoors; air pressure  differentials between
 the basement  and upstairs, between the basement  and outdoors, and between
 the basement  and  the  sub-slab region; wind  speed and  direction;  rainfall
 amounts;  HAG  operation; and  continuous  radon  levels both in the basement
 and on  the first  floor.  The monitoring period was  quite long, ranging  from
 several weeks to  months in duration.   Due to  space  limitations,  the bulk of
 this  monitoring data cannot  be listed and explained in this paper.

     A major  point considered for  this  type of mitigation  system  was the
 amount  of sealing required in the basement to achieve sufficient

-------
 overpressurization while keeping the  airflow rate  down to  less  than 300-
 400 cfm (8.51 -  11.35 m3/min).  Also considered were the effects of such a
 system on the air infiltration of the house.  Air  infiltration  was  measured
 to determine how much of an energy penalty the  system  imposes on the house.
 Results of these studies will be included in the EPA final report.

     In the  remaining basement house  (DW  35)  an off-the-shelf activated
 charcoal air cleaning unit  (Radsorb-222,  RAd Systems,  Inc., Westborough,
 MA)  was used to  remove the  radon gas  from the basement air.  In this device
 the radon gas adsorbed by the charcoal is removed  by reverse flushing  to
 the outdoors before the radon completely  decays to its daughter products
 and remains  trapped in the  charcoal bed.   The radon levels in both  the
 basement and the first floor were continuously monitored over a period of
 several months.

                          HOUSE MITIGATION RESULTS

     These houses were  mitigated  during 1989.  Because of  the length of
 monitoring time,  the  mitigations  extended from January through  November of
 1989.   Prior to  mitigation,  long term alpha  track  detectors (ATDs) were
 installed in the houses.  ATDs were obtained from  Tech-Ops Landauer, Inc.
 (Glenwood, IL).  The results  of these  pre-mitigation measurements are shown
 in Table 1.   The radon levels in the  basements ranged  from 9.4  to 24.4
 pCi/L  and the levels  on the  first floor ranged from 4.9 to 10.4 pCi/L.
 Short  term (3 day), closed house,  charcoal canister (CC) measurements  were
 carried out  just prior to mitigation.  Charcoal canisters  were  provided and
 analyzed by  Scientific Analysis,  Inc.  (Montgomery,  AL). These results  are
 shown  in Table 2. The agreements  between  the  ATD results and the CC values
 are  not as good  as  might be  expected.  The differences are attributed  to
 differences  in the  time of year  and to weather. The average levels  in  the
 basement and on  the first floor  for both  the  CC and the ATD measurements
 are  summarized in Figure 1.

 House  DW 47

     The  first house  to be mitigated was DW 47 using the HPV-80 device.
 The  radon levels  in the basement  and on the  first  floor along with the
 house  parameters were continuously  monitored  beginning on  1/31/89.  The
 continuous radon monitor (CRM) values during  the pre-mitigation period
 averaged 13.1 pCi/L in the living room and 18.1 pCl/L  in the basement.  The
 HPV-80  system was installed  on 3/15/89, and  the openings between the
basement  and upstairs were sealed the following day.  The major openings
 that were closed were the bypass  to the chimney flue, the water and sewer
pipe penetrations,  and  an opening in the sub-floor below the HAC.  Since
 the basement was  largely unfinished, sealing  these  openings was  simple and
 effective.   The HPV-80  fan was turned on at 6:40 pm on 3/15/89.   The radon
 levels began to decline within 6 hours, and within  12 hours the  levels were
 less than 2  pCi/L in  both the  basement and on the first floor.   The CRMs
remained  in  the house until 4/13/89.  Over this post-mitigation period the
average radon levels  were 1.8  pCi/L in the basement and 1.7 pCi/L on the
 first  floor.

     Prior to turning on the HPV system,  the basement pressure was
averaging about 2.5 Pa below ambient outdoor pressure.   With the system on,

-------
the basement pressure was approximately 2.5 Fa above ambient outdoor
pressure.  The HFV system flowrate was 141 cfm (4.00 m3/nin) measured at
the exhaust outlet in the basement.  On 4/4/89 the HPV system was shut down
for approximately 48 hours and then was turned back on.  The pressure
changes over this period are shown in Figure 2, and the changes in the
radon levels are shown in Figure 3.  During the time the HPV system was on,
the sub-slab pressure was averaging about 2 Pa below the basement pressure,
and the basement was about 2.5 Pa above the outdoor pressure.   With the HFV
system off, the sub-slab pressure increased to 1 Fa or less (below the
basement pressure) and the basement pressure decreased to about -3 Pa
relative to the outdoors.  The radon levels in both the basement and on the
first floor increased by factors of 5 or more.  When the system was turned
back on, the levels decreased sharply as seen in Figure 3.  Overall the
reduction in radon levels was approximately 92% in the basement and 83% on
the first floor.

House DW 38

     The second house to be mitigated using basement pressurization was DV
38.  In this house an in-line duct fan was used to pull air from the
kitchen upstairs.  The air was then distributed in the basement in two
zones, the finished den and the unfinished workshop.  The mitigation system
was installed and energized on 4/25/89.  Prior to mitigation,  the radon
levels and house parameters were monitored beginning on 4/13/89.  The
average pre-mitigation radon levels measured with the CRMs were 10.6 pCi/L
on the first floor and 14.6 pCi/L in the basement.  Also over this period,
the average basement pressure was approximately 1 to 2 Fa below the outdoor
ambient pressure, and the sub-slab area pressure wa& approximately 2 Pa
more positive than the basement.  During the initial mitigation a small fan
[158 cfm at 0 in. H20 (4.48 m3/min at 0 Pa), model R125, Fantech, Inc.,
Sarasota, FL] was used.  It was soon apparent that the fan could not supply
sufficient airflow to overcome the leaks to the upstairs and to the
outdoors.  A larger fan was installed [410 cfm at 0 in. H20 (11.63 ms/nin
at 0 Pa), model T3A Turbo 8, R.B.  Kanalflakt. Inc., Sarasota, FL] to
pressurize the basement.

     Sealing this house was difficult because of limited access to the
overhead from the basement.  The finished basement ceiling had acoustic
tile attached directly to the floor joists.  Also, the homeowner had made
several changes in the electrical and plumbing systems over the years and
consequently left many openings.  The MAC system located in the unfinished
part of the basement not only had many leaks in the cold air return but
also had a return vent in the finished area of the basement that had to be
sealed off.  To prevent the pressurization air from leaking back upstairs
through the return duct, a back-draft damper was constructed at the return
opening upstairs.  Also, to prevent back-flow through  the supply vents in
the basement, back-draft flaps of vinyl were attached  to the outside of the
supply grills.  Also, a direct path to the attic in the basement stairwell
required sealing.  In summary, this house was extremely difficult to seal.
The basement and the upstairs were never completely isolated.

     Following installation of the larger pressurization fan and sealing of
as many of the openings as possible, the pressure difference between the
basement and the sub-slab region was on the average reduced to zero.  The

-------
basement was never pressurized above the outdoor ambient conditions.
However, by achieving a neutral pressure difference relative to the sub-
slab, radon gas flow into the basement was significantly reduced.   No
condensation problems were observed with the kitchen air pulled down into
the basement. The radon levels in the basement, as measured with the CRM,
averaged 1.3 pCi/L.  These levels were confirmed with CC measurements
carried out on 8/22-24/89.  The CC results, shown in Table 2, were 1.2
pCi/L in the basement and 0.9 pCi/L on the first floor.

House DW 01

     Radon in this house was mitigated over the period from 10/2/89 to
10/6/89.  Prior to mitigation the datalogging system was installed on
9/13/89 and remained until 11/6/89.  The pre-mitigation radon levels in
this house were 19.5 pCi/L in the basement and 9.8 pCi/L on the first floor
as measured with CCs and shown in Table 2.  The average radon levels
recorded with the CRMs (model R210F radon monitor, Femto-Tech, Inc.,
Carlisle, OH) were 22.5 pCi/L in the basement and 8.5 pCi/L on the first
floor.  The basement pressurization mitigation system was installed and
turned on at approximately 4:00 p.m. on 10/3/89.  Sealing of the basement
openings to the upstairs and to the outdoors was delayed until 10/5/89.
With only the pressurization fan installed and no additional sealing, the
radon levels dropped within 6 hours to 2.8 pCi/L upstairs and 4.2 pCi/L in
the basement.  This is shown in Figure 4.  The average pressure in the
basement relative to: the outdoors changed from -0.4 to -0.2 Pa, the
upstairs from +0.03 to +2.30 Pa, and relative to the sub-slab area from
-0.1 Pa to zero.  With a limited amount of sealing the radon levels dropped
further to 1.6 pCi/L upstairs and 1.9 pCi/L in the basement.  The average
pressures in the basement increased to +1.1 Pa relative to the outdoors,
+2.4 Pa above the upstairs, and +0.14 Pa above that under the slab.

     On one occasion, the homeowner inadvertently left the door between the
upstairs and basement open for most of a day.  During this time the
pressurization was lost in the basement and the radon levels in both the
basement and the first floor increased sharply.  These results are shown in
Figures 5 and 6, respectively.

     The radon levels and the house parameters were monitored until
11/6/89.  The radon levels in the basement averaged 1.2 pCi/L and on the
first floor 1.1 pCi/L.  These levels were confirmed with CC measurements
carried out over the period 11/6-8/89.  These tests gave the results (shown
in Table 2) of 0.2 pCi/L in the basement and upstairs.

House DW 35

     This house was mitigated with the charcoal adsorption unit (Radsorb-
222).  The unit was installed on 5/30/89.  Continuous radon monitors (model
R210F radon monitor, Femto-Tech, Inc., Carlisle, OH) were installed prior
to installation on 5/18/89.  Pre-mitigation radon levels averaged 11.2
pCi/L in the basement and 5.4 pCi/L on the first floor.  The charcoal

-------
adsorption unit was operated more-or-less continually for the next 6
months.  Monitoring of the radon levels in the basement and upstairs
continued throughout most of this period.  Monitoring was interrupted on
several occasions by failure of the CRM devices.  The failures were due to
excessive moisture levels in the basement.  Over the last period for
which data were available for this paper, the average levels measured by
CRMs (model AB5 with PRD1 passive cells, Pylon Electronics Development Co.,
Inc.,  Ottawa, Canada) were 5.5 pCi/L in the basement and 4.2 pCi/L
upstairs.  The CRMs also indicate that during heavy rains the levels can
increase drastically.  During one such rain storm on 10/17/89 the levels
increased to about 54 pCi/L in the basement and 25 pCi/L upstairs.
Following the rain storm, the charcoal device reduced the levels to 6.4
pCi/L in the basement and 3.5 pCi/L upstairs within 24 hours.  The radon
levels in both the basement and the living area over this period are shown
in Figure 7.

     Post-mitigation CC measurements were carried out over the period 11/6-
8/89 as shown in Table 2.  These levels are higher than those measured with
the CRMs over a much longer period.  However, the CC measurements were
begun shortly (about 6 hours) after a storm dropped approximately 1.2 in.
(3.05 cm) of rain in the area.  CRMs will be used to monitor this house for
some additional period of time.  Also, long term ATDs have been installed
both in the basement and upstairs.

CONCLUSIONS

     The basement pressurization technique has been applied to three houses
of differing age and construction type.   In one house heat was extracted
from the upstairs air before pressurizing the basement.  The degree of
success in reducing the radon levels with this method is determined by how
tight the basement can be made to air leaks to the outdoors and to the
upstairs.  Changes in house tightness are currently being quantified. For
houses in which the HAC system incorporates continuous ductwork in good
condition, the sealing is not too difficult.  For houses in which the floor
joists are used as the sides of cold air  return ducts, sealing can be very
difficult and time consuming.  Each house has to be evaluated on  its own.
Use of diagnostic tests such as blower door measurements can aid  in the
evaluation prior to mitigation system design.  The charcoal adsorption
technique is currently under test in one  house.  The device  is capable of
reducing moderate levels  (10 to 15 pCi/L) of radon to levels that approach
4 pCi/L.  However, the device cannot control rapidly increasing levels as
produced during water capping of the surrounding soils.  The long-term
effectiveness of the device is currently under  study.

-------
                  TABLE 1.  ALPHA TRACK DETECTOR MEASUREMENTS FOR
                         NASHVILLE PHASE II BASEMENT HOUSES
      Location   Date       Date     Days    Radon
        Code    Installed  Removed  Exposed  Level
                                             (pCi/L)
                               Comments
      DW 01-B* 09/22/88
      DW 01-B  09/22/88
      DW 01-L* 09/22/88
      DW 01-L  09/22/88

      DW 35-B  09/20/88
      DW 35-B  09/20/88
      DW 35-L  09/20/88
      DW 35-L  09/20/88

      DW 38-B  09/21/88
      DW 38-B  09/21/88
      DW 38-L  09/21/88
      DW 38-L  09/21/88

      DW 47-B  09/21/88
      DW 47-B  09/21/88
      DW 47-L  09/21/88
      DW 47-L  09/21/88

      DW 47-B  04/13/89
      DW 47-B  04/13/89
      DW 47-L  04/13/89
      DW 47-L  04/13/89
09/14/89
09/14/89
09/14/89
09/14/89

05/30/89
05/30/89
05/30/89
05/30/89

05/04/89
05/04/89
05/04/89
05/04/89

03/16/89
03/16/89
03/16/89
03/16/89

10/03/89
10/03/89
10/03/89
10/03/89
357
357
357
357

252
252
252
252

225
225
225
225

176
176
176
176

173
173
173
173
12.6
16.2
 5.4
 5.0

 9.4
 9.6
 5.4
 4.9

11.1
11.2
 9.9
 9.3

12.5
24.4
 8.1
10.4

 1.6
 1.5
 1.7
 1.5
Pre-Mitigation
Pre-Mitigation
Pre-Mitigatlon
Pre-Mitigation

Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation

Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation

Pre-Mitigation
Pre-Mitigation
Pre-Mitigation
Pre-Mitigation

Post-Mitigation
Post-Mitigation
Post-Mitigation
Post-Mitigation
     * B -  Basement,  L -  1st Habitable Level Above  Basement

                     TABLE 2.   CHARCOAL CANISTER MEASUREMENTS  FOR
                          NASHVILLE PHASE II BASEMENT HOUSES

                                    Radon Level (pCi/L)
                                                  1st Habitable
                                                       Level
                                                  Test 1  Test 2
House
ID
DW 01
DW 01
Date
Started
9/13/89
11/6/89
Date
Stopped 1
9/15/89
11/8/89
Basem
test 1
19.0
0.3
ent
Test 2
20.0
0.1
                                                     9.6
                                                     0.1
                                9.9
                                0.2
                            Comments

                           Pre-Mitigation
                           Post-Mitigation
  DW 35*    11/6/89  11/8/89      9.6   9.2

  DW 38     4/4/89    4/6/89     17.5  17.8
  DW 38     8/22/89  8/24/89      1.3   1.1

  DW 47**   3/14/89  3/16/89      4.6   4.7
                           6.3    6.1      Post-Mitigation
                          13.6   13.6
                           1.0    0.8

                           1.8    1.9
                           Pre-Mitigation
                           Post-Mitigation

                           Pre-Mitigation
  * Time constraints did not permit pre-mitigation charcoal measurements
 ** Post-mitigation charcoal measurements could not be made due to homeowner's
constraints

-------
 O
  O
25-
20-
10-
5-
n.
CUATD-BSMT raATD-1 FLR ^CC-BSMT ZZ2CC-1 FLR




I.
t
'

i
i
i
i
i
;


1
>


X
1
               DW 01    DW 35       DW 38     DW 47

                              HOUSE ID

    Figure 1.  Pre-mitigation AID and CC results (no CCs at DW 35).
   10
o
a_
o
•z.
LJ
a:
o

LJ
a:
     0
   -5
  -10
                     — Basement Pressure Relative to Outdoors

                    —• Basement Pressure Relative to the Sub—slab
              Fan On
                                     Fan Off
                                                     Fan On
                                                    4/7/89
             4/3/89            4/5/89

                         DATE (Midnight)

Figure 2.  AP changes as the fan was turned off then back on in DW 47.

-------
0
 Q.
   20
    15
    10
O
<
     0
            Fan Turned Off
                                         Fan Back On
•—'Basement
  — First Floor
            4/3/B9           4/5/B9
                         DATE (Midnight)
                                   4/7/89
Rgure 3. Radon level changes as the fan was turned off then back
         on in DW 47.
                                        o—o First Floor
                                         —  Basement
             10/3/89
                                           10/7/89
                            10/5/89
                    DATE (Midnight)
Figure 4. Changes fn radon levels as the fan was turned on
         and as the basement was sealed in DW 01.
                               10

-------
                                —o Basement Relative to Outdoors
                                	Basement Relative to First Floor
                                — • Basement Relative to Sub-slab
      Door Closed
Door Closed
            10/24/89       10/25/89       10/26/89
                   DATE (Midnight)
   Figure 5.  Effects of pressure loss in basement through
             door left open in DW 01.
                                           o — o First Floor
                                             — Basement
             10/24/89       10/25/89       10/26/89
                   DATE (Midnight)
Figure 6.  Effects of pressure  loss on radon levels in  DW 01
                              11

-------
   60
   50
£40
LJ
   30
§  20
   10
  10/16/89
•—'First Floor
  — Basement
   • Rainfall
              10/18/89
                        10/17/89
                     DATE (Midnight)
Figure 7.  Effects of heavy rainfall on radon levels in house
          DW 35 mitigated with charcoal adsorption device.
                            12

-------
                                                                     Paper VH- 5
            ONE-YEAR FOLLOW-UP STUDY OF PERFORMANCE OF
                  RADON MITIGATION SYSTEMS  INSTALLED IN
                           TENNESSEE VALLEY HOUSES

                  C S. Dudney, D. L Wilson, R. J. Saultz and T. G. Matthews

                             Health and Safety Research Division
                               Oak Ridge National Laboratory
                                 Oak Ridge, TN 37831-6113
                                       ABSTRACT

       Subbarrier depressurization systems were installed for radon mitigation in two basement ranchers
in Oak Ridge, TN, and in two ranchers with partial basements in Huntsville, AL.  System performance
parameters, including pressure field extension, subslab permeability, and indoor radon concentrations, were
followed in each house for a year or longer. In general, these performance factors were stable over the
year. In one house built on a slab without underlying aggregate, the subslab permeability has increased in
two of four suction pits.  In another house, the extension of the field of depressurization is markedly
anisotropic. In three out of the four houses, mitigation measures that do not consume electrical power seem
to have provided 30 to 60% reductions in indoor radon concentrations.

       This paper has been reviewed in accordance with the U.S. Environmental Protection Agency's peer
and administrative review policies and approved for presentation  and publication.

-------
                                         INTRODUCTION

         Radon  is widely perceived as one of the most significant health risks present in indoor air (1).
 According to recent state surveys, radon levels in excess of the Environmental Protection Agency (EPA)
 guideline of ISO Bq m'3 are widely distributed throughout the US. (2). As a result, research programs have
 been supported on a regional basis to more quickly develop mitigation  technologies required for variant
 macro-geologic  and house-construction characteristics.  A radon research program in the Tennessee Valley
 has been sponsored by multiple agencies to:

         1.      better understand the physical basis of factors  underlying radon entry and its control,

         2.      refine diagnostic methods for implementation of effective radon control methods, and

         3.      improve radon control  technologies while systematically reducing  radon levels in study
                houses.

 In addition, a study of performance of radon mitigation  systems for one year following installation was made
 in two houses in Tennessee and two houses in Alabama.

         Several  research groups have evaluated mitigation system performance over extended periods of time.
 The principal means of evaluation to date have included post-mitigation radon monitoring with alpha track
 devices  and  follow-up  visits  to inspect the system for operational  integrity.  Nitschke (3) discovered
 numerous design flaws in early subsurface depressurization systems installed in New York houses. Prill (4)
 discovered in six houses in  Washington and Idaho that performance of subsurface  pressurization systems
 declined due to flow obstruction at  the outlet and due to occupant interference with vents and fans.

        Because of the substantial seasonal variation of indoor  radon concentrations that we  have seen in
the Huntsville study houses (5), the  pressure and flow characteristics of the systems were followed closely
to monitor system performance. In each of 4 houses  with basements,  12 or more sampling holes were
installed in the concrete slab and repeated measurements of radon concentration and air pressure were taken
at these  sites. In addition, flow and subslab permeability in  the  system  were followed.  This paper will
discuss what  changes were seen in the  pressure field and  indoor radon concentrations during  the first year
of operation.

                               DESCRIPTION OF STUDY  HOUSES

         The  study  houses were selected based on:  (1) the above research goals, (2) indoor radon survey
 data, (3) the availability of tightly clustered houses with similar construction characteristics, and (4) predicted
 homeowner cooperation with a 1-2 year, highly instrumented,  research program. The Oak Ridge/Knoxville
 area of eastern Tennessee, and the Huntsville area of Madison County, AL, were selected because of their
 comparatively high indoor radon levels in  state surveys (6), different  underlying geologies, and similar
 housing stock.  A  data base of prospective houses was generated in both locales from radon surveys of
 approximately 500  houses using charcoal monitors, and door-to-door screening using  continuous radon
 monitors.

        For the  larger study of radon entry and control, four basement ranchers in Oak Ridge, TN, and four
 crawlspace houses in the Garth Mountain area of Huntsville, AL, were selected. All eight study houses had
 substructural (i.e., crawlspace and/or basement) and superstructural levels of equal floor area to minimize
variability in  radon  entry, transport, and exhaust properties. The Huntsville houses (HU11-HU14) and Oak

-------
Ridge houses (OR1S-OR18) were clustered within a 0.4 and 1.2 km radius, respectively, to minimize the
variability in underlying geology and the time dependence of meieorologic events. The four houses included
in this study were:

House HU13   This is a single-story house, built around 1957 and rebuilt after a fire in 1978. The house
               has a substructure composed of crawlspace and walkout basement areas below  the living
               area.   There was no vapor barrier in  the  crawlspace and a continuous wall between
               crawlspace and basement areas with two doors.  There are two single garage doors opening
               into the basement area.  There is an attached  parking area with slab-on-grade construction.
               An electric furnace with forced-air distribution system is located in the basement.  An
               electric central air conditioning unit is located beside the rear wall.  The occupants routinely
               use a wood stove in the upstairs living area for heat during the winter months. The house
               is built on flat land  with few trees in the backyard. There are three adult occupants, one
               of whom was home most of the time.

House HU14:  This is a single-story house, built around 1962, with a split foyer entrance and a substructure
               composed of crawlspace and walkout basement areas below the living area.  There was no
               vapor barrier in the crawlspace.  There is a continuous wall, with a door, between the
               crawlspace and basement areas.  There a double garage door that opens into the basement
               and is opened twice per day for brief periods.  An electric  heat  pump  with  forced-air
               distribution system is located  in the crawlspace. An electric central air conditioning unit
               is  located beside the rear wall.  The occupant routinely uses an upstairs wood-burning
               fireplace in the wintertime.  The house is built on sloping land with many trees in the
               backyard. There is  one adult occupant who  was home evenings and weekends.

House OR15:  This is a single-story house, built in 1963, with a  full walkout basement below  the living
               area.  Three-quarters of the basement area is finished with panelled walls and tiled floors.
               There is an attached carport with slab-on-grade construction.  Beneath the carport slab is
               a fallout shelter  that can be entered from the basement.  An electric furnace with forced-
               air distribution system is located in the basement.   An electric central air conditioning unit
               is  located beside the side wall.  The house is built on sloping land with few trees in the
               backyard. There are two adult occupants, one of whom was home most of the time.

House OR16:  This is a single-story house, built in 1963, with a full  walkout basement below the living
               area.  Three-quarters of the basement area is finished with panelled walls and tiled floors.
               There  is an attached carport with slab-on-grade  construction.  An electric furnace  with
               forced-air distribution system is located in the basement. An electric central air conditioning
               unit is located beside the rear wall. The house is built on sloping land with many  trees in
               the backyard.  There are two adult occupants,  both of whom  were home during evenings
               and weekends.

                          DESCRIPTION OF MITIGATION MEASURES

WEATHERIZATION, HAC SEALING, AND SLAB SEALING

        The first stage of radon  mitigation performed in the Huntsville houses was simple weatherization
measures  and sealing of the heating  and air conditioning (HAC) system. All  houses were weatherized  to
a similar degree, consistent with standard Tennessee Valley Authority (TVA) recommendations and practice.

        In both Huntsville houses the return-side ductwork extended into all substructure! compartments.
Prior to house weatherization, inspection of the HAC systems  revealed leaky ductwork. For the return duct,

-------
leak testing was performed by two methods, physical inspection (i.e., for large holes or gaps) and by utilizing
a smoke pencil.  For small leaks, latex caulk, fingertip caulk, and duct tape were applied to seal the leaks.
In cases where joint leakage was due to nonsupport of the duct, sheet metal support bands were added.
In more severe cases (i.e. large holes or missing connections), heavy gauge sheet metal form connectors,
support cuffs,or cover plates were fabricated and permanently attached with sheet metal screws. These joints
were sealed with latex caulk as well.  In both houses, the pan-type return ducts were renailed to the floor
joist and the edges sealed with latex caulk  In addition, all other accessible metal to metal joints in the
return ducting were sealed with latex caulk.  Confirmation of sealing was made by using a smoke pencil.

        In both  Huntsville houses, the  central HAC unit was located in the crawlspace.  Smoke pencil
testing of the units revealed that the units were very leaky and drew crawlspace air into the HAC system.
Since  100%  sealing of  the units was impossible,  user-removable service panels were sealed with a
combination of modeling clay, fingertip caulk, and duct tape. Similar seals were fashioned for all inter-panel
joints and gaps  between the chassis and user-removable panels.  Other non-removable panels or non-
serviceable areas were permanently sealed with latex or silicon caulk. Effectiveness of the sealing effort was
judged by smoke stick testing. After all  direct surface seals were completed, the entire cold-air return duct
system was permanently covered with a  vapor-seal/fiberglass duct-jacketing.

        In all of these houses, the slabs were in  reasonably good condition with most deficiencies  being
small "hair line" cracks.  However, in HU14 there were some large (0.5 cm) cracks.  For houses OR1S and
OR16 access to the slab was limited since most of the basement area was  finished. All stress cracks in the
slabs  were expanded to a  V-shaped channel, 0.5 to 1.5 cm  in width and  1.5 cm  in depth  using a
percussion-type bit and  filled with flowable urethane sealant. If needed, an application of latex caulk or
liquid adhesive was used to fill large gaps to prevent bleed-through of the flowable sealant.  After the first
application of the flowable  sealant  had  aged for 3 days, an  additional layer was applied to  fill any
depressions or gaps which  had occurred during the curing process.  Cracks and perforations in the  block
walls were sealed with latex caulk or expansion foam. Wall outlets were sealed with a combination of latex
and fingertip caulk.  Slab-wall expansion joints were filled with  backer rod followed by flowable urethane
caulk.   In House  OR1S, the fallout shelter was  isolated from  the basement with a removable, weather
stripped plywood door.

SUBSLAB VENTILATION SYSTEMS

        All pits  for subslab ventilation were square in design, approximately 60 x 60 cm in size, and with
a depth chosen according to permeability test results.  Excavation depths of each pit depended upon  the
subslab permeability and the depth permeability profile data.  Standard 10 cm (4") or IS cm (6") schedule
40 PVC pipe was inserted within the pit to a depth of 30 cm beneath the bottom of the slab and  filled with
gravel (1 to 2 cm). A 60 x 60 cm cardboard shield was then inserted around the pipe and concrete was
poured and finished. The pit pipe was then attached to the main 10 cm (4") PVC pipe trunk equipped with
a Kanalflakt  pipe  fan.  Exhaust locations for HU13 and HU14 were  through the side of the house and
OR15 and OR16 were through the roof.  A detailed description  of the system  installed at each house is
given below.

House HU13  The substructure consists of a basement/crawlspace combination with a hollow  block wall
               separating the two compartments.  The basement construction consists  of hollow concrete
               block walls with interior walls  built over the slab.  A thick IS cm layer of aggregate was
               found under the slab, and excellent communication and permeability (140 to 190 x 10'7 cm2)
               were  observed between diagnostic holes and slab suction  points  within the  basement.
               Subslab to wall communication was not detected in the basement during subslab diagnostic
               tests. Due to the excellent communication and permeability found under the slab, only one
               50 cm deep mitigation pit was installed near the center of the basement slab.

-------
House HU14   The substructure consists of a basement/crawlspace combination with a hollow block wall
               separating the two compartments.  The basement construction consists of hollow concrete
               block walls with interior walls built over the slab.  Some (~8 cm) aggregate was found
               under the slab.  Subslab permeability was reasonable ranging  from 88 to 128 x 10*7 cm2
               between  diagnostic  holes  and slab suction points in the basement.  Subslab  to  wall
               communication was not detected in the basement during the subslab diagnostics experiments.
               Due to the relatively good communication among test holes in the slab, only one 65 cm
               deep subslab mitigation  pit was installed in the basement.

House OR1S   The substructure consists of a mostly finished basement with hollow block wall construction.
               All interior walls are built over the slab. Varying layers of aggregate were found under the
               slab (0 to IS cm) with  permeability ranging from 4 to 123 x Iff7 cm1 at different sites
               under the slab. Subslab to wall communication was not detected in the basement during
               subslab diagnostic experiments. Due to the varying depths of gravel found under the slab
               and wide ranges of permeability, two mitigation pits were installed.  Pit 1 was 40 cm deep
               and Pit 2 was  120 cm  deep.  Post mitigation radon measurements indicated that  the
               structure had not been sufficiently mitigated although diagnostic tests revealed an excellent
               pressure field extending  over 90% of the slab area.  Based on collected grab sample data,
               an additional source of radon within the structure was determined to be the attached fallout
               shelter (in excess of 3700  Bq m°).   Attempts were made to isolate the shelter from the
               structure by the sealing of the basement/shelter wall and  the entrance door.  When these
               measures failed, an 8 cm (3") PVC pipe was connected from the air exhaust of the shelter
               to the subslab mitigation system to depressurize the shelter. A control valve was installed
               in the 3"  line to allow  the room to be  maintained at a constant -10 Pa.  This action
               reduced the radon levels within the occupied structure to less than 100 Bq m°.

House OR 16   The substructure consists of a mostly finished basement with hollow block wall construction.
               AH interior walls are built over the footing with a poured concrete floor.  Varying layers
               of aggregate were found under the slab (0 to 5 cm) with permeability ranging from 0.1 to
               1.7 x 107 cm1 throughout the slab.  Subslab to wall communication was not detected in the
               basement during subslab  diagnostic experiments. Due to the varying depths of gravel found
               under the slab and the overall poor  permeability, a total of four mitigation pits of 120 cm
               in depth were installed.

CRAWLSPACE VENTILATION AND  MODIFICATIONS

        Passive ventilation of the crawlspace areas in HU13 and HU14 was attempted by the installation
of standard 20 cm (8') x 40 cm (16") vents.  The addition or  replacement of the vents did  require some
modifications to the substructures. Where below grade vents were required, gravel lined brick-walled wells
were installed.  Three below grade crawlspace vents were added on the side of the crawlspace opposite the
basement  in House HU13. Two additional above grade crawlspace vents were added to the front of the
house.  Three below grade crawlspace vents on the side  of the crawlspace opposite the basement and one
below grade vent on the north, rear corner of the house were added in House HU14. Two additional above
grade vents were added to the front of House HU14.

        Vapor barriers and active subbarrier ventilation systems were installed in the crawlspaces of the
houses. The crawlspace walls were prepared for the adhesion of the barriers by  scraping and then painting
the inner  walls and upper floor supports with gloss or semi-gloss masonry enamel.  A straight length of
8 cm (3*)  schedule 40 PVC pipe was laid roughly along the center length of the crawlspace floor with a
8 cm (3")  T placed approximately in the  middle of the pipe.  The ends of the main 8 cm (3') trunk were

-------
then capped and 2.5 cm (1*) horizontal holes were drilled in the side of the trunk at staggered 2 to 3 m
intervals.  Perforated 2.5 cm (1*) PVC pipe (0.6 cm hole per 20 cm of pipe)  was then inserted into the
trunk and cemented in place with epoxy and PVC cement. The barriers were then placed over the piping
system with 90 cm overlaps. Care was taken to custom trim the barrier to conform to the various supports
and obstacles in the crawlspace.  Sealing of vapor barrier material to itself and to the painted portions of
the walls was accomplished with double sided tape.  From the central trunk T, a series of adaptations were
made to connect the 8 cm (3") trunk to the 15 cm (6*) pipe fan.  Flexible ducting was then attached to the
exhaust side of the fan and run to an existing vent  to exhaust outdoors.  The flexible duct was hung from
the ceiling using permanent galvanized band-strapping.

                                  EXPERIMENTAL METHODS

       The study design included a variety of radiological, environmental, and house dynamic measurements
with  near-continuous and episodic data  acquisition  schedules.  Continuous measurements of radon,
temperature, and relative humidity were  performed in substructure! (Le., crawlspace and basement) and
superstructural  (Le., upstairs) zones.  Differential pressures were monitored between:  (1) substructure! and
superstructure!  zones, (2) substructure! and outdoor zones, and (3) basement and subslab zones.  These
monitors were also used on an episodic basis to  evaluate the impact of clothes dryers and HAC systems on
differential pressures across various building membranes.  Meteorologic measurements of wind speed, wind
direction, and rainfall were made in two locations per housing group; barometric pressure was measured in
one location. Episodic measurements included subslab radon concentration and permeability, and pressure
difference across the slab.  These techniques have been described by Dudney et aL (7).

       Subslab permeability was determined by applying a 10-fold or greater range of pressures to a hole
in the slab and measuring the resultant flow of exhausted air (8).  In premitigation  studies, suction was
applied with a  variable speed shop vacuum to a 3.8 cm ID hole in the slab.  Airflow was measured with
a hotwire anemometer inserted into a 2.6 cm ED PVC exhaust pipe.  In several houses, the permeabilities
of the subslab soil and soil/gravel mixtures were investigated as a function of depth  by coring the 2.6 cm
ID holes to depths of approximately 10, 40, 80 and  120 cm.  In post-mitigation studies, suction was applied
with a variable  speed mitigation blower through  10.2 and 15.2 cm holes in the slab. These holes connected
directly to SO x 50 cm gravel-filled pits having typically 40 and 150 cm depths for houses with and without
subslab gravel layers, respectively. Airflow was monitored with a hotwire anemometer inserted into a 13.9
or 15.2 cm ID exhaust pipe.  Induced differential pressure was monitored  at the point of entry into the slab
with an electronic vacuum gauge.

       The extension of the subslab pressure field was determine by applying controlled pressures to pre-
mitigation holes or post-mitigation pits and monitoring the induced pressure through  1 cm ID holes m the
slab  at several remote locations.  In the pre-mitigation diagnostic phase, negative pressures of 0.5, 2.0. and
5 0 kPa were used whenever possible to enhance comparisons between different houses and time periods (e.g.,
pre-  and post-mitigation phases). After mitigation, pressures were limited to about 0.35 kPa due to the
blower capacity.

        Figure  1 shows scale drawings of the basements and crawlspaces of the four houses described above.
The  locations of pits for the subsurface depressurization systems and of all sampling holes are indicated.

                                             RESULTS

        In a manner similar to Fowler (9), we monitored pressure field extension for nearly a year in these
houses.  Figure 2 shows the results from  the houses.  There was no discernible change over time in  the
induced pressure fields when the mitigation systems were operated normally.  What is most  noteworthy is
the clear  difference in pressure  field extension  under different sectors of the  slab.  In  House HU13, the

-------
 sampling sites closest to the driveway (1, 2,7, and 17) showed a gradual drop in pressure with increasing
 distance from the pit  The sampling sites on the crawlspace side of the pit (3, 9,10,13, IS, 16 and 18)
 showed a much steeper decline. At a distance of 7 m, there was measurable pressure on the driveway side
 but not on the crawlspace side.  Because of the lack of a close sampling site on the crawlspace side of the
 slab at House HU14, it is unclear whether the pressure field extension varies significantly with direction.
 Because multiple pits were installed at the Oak Ridge houses, only sampling sites at the ends of the houses
 were included in analysis of pressure field extension. Distances were computed from the nearest pit  For
'House OR16, distances were computed from Pits 2 and  3, which are those on the uphill side of the house.
 At House OR1S, the pressure field is easily measurable  at a distance of 5 m. It is dear from the data that
 there is an enhancement of pressure extension on the side of the house where the fallout shelter is located.
 Since a -10 Pa depressurization was applied to the  shelter,  the enhanced  extension  may  result from
 communication between shelter and subslab zones.  In House OR16, the pressure field falls off rapidly on
 the end by Pit 3 but not so  rapidly  on the end by Pit 2.

        For both  Huntsville houses (HU13 and HU14), data from sampling sites closest to pit showed a
 decline in induced pressure during the first several months of operation (see Table 1). This effect was only
 seen in the holes where the induced pressure was largest and easiest to measure. This decline in pressure
 following initial start-up may reflect the removal of fine paniculate material from the flow paths and the
 consequent absence of some flow restrictions and pressure drops near the point of applied depressurization.

        Using the techniques described by Matthews (8), permeability was monitored at all suction pits on
 a recurring basis.  Detailed  analysis has not been completed on all available data, but there is no clear
 evidence of change in subslab  permeability except at House OR16 (see Table 2).  The slab at this house
 was poured directly onto packed clay.  We suspect that severe drought conditions during 1988 may have
 contributed to shifting of the  soil beneath the  slab, resulting in the opening of channels leading to the
 suction pits.  Continued monitoring during 1989 (data  not shown) has revealed continued increase in the
 permeability in Pits 1 and 4.  The  1989 results and results from tracer gas experiments will be reported
 separately.

         Radon  reduction efficiency was quite  good in these houses following initial installation and
 adjustment of the systems (see Table 3).  Reductions ranged from 89 to 97% in 1988. During 1989, there
 was an extended period during which the electric power to the systems was cycled on for a week followed
 by a week without power.  In Figure 3, the data from the period around  Julian Dates 80 to 160 (1989)
 include most of the period during which the power was cycled on and off.  The comparison of power-off
 radon concentration to a priori radon concentrations suggests that, except for House HU13, from  40 to 80%
 of the reduction may be due to the passive effects of modifications that were made to the houses.  The
 reader should note that this comparison is suspect due  to strong seasonal variations that can occur in this
 area. This effect is easily seen  (see Figure 3) in the changes in indoor radon that occurred in House HUM
 during the first 50 days of monitoring.

         In summary, this study has  shown that there is generally  very good stability in the performance of
 radon mitigation systems that depressurize the soil beneath concrete slabs in three of these existing houses.
 One house built  on a slab  without aggregate (OR16)  was a notable  exception.  There was evidence of
 domains beneath the slab of House HU13 with distinctly different extensions of the field of depressurization.
 There is evidence present that in three of these houses passive measures,  such as  slab sealing, can yield
 considerable radon reduction.

-------
                                  ACKNOWLEDGEMENTS

       The  authors are very grateful to the  homeowners in whose houses these data were collected.
Without their patience, this study would not have been possible. This research was sponsored by the Office
of Health and Environmental Research of the U.S. Department of Energy, by the Tennessee Valley
Authority under Interagency Agreement No. 40-1602-85, and by the U.S. Environmental Protection Agency
under Interagency Agreement No. 1824-1709-A1 under Martin Marietta Energy Systems, Inc., contract DE-
AC05-84OR21400 with the U.S. Department of Energy.
                                                      us
                                         torn of Ha
                                    «h« te fc ». to

-------
                                        REFERENCES

1.     National Research Council. Health risks of radon and other internally deposited alpha-emitters.
       Washington. DC:  National Academy Press; 1988.

2.     U.S. Environmental Protection Agency and Centers for Disease Control   A citizen's guide to
       radon:  What it is and what to do about it Report #OPA-86-004 (GPO 055-000-00258-4). 1986.

3.     Nitschke, L A.; Clarion, M. E.; Brennan, T.; Rizzuto, J. E>; "Long-term assessment of residential
       radon-mitigation systems.*  Paper #107.5 in Proceedings of 1988 meeting of the Air Pollution
       Control Association; 1988.

4.     Prill, R. J.; Fisk, W. J.; Turk, B. R;  "Monitoring and evaluation of radon mitigation systems over
       a two-year period.1 Report #LBL-25909 (OTIS DE 89 00291).  1988.

5.     Gammage, R. B.; Wilson, D. L; Dudney, C. S.; Sanltz, R. 14 Summertime elevation of "Rn levels
       in Huntsville, AL; submitted to Health Physics. 1990.

6.     Ronca-Battista, M.; Moon, M.;  Bergsten, J.; White, S. B.; Alexander, B.; Holt, N.  Radon 222
       concentrations in the United States • Results of sample surveys in five states.  Radiation Protection
       Dosimetry 24:307-311; 1988.

7.     Dudney, C S.; Hubbard, L M.; Matthews, T. G.; Socolow, R. R; Hawthorne, A R.; Gadsby, K. J.;
       Harrje, D. T.; Bohac, D. L; Wilson, D. L  Investigation of radon  entrv and effectiveness of
       mitigation  measures in seven houses in New Jersey. Report 0ORNL-6487.   1989.

8.     Matthews,  T. G.;  Wilson, D. L.; TerKonda. P. K.; Saultz,  R. J.;  Goolsby, G.; Burns, S. E.;
       Haas, J. W.  Radon diagnostics:  Subslab  communication and  permeability measurements.   In
       Proceedings of the 1988 Symposium on Radon and Radon Reduction Technology, Volume 1, EPA-
       600/9-89-0063 (NT1S PB89-167480),p. 6-45, March. 1989.

9.     Fowler, C  S.; Williamson, A D.; Pyle, B. E; Belzer, F. E.; Sanchez, D. C; Brennan, T.; 'Sub-slab
       depressurization demonstration in Polk County, Florida, slab-on-grade houses.* In Proceedings of
       the 1988 Symposium on  Radon and Radon Reduction Technology, Volume 1, EPA-600/9-89-006a
       (NT1S PB89-167480),p. 7-65, March, 1989.

-------
HIM 3
*1
Garage ,*


f 2



Garage
7

Scale: -^ 	
::*:::
*T*
SSO Pit J
© J 	
^ V
3 L 	
* j
,18 t
•• r


Slab Soil w/ Barrier

*
10


*
13
*
15


           5m
HU14
*
3
*
2



+ (

Scale:
5 fn
SSDP
* ©
5



3arage

*
6
t



*
4


*



.••'

7
8
...j.
I
..J.
'•• 1 *
1 9
1
Crawlspace
^ Soil w/ Barrier






OR15
           ,,*  (£
                L_

               I
                        \m
                                      Scale:
                                       4m
                                *
                                i
OR16
                                                                                       Scale:
* I*
6 20
5
4
, ,*,,-./
* 1R
-* J^
*8
rF"11
Pit 2
TITf

1 Pit 4
2
^3*
V
Vs
* 1O
12 *
11
*
13f-
   Figure 1.  Schematic of substructures indicating depressurization points and sampling locations.

-------
HU13
                                   Data from Sites:
                                     1.Z7.17
                                   Data from Sites:
                                3. 9.10. ia 15.16.18
HUM
OR15
OR16
                                      Data from Sites:
                                      8.9.10. 11.12
                                      Data from Sites:
                                         1.Z3.4

                                                                                               Data from Sites:
                                                                                               4.5.6.8.16.20
                                   Data from Sites:
                                7.10.11. 1£ 13,14. 15
                  Figure 2.  Variation of pressure with distance under slab for four houses.

-------
cr
m
(0
oc
          Mt   M
       1987
                                               1989
                          Julian Date
                           OR15
     "?9B7
                           1 _______  •
                              at   M   at
                        Julian Date
                                                                                                              1989
                                                                                          Julian Date
X)   X3   M


  1987
                           Figure 3.  Daily average radon concentrations in four houses.
                                                                                       Julian Date

-------
           Table 1. DIFFERENTIAL PRESSURE (Pa)
                  AFTER INSTALLATION.
                    House/Sampling Site
             Date      13/3         14/5

              8/88   -28 (113%*)   -43 (118%)
             10/88   -25 (103%)   -39 (106%)
             11/88   -23 ( 96%)   -35 ( 97%)
             12/88   -22 ( 89%)   -36 ( 98%)
              1/89   -23 ( 93%)   -38 (105%)
              2/89   -22 ( 89%)   -35 ( 97%)
              7/89   -24 (100%)   -37 (100%)

             ' Fraction of 7/89 value, expressed
               as a percentage.
Table 2. VARIATIONS IN SUBSLAB PERMEABILITY (x 107 cm'2)
                      AT OR16 IN 1988.


  Date  Pit 1         Pit 2        Pit 3         Pit 4

   5/88   4.0 ± 0.5      7.4 ±  1.8     5.1 ± 1.3

   6/88    —          —          —        10.3 ± 1.2

  11/88  16.7 ± 1.0      5.5 ±  0.5     5.1 ± 1.3     20.0 ± 1.1

  12/88  14.6 ± 0.9      5.2 ±  0.5     5.1 ± 0.8     18.8 ± 2.0

-------
Table 3.  RADON REDUCTION EFFICIENCY: 1988 AND 1989.


          1988          1988      1989        1989
  House Pre-Mitipation Post-Mitipation System Off System On

  HU13    191 Bqm4      21      235        32

  HU14    958           30      220        63

  OR15    585           26      218        35

  OR16    724           78      315        100

-------
                                                                            VII-6
A COST-EFFECTIVENESS COMPARISON OF PRIVATE-SECTOR RADON REMEDIATION

             WITH TRADITIONAL RADIATION PROTECTION ACTIVITIES

                            by:    Daniel J. Strom
                                  William D. Ulicny
                                  Department of Radiation Health
                                  Graduate School of Public Health
                                  University of Pittsburgh
                                  Pittsburgh, PA 15261

                                  John B. Mallon, Jr.
                                  Richard W. Benchoff
                                  Radon Detection & Control
                                  P.O. Box 419
                                  South Heights, PA 15081



                                      ABSTRACT

  Private-sector radon remediation can be well over 100 times more cost-effective than the
  minimum expenditure of $1000 to avert a whole-body person-rem mandated by the U.S. Nuclear
  Regulatory Commission for marginal control of exposures to radiation due to effluents from
  nuclear power plants (10 CFR 50 Appendix I). Using data from the Pennsylvania Department
  of Environmental Resources (DER) and from Radon Detection & Control (RDC; a commercial
  radon remediation firm) spanning more than two years, we relate differences in radon
  concentrations measured before and after remedial action to capital and operating costs.  Simple
  demographic assumptions and concentration-to-collective dose equivalent  conversions lead to a
  quantitative demonstration of the dramatic imbalance in society's valuation of radiation
  protection by nuclear utilities when compared to that in the home.  The  DER data show that an
  effective person-rem can be averted  for $3.01 in high-radon houses (104 houses; average, 189
  pCi/L; median 112 pCi/L; GSD 2.75), while the RDC data show that an effective person-rem
  can be averted for  $9.41 in more typical houses (201 houses; average 28 pCi/L; median 19;
  GSD 2.31).

-------
                                    INTRODUCTION

       A basic tenet of public health activities is that limited resources should be used in a way
that is most beneficial to  the most people.  Studies on the value of a human life have
demonstrated the great incongruities in how money is spent for increments of health and safety
(1,2). In this paper, we demonstrate that radon remediation activities can be 100 to 300 times
more  cost effective than radiation protection activities at nuclear power plants required by U.S.
Federal regulations (3).

       In the case of nuclear power, the risk is perceived as dread, unknown, uncontrollable,
imposed, and manageable with someone else's money.  In the  case of radon exposure in the
home, the risk is perceived as less dread, better known, more controllable, voluntarily accepted,
and manageable only with one's own money.  The traditional perception of the home as a "safe
haven" from the ills of the outside world runs counter to the acceptance of a serious
radiological risk that may exist from natural (or, rather, technologically-enhanced  natural) causes.
The public health perspective, however, dictates that society's resources should be refocused on
indoor radon, rather than  on tiny increments of safety achieved at tremendous cost in controlling
effluents  from nuclear power plants.

                                         DATA

       Two sets of data are used in this study.  The fust is from the  Pennsylvania Department
of Environmental  Resources (DER) report summarizing the Pennsylvania Radon Research  and
Demonstration Project (4).  This report lists radon concentration data from 104 houses for which
measurements before and after remediation were available.  The DER  data arose  from research
and development work, and the houses to be remediated had very high radon concentrations
(many over 200 pCi/L).  These data are  not a random sample  of houses, and remediations were
done  in the mid-1980s  when the technology was in its infancy. Concentration measurements
were  made using grab or continuous flow-through  alpha scintillation measurements.  There is no
evidence that measurements were made using integrating detectors such as charcoal,  diffusion
barrier charcoal, or alpha track.

       The second set  of data are from Radon Detection & Control (RDC).  These  data
included  201  houses  remediated between 1987 and  1989, mostly in the Allegheny and Beaver
County areas of western Pennsylvania.  The RDC  data represent a more typical set of radon
remediations in that they  were done  for real estate transactions, for relocation companies,  or for
health concerns of the occupants.  The RDC  radon measurements were made using activated
charcoal screening detectors' exposed for two to four days and  processed commercially.  In
some  cases, the concentrations are averages of two or more measurements.  The  radon
concentrations were measured at the same location before  and  after remediation.

                             ANALYSIS AND DISCUSSION

       The DER  data are summarized in Table la, while  the RDC data are  summarized in
Table  Ib.  From left to right, the columns  are the initial or "Before" remediation radon
    'Key Technology, Inc., P.O. Box 562, Jonestown, PA 17038; or Air Chek. Inc., P.O. Box 2000, Ardcn,
NC 28704.

-------
 concentration; the final or "After" remediation concentration; the remediation cost; the percent
 reduction of radon concentration; the percent reduction in radon concentration per $1000 spent
 on remediation; the change in concentration; the change in concentration per $1000  spent on
 remediation; the ratio of the "Before"  to the "After" concentrations; and the ratio reduction
 achieved per $1000 spent on remediation.  Note that minimum and maximum entries in different
 columns may be from different  houses; these values were the minima and maxima in the entire
 data set.
  Table la. Summaries of the Pennsylvania D.E.R.data (Weston & Simon 1988).
               Radon
                Cone.
               Before
 % Re-
duction
 Cone.
Change
Reduct.
 Factor
 Cone.   Cone. Reined-           duction     Cone.    Change            Factor
 Before    After   iation   %Re- per 1000$   Change  per 1000$  ReducL per 1000S
(pCi/L)  (pCi/L) Cost($)  duction     %/k$   (pCi/L)  (pCi/L)/k$   Factor     (/k$)
No. of houses
Minimum
Maximum
Mean
Std.Dev.
Geo. Mean
G.S.D.
r*2 (lognorm.)
104
11.6
1653.2
189.1
240.2
112.3
2.75
0.992
104
0.3
178.8
10.6
27.7
2.5
4.43
0.930
104
1301
16185
3729
2156
3324
1.57
0.966
104
-53.2
99.9
90.4
23.2
92.7
1.21

104
-13.0
73.7
29.6
14.2
27.8
1.59

104
-62.1
1576.2
178.5
231.3
107.4
2.91

104
-15.1
312.3
47.4
49.7
32.3
2.61

104
0.65
893.0
112.4
145.0
44.8
4.96
0.972
104
0.16
222.9
33.3
40.7
13.5
4,98

 Table Ib. Summaries of Radon Detection & Control data, 1987-89.
               Radon   Radon                    % Re-
               Conc.   Cone. Remed-           duction
               Before    After   iation   %Re- per 1000$
              (pCi/L)  (pCi/L) Cost($)  duciion     %/k$
                    Cone.
          Cone.    Change
         Change  per 1000$
         (pCi/L) (pCi/L)/k$
                 Rcduct.
                  Factor
       Reduct.  per 1000$
        Factor      (/k$)
No. of houses
Minimum
Maximum
Mean
Std.Dev.
Gco. Mean
G.S.D.
r*2 (lognorm.)
201
4.2
156.0
28.1
29.0
19.2
2.31
0.969
201
0.0
18.3
2.0
2.4
1.3
2.53
0.993
201
150
2905
1030
325
957
1.55
0.862
201
-2.7
100.0
87.3
15.8
86.1
1.24

201
-1.4
634.4
99.0
63.4
90.3
1.47

201
-0.3
155.6
26.1
28.4
16.6
2.59

201
-0.2
380.7
27.7
37.3
17.4
2.50

199
1.0
390.0
29.3
46.8
14.5
3.21
0.997
201
0.0
421.1
30.3
52.9
15.4
2.97

       Under each column are the number of houses used in that column; the minimum,
maximum, arithmetic mean, and standard deviations of each quantity; and the geometric mean
and geometric standard deviations (GSD; dimensionless) of the  quantities.  For the DER data,
the geometric means and GSDs for the difference, % reduction, concentration change per $1000,
and percent per $1000 were calculated from 101 data pairs, omitting the 3 negative values.  For
the RDC data, geometric means and GSDs were calculated from  197 to 199 non-negative values

-------
for those columns with negative entries, or, in the case of the ratio, with zero final
concentration.

       The DER houses had  larger average initial ("Before" remediation) concentrations than did
the RDC houses, with arithmetic means of 189 ± 240 and 28 ± 29 pCi/L, respectively.  This is
due to the different selection  processes, and to the fact that  many  of the DER houses were in
the Reading Prong area of southeastern Pennsylvania, a region with exceptionally high indoor
radon levels.  Both data sets  showed lognormal distributions, as seen in Figures 1  and 2, with
geometric means of 112 I 2.75 (r2  =  0.992) and 19 ; 2.31 (r7  =  0.969).  Lognormal fits  were
done by the LPROBIT code using  Finney's weighting method (5).

       The maximum in  the  DER  data set, 1653 pCi/L, was more than ten  times higher than
the maximum in the RDC data set, 156 pCi/L.  The minimum "Before" value (i.e., 4.2  pCi/L)
in the RDC data may indicate the  force that the real estate and  relocation companies have on
radon remediation decisions.

       Final ("After" remediation)  concentrations averaged 10.6  and 2.0, respectively, for the
DER and RDC data.  The lower final concentration  for RDC  is understandable in  light  of the
fact that they usually guarantee a concentration  below 4 pCi/L, and usually  continue to
remediate  until  the concentration is below this level.  Again, both  data sets  showed lognormal
distributions, with geometric means of 2.5  ; 4.43 (r2 = 0.930) and  1.3 : 2.53 (r2 =  0.993). The
wide GSD, 4.43, for the DER data, along with its moderate deviation from  lognormality may
reflect the varying conditions applied to remediation during the DER's research and development
process.

       The DER data show significantly higher average remediation costs than the RDC data:
$3729  ± $2156 vs.  $1030 ± $325.  The geometric means were $3324 : 1.57 (r2 = 0.97) vs.
$957 I 1.55 (r*  = 0.86).  These numbers are understandable  in light of the differing nature of
the two endeavors:  the DER project was a research and development activity, done to develop
technologies; the RDC work is done  commercially in a competitive atmosphere, and benefits
from well-established practices in remediation.

       We chose three ways  of looking at the data:  percent  reduction, concentration change,
and the reduction factor.  The percentage reduction is the difference in  radon concentrations
divided by the  "Before" concentration, expressed as a percent; the  concentration change  is
simply the difference between "Before" and "After" concentrations; and  the  reduction factor  is
the ratio of the "Before"  to "After" concentrations.   All three methods are somewhat flawed
because no concurrent long-term background concentration measurements were available. Since
indoor radon levels cannot be reduced below ambient outdoor levels using any of  the
remediation methods employed for DER or by RDC, both percent reduction and reduction factor
have values limited by background concentration.  The limits  are unimportant for high initial
concentrations,  but can restrict measures of success when starting from lower concentrations.

       The percentage reduction summaries are given for each data set in Table 1.  The means
are 90.4% for the DER data, and 87.3% for the RDC data.   The mean  percentage reduction is  a
somewhat misleading statistic, since its maximum value is 100,  with many houses very  near this
value.  The mean percentage reduction per $1000 spent on remediation is 29.6%/$1000  for  DER
vs. 99.0%/$1000 for RDC,  which primarily reflects the different average remediation costs.

-------
                             Pennsylvania DER Data


                         With Finney Regression Predictions
    o
    ^ft


     ti
     o
    •l-t
    -p
     cd
     o
     a
     o
    u

     ti
     o
    -d
     cd
        1000
100
          10
          0.1
                      Pre
            0.13
            2.28
                              15.87     50      84.13     97.72    99.87

                                    Percentile


Figure 1.  Log-probability :plot of "Before" and "After" radon concentrations in 201 houses

remediated by Radon Detection & Control, with Finney-weighted regression lines.


                        Radon Detection & Control Data


                         With Finney Regression Predictions
                                            Avg.  Geo.Mean  GSD


                                  Before    28.1     19.2    2.31


                                  After       2.0      1.3    2.53
                      2.28    15.87    50     84.13    97.72    99.87


                                    Percentile
Figure 2.  Log-probability plot of "Before" and "After" radon concentrations in 201  houses

remediated by Radon Detection & Control, with Finney-weighted regression lines.

-------
       The concentration change summaries show arithmetic means  of 179 ± 231.3  pCi/L for
 the DER data and 26.1 ± 28.4 pCi/L for the RDC data, the difference largely reflecting the fact
 that the DER houses had higher concentrations to begin with.  The geometric means were 107
 pCi/L ; 2.91 and 16.6 pCi/L I 2.59, respectively. The mean concentration change per $1000
 was 47 ± 49 [pCi/L]/k$ and 28 ± 37 [pCi/L]/k$  for the DER and RDC data, respectively,
 indicating that, while the concentration changes in the DER work averaged nearly 7  times
 greater than the RDC work, the remediation cost per unit concentration change averaged only
 1.7 times greater for the DER work.  Note  that  this  analysis does not include operating and
 maintenance  costs.

       Perhaps more interesting than the percentage  reduction is the reduction factor.  Reduction
 factors as great as 893  were achieved in the DER program, and 390 in the RDC work.
 Average reduction factors were 112 and 29, respectively, with geometric  means of 45 I 4.96 (r2
 = 0.972) and 29 ; 3.21 (r2 = 0.997), indicating both  more vigorous remediation efforts in  the
 DER demonstrations and the limiting effects of ambient radon concentrations when remediating
 houses with initially low radon levels. Of great interest are the mean reduction factors per
 $1000 spent  on remediation: these were 33/S1000 and 30/S1000 for DER  and RDC,  respectively,
 with geometric means of 13.5/S1000 and 15.4/S1000, indicating that  the work was quite
 comparable in terms of cost-effectiveness when expressed  in terms of reduction factors!

       The dose equivalent averted through remediation is directly proportional to the
 concentration change.  To evaluate the overall cost-effectiveness of the two remediation projects,
 we summed the "Before" and "After" concentrations, as well as the costs, as shown  in Table 2.
 These concentration  sums can be considered as "Collective Concentration" by  analogy with
 "Collective Dose" and "Collective Dose Rate" as defined by the International Commission on
 Radiological  Protection  (6).

 Table 2. Sums of concentrations and costs, with analysis statistics for PA D.E.R. and RD&C data.

              Sum of  Sum of                  Overall              Overall           Overall
              Radon   Radon   Sum of            %Re-    Overall     Cone.           Reduce
               Cone.   Cone.  Remed- Overall    duction     Cone.    Change  Overall     Factor
              Before   After   iation  %Re-  per 1000$    Change  per 1000$  Reduct  per 1000$
              (pCi/L) (pCi/L) Costs ($) duction     %/k$    (pCi/L)  (pCi/L)/kS   Factor      (/k$)

 PAD.E.R.      19664    1102  387797    94.4      25.3     18562      47.9    17.8        4.8
 R.D.&C.       5649     402  206936    92.9      90.2      5248      25.4    14.1       13.7
       The columns in Table 2 that begin with the word "Overall" differ from the mean or
geometric mean values presented in Table 1 because they are computed only from the sums
given in  Table 2, rather than from the actual distributions.  The overall percent reduction in
each data set (94.4% and 92.9%, respectively, for DER and RDC) is greater than the mean
because greater reductions were achieved in  high initial concentration houses than in low
concentration houses.  The percent/Si 000 values (25% and  90%) are based  on the mean
remediation costs in Table  1, and are lower  than the means in that table.

-------
       The collective dose equivalent averted by remediation is proportional to the overall
concentration change, which  was 18562 pCi/L for the DER work and 5248 pCi/L for the RDC
remediations.  While the DER overall concentration change is 3.5 times higher than the RDC
change, the cost-effectiveness value for DER is only 1.9 times that for RDC (47.9 vs. 25.4
[pCi/L]/$1000) based  on average remediation costs from Table 1.

       The overall reduction factors were 17.8 and 14.1, respectively, for DER and RDC, while
the reduction factor per $1000 values were 4.8/$1000 and 13.7/$1000, indicating an almost
three-fold better investment with the later work when viewed in  these terms.

       Rather than dealing with risk, as the  DER report does, we compute the effective dose
equivalent averted by remedial actions,  and compare the cost per unit effective dose  equivalent
averted by remedial actions.   The many assumptions needed to do this are summarized in Table
3.  We calculate  1.93 lung-rems (to an adult) from a 1-year exposure to a concentration of 1
pCi/L  of radon from the product (1  WL/[200 pCi/L])(8766 hours/environmental year)(l
occupational month/170 hours)(0.75  fraction of time spent at home)(0.5  rad/WLM)(20 rems/rad =
Quality Factor for alpha radiation).  Effective dose equivalent rate per unit concentration (0.23
effective rems/([pCi/L]-year) is  obtained by multiplying by the ICRP weighting factor for the
lung (6), i.e., w,., = 0.12.


Table  3.  Assumptions used in calculating the expenditure per person-rem.

  200   [pCi/L]/WL, i.e., 50% equilibrium
   0.5   rad/WLM (ref. 7)
   20   Q (rem/rad;  ref. 6)
  0.12   w,,,,,,, i.e.,  effective rem  per lung rem (ref. 6)
  0.75   fraction of time spent at home
 8766   hours/environmental  year
   170   hours/occupational month
    4   persons per  house
    10   years life  expectancy for radon  control system
   125   $/year energy penalty (ref. 11, modified)
   29   Watt fan'
  0.10   S/kWru*

  1.93   lung rems/([pCi/L]-year)
  0.23   effective rems/([pCi/L]-year)
  62.7   effective microsieverts/([Bq/m3]-year)
25.42   $/year, fan operation cost


1  Dayton Electric Mfg. Co., 5959 W.  Howard St., Chicago IL 60648;  Model 4C720.
b  Duquesne Light Co., Pittsburgh, PA; 1989
        The value of 0.5 rad/WLM  is taken from Harley and Cohen (7), and is probably an
 underestimate of the conversion for children (8,  9), for which estimates as high as 1.2 rad/WLM
 are available (8).  Additionally, the value of 0.75 of one's time in the house is somewhat higher

-------
than the values of 0.64 to 0.73 used by some at the EPA (10); however, the EPA work is based
on data which dramatically under-represent infants, small children, and the elderly.

       Additionally, we assume that there are 4 persons per house, and a ten-year life
expectancy for the remediation system.  We further assume $125/year for operating costs and
the cost of loss of conditioned air.  Due to the  use of a 29 W fan by RDC, rather than the
more common 90 W fan, we have lowered our estimate of annual costs below the $150
estimated by Henschel (11).  It is important to  point out that the annual cost is not well  known,
and is  limited  at the lower end by the fan operation cost ($25.42), but may be much  greater
than our estimate in some cases.

       A critical assumption in our calculations is that the  change in radon concentrations  are
representative of the air that house occupants actually breathe. We do not know  the proportion
of houses  for which this is the case.  In some cases, the radon measurements are simply
screening values measured in the lowest livable area of a house;  in other  cases, they are
representative measurements  of air that the occupants  breathe.  If a ratio for basement
concentration to first floor concentration of 2.5  is used (12), then our calculated average  cost per
(unit collective effective dose equivalent [CEDE] averted) could be multiplied by no more  than
this factor.

       The average cost per unit CEDE averted is tabulated in Table 4.  The remediation  cost
plus the operation and maintenance costs are added to yield totals of $4979 for the DER work
and $2280 for the RDC remediations.   Average concentration changes are multiplied by 0.23
effective rem per ([pCi/L]-year) and multiplied  by 4 persons and 10 years to get  the 10-year
CEDE averted values of 1657 person-rems per  house  for the DER houses and  242 person-rems
for the RDC houses.  These values, divided by the total 10-year cost, give the average cost per
unit CEDE averted, in dollars per person-rem.  The values  for the DER and RDC houses are
$3.01/person-rem and $9.41/person-rem, respectively.
Table 4.  Average cost per unit collective effective dose equivalent averted in radon
remediation and as  required of the nuclear power industry.
                                                                    Average
           Cost of                Average     10-year CEDE*    Cost per unit
            10-year              Rn Cone.
            Energy  Capital        Change
            Penalty  Outlay  Total  (pCi/L)
             $1250   $3729 $4979    178.5
             $1250   $1030 $2280     26.1
PA D.E.R.
This Study
10 CFR 50"
10 CFR 50?
10-year CEDE'
      Averted
  (person-rems
       /house)
         1657
          242
       CEDE'
      Averted
($/person-rem)
        $3.01
        $9.41
     $1000.00
    $33333.33
1 Collective Effective Dose Equivalent
b (App.I Sec.H.D.; whole body)
c (App.I Sec.H.D.; thyroid)
       The U.S. Nuclear Regulatory Commission has set values of $1000 per whole body rem
 and $1000 per thyroid rem as the minimum amounts that must  be  expended to limit offsite

-------
releases of radioactivity  by nuclear power plants (3).  The latter value can  be converted to
effective dose equivalent by dividing by the ICRP risk-based weighting factor for the thyroid,
i.e., 0.03, giving $33,333.33.  The NRC cost per unit CEDE averted values are roughly 300 and
10,000 times greater than the DER values, and 100 to 3,000 times the more realistic RDC
values.  Clearly, these numbers are inconsistent.

       If a cancer fatality rate of 2 x  10"* per person-rem is assumed, dollar values to save a
human life can  be computed.  These are $15,000 and $47,000 for the DER and RDC
remediations, and $5,000,000 and $167,000,000 for the NRC release limits.  The radon
remediation costs are in line with other common societal expenditures for life saving (1,2), but
the NRC  values are not.


                                       CONCLUSIONS

       By analyzing two sets  of radon remediation data, we have rhown that remediation can
be a cost-effective activity when compared to other life-saving  measures  in our society.
Expenditures in other areas of radiation protection, such as those  required  of nuclear power
plants by the U.S. Nuclear Regulatory Commission, can be hundreds or even thousands of times
less cost-effective.

                                        REFERENCES


1      Cohen, BJL.  Society's valuation of life saving in radiation protection and other contexts. Health
       Phys. 38:33-51, 1979.

2      Graham, J.D., and Vaupel, J.W.  Value of a life:  What difference  does it make?  Risk Analysis
        1:89-95,  1981.

3.     U.S. Nuclear Regulatory Commission.  Title 10, Code of Federal Regulations, Part 50. Appendix I.
       Originally published in 40 FR 40816, but still in effect in 1990.

4.     Roy F. Weston, Inc., and R.F. Simon Company, Inc.  Final report of the Pennsylvania radon
       research  and demonstration project. Pennsylvania Department  of Environmental Resources, bureau
       of Radiation Protection, Harrisburg, PA, 1988.

5.      Strom, D.J. LPROBIT.  Health Phys. 57:VII-VIII. July, 1989.

6.      International Commission on Radiological Protection.  Recommendations of the International
        Commission on Radiological Protection. ICRP Publication No. 26.  Oxford: Pergamon Press. 1977.

7.      Harley, N.H. and  Cohen, B.S.  Updating radon daughter bronchial dosimetry.  In:  P. Hopke  (ed.),
        Indoor Radon.  American Chemical Society Symposium Series 331. Washington, DC: American
        Chemical Society,  1987.

8.      National Council on Radiation Protection and Measurements.   Evaluation of occupational and
        environmental exposures to radon and radon daughters in the United States. NCRP Report  No. 78.
        Bethesda, MD:  NCRP Publications, 1984.

-------
 9.      Hofmann, W., Steinhausler, F., Pohl, E. Dose calculations for (he respiratory tract from inhaled
        natural radioactive nuclides as a function of age - 1.  Health Phys.  37:517-532, 1979.

 10.     Exposure Assessment Group,  U.S. Environmental Protection Agency.  Exposure Factors Handbook.
        EPA/600/8-89/043.   Washington,  DC: United States Government Printing Office, 1989.

 11.     Henschel, D.B. Radon reduction techniques for detached houses. Technical guidance.  2nd ed.
        EPA/625/5-87/019.   Washington,  DC:  United States Government Printing Office. 1988.

 12.     Cohen, B.L.  Variation of radon levels in U.S. homes with various factors. Journal of the Air
        Pollution Control Association  38:129-134, 1988.
       The work described in  this paper was not funded by the  U. S.
Environmental  Protection Agency and therefore  the  contents do not  necessarily
reflect  the views  of the Agency and no official endorsement should be
inferred.

-------
     THE  EFFECTIVENESS  OF  RADON  REDUCTION  IN NEW JERSEY   VII-7

           By:   Nick  DePierro, Tonalee Key  and Jennifer Moon
           New Jersey Department  of Environmental Protection
                  CN  415, Trenton, New Jersey  08625

                              ABSTRACT


     The  results  of  473 homeowner-funded radon mitigations conducted
 in New Jersey from April,  1988 to November, 1989 are reported and
 compared  to similar  results previously reported for the period
 January,  1986 to  March  1988  (1988 Study).  Data presented is
 compiled  from the New Jersey Department of Environmental
 Protection's (DEP) post remediation testing program.  A review of
 the data  compiled from  the post  remediation testing program
 indicates that  soil  depressurization is still the predominant
 mitigation method employed by certified firms, while soil
 depressurization  and sealing only methods tie for predominant
 mitigation method employed by homeowners and non-certified firms.

     Lowest floor radon concentrations following mitigation were
 below 4 picocuries per  liter (pCi/1) in 56% of the homes, compared
 to only 36% in  the 1988 study.   This increase may be explained by
 the fact  that radon  mitigators are more experienced and knowledgeable
 in radon  diagnostics and mitigation design.  The average cost of
 mitigation, as  determined from reports submitted by mitigation
 firms, was $1,100 compared to $1,300 in the 1988 study.


                             Introduction

     In order to monitor and evaluate effective radon reduction
 techniques for  homes, the DEP has developed a program to track radon
 mitigation efforts.   Mitigation  techniques monitored through this
 program include sealing only, soil depressurization, and forced air
 and crawl  space ventilation.  The latest procedures and guidelines
 for this  technology  have been provided to the public and private
 sectors of New Jersey through the dissemination of The United States
 Environmental Protection Agency  (EPA) documents,    and phone
 consultations provided by DEP technical staff.

     The  DEP instituted a voluntary mitigation certification
 program in 1986.  Successful program applicants are listed in a
 brochure which  is distributed to the public.   One requirement of
 this program is the successful completion of a DEP approved
 mitigation training course, which, also serves as a means of
 technology transfer.   A mandatory certification program will be
 instituted during 1990,  and will require all  individuals or firms
who test  for or reduce radon levels in the State to be certified by
 the DEP.

     The DEP has compiled data on homeowner funded radon reduction
efforts since January, 1986.   Evaluation of this data has enabled
the DEP to determine   (1) what mitigation methods are being selected

-------
by New Jersey residents, (2) who is performing the mitigation work,
(3) the effectiveness of mitigation methods installed by both
homeowners and voluntarily certified firms and, (4) the average cost
of each mitigation technique.

     The purpose of this paper is to report the results of data
collected on 473 mitigations performed in New Jersey from April,
1988 to November 1989 and to compare these results with previously
reported data collected from January, 1986 to March, 1988.
Evaluation of the effectiveness of mitigation techniques will help
the DEP to identify and target those areas where more training and
information transfer may be needed.

     Use of the term "certified firm" in this paper means a radon
mitigation firm which has successfully participated in the DEP
Voluntary Certification Program.  "Non-certified firm" refers to
those firms and individuals that either elected not be participate
in the DEP voluntary program or failed to meet minimum requirements.
"Effectiveness" means those post remediation radon levels in the
lowest liveable area which are less than four picocuries, per liter
(4 pCi/1).


                           Sources of Data

     The DEP's post remediation testing program is designed to track
homeowner-funded radon reduction efforts in the State.  The purposes
of this program are to provide homeowners, at no cost, with radon
testing to assess the effectiveness of mitigation measures installed
and to monitor the performance of the mitigation industry in the
State.  Testing in the program was initially conducted through field
visits by DEP staff to remediated homes.  Two carbon canisters were
deployed on the two lowest liveable levels of the home and a
post-remediation survey form was completed.

     Since March, 1988, post remediation testing has been provided
through a mail-out program.  Accordingly, homeowners are encouraged
to contact the DEP and arrange for post-remediation testing.  DEP
staff complete phone information forms which include the resident's
name, address, pre-mitigation test results, if available, mitigation
technique installed, and the person or firm who installed the
system.  The resident is then required to submit to the DEP, a copy
of his or her mitigation contract or a description of the system
installed.  Upon receiving this information, the homeowner is mailed
two charcoal canisters with specific instructions on how to deploy
and return the devices.  A mitigation installation is determined to
be effective when post remediation radon concentrations, under
closed house conditions, are less than 4 pCi/1 in the lowest
"liveable" floor.  The review of the data which follows has been
derived from 473 post remediation surveys conducted by DEP from
April, 1988 to November, 1989.  Comparisons are made with similar
data collected on 716 remediations from January, 1986 through March,
1988 (the 1988 Study).

-------
     A second source of data on mitigation activities was collected
through the DEP's voluntary mitigation certification program.  Firms
participating in the program are required to submit quarterly
survey reports on all radon remediations conducted in New Jersey.
In addition to the information obtained through the DEP
post remediation testing program, firms are required to include the
cost of each mitigation system installed.  Data has been compiled
from firm reports submitted to DEP for 3,197 homes mitigated from
April, 1988 to October, 1989.  Because evaluation of this data
relies on accuracy and the honesty of firm reporting, it is used
only as a comparison to DEP post remediation program data and in
determining mitigation costs.


                         Mitigation Methods

     Methods employed to reduce indoor radon levels in the 473 homes
monitored through the DEP post remediation testing program have been
grouped into four major categories which include (1) sealing only
techniques, (2) soil depressurization methods, (3)  air to air heat
exchangers and, (4) other methods.

     Sealing only techniques typically consist of covering or
filling in the sump pump pits and perimeter french drains, and
sealing of floor/wall cracks and utility openings.   The extent to
which sealing measures were used varied considerably, especially
with techniques and materials.  According to phone conversations,
homeowners/non certified firms were less likely to follow guidance
measures outlined in EPA documents.  Cracks were often not properly
prepared before applying sealants and in some cases ordinary cement
was used as the sealant rather than non-shrink urethane caulks.

     Soil depressurization methods include the use of active and
passive sub-slab ventilation, drain tile, sump pump, and block wall
suction.  Active sub-slab ventilation was the most widely employed
soil depressurization method and was generally accompanied by
sealing of major radon entry routes.  However, after examining
mitigation contracts and design descriptions, it became apparent
that some certified firms and homeowner/non-certified firms did not
always follow EPA installation guidelines.  The locations of exhaust
fans and exhaust ports was the most often noted non-conforming
practice.  Basement mounted fans often leak and improperly located
exhaust ports may allow radon gas to re-enter a home.  This could
explain the ineffectiveness of some systems.  Also, some homeowners
and non-certified firms did not properly seal radon entry routes,
thus reducing system effectiveness.

     Air to air exchangers include those systems with heat recovery
only.  The principle of this design is to increase the ventilation
rate in the house thus diluting the indoor radon concentration while
decreasing the natural exfiltration of indoor air and reducing the
pressure differential which causes soil gas to be sucked into the
home.  Only 14 such mitigation attempts were observed in this study.
This may be explained by the fact that the installation of air to

-------
air heat exchangers is often expensive and requires professional
services for proper sizing and installation.   Additionally, this
type of mitigation is less effective than soil depressurization in
reducing radon levels and is more costly to operate and maintain.

     The last mitigation type considered in this study, "other"
methods, includes forced air and crawl space ventilation.  This type
of mitigation includes passive venting or installing fans or blowers
in basement and crawl space walls and windows.  These measures
usually require a minimal amount of technical skill and money, which
may explain why they were most often installed by homeowners.  Only
18 such mitigation attempts were reported.

     Selection and installation of mitigation methods were
influenced by who was performing the work.  Homeowners/non-certified
firms with some knowledge of buildings construction often performed
their own mitigation work employing mainly sealing techniques.
Certified firms overwhelmingly selected depressurization techniques
as a means of remediation (166 out of 227 installations).  Table 1
shows the homes remediated for each of the 4 mitigation methods
discussed above.  As shown, the most widely employed technique
selected by homeowners/non-certified firms and certified firms was
soil depressurization (58% of homes).  Sealing only as a mitigation
technique was used in 35% of the homes, air to air exchangers were
installed in 3% of the homes and other techniques in 4% of the
homes.

     When compared to Table 2, which shows the 716 homes mitigated
in the DEP's 1988 Study, some obvious conclusions can be drawn.
Installation of soil depressurization systems remained relatively
unchanged with 58% of the homes in this study and 59% of the homes
in the 1988 study selecting this method.  Sealing only was employed
in 35% of the homes in this study and in 28% of the homes  in the
1988 study, and was again the dominate method employed by
homeowner/non-certified firms.  Certified firms installed mitigation
systems in 48% of the homes in this study and as in the 1988 study
selected soil depressurization techniques most frequently,
comprising 73% of all work which they performed.  An interesting
observation is in the difference in the percentage of homes that
were mitigated by homeowner/non-certified firms and DEP certified
firms.  In the 1988 study, 47% of the homes were mitigated by
homeowners/non-certified firms, and 53% by certified firms.  In this
study, 52% of the homes were mitigated by homeowner/non-certified
firms and 48% by certified firms.  Perhaps this is a result of
further dissemination of radon reduction information to the public
sector.

                 Effectiveness of Mitigation Systems

     The performance of the mitigation systems installed in the 473
homes in this study was evaluated using post-mitigation radon levels
measured by the DEP.  Two canisters were provided to the homeowner
with placement instructions in accordance with DEP/EPA testing
protocols.  Mitigated homes with radon concentrations less than 4
pCi/1 in the lowest liveable level were considered effective

-------
remediations.  It should be emphasized that post remediation testing
was conducted using short term charcoal canisters and therefore,
these results are not indicative of annual averages.

     The performance of all mitigation methods installed in this
study is compared to similar data on the 716 homes surveyed in 1988
and is shown in Figure 1.  The histograms show the distribution of
post remediation radon levels in homes mitigated by homeowner/
non-certified firms and certified firms.  The distribution of post
remediation radon levels shows that in this study, 56% of the homes
mitigated had radon levels below 4 pCi/1.  This was a significant
increase when compared with the 1988 study which shows only 36% of
the mitigated homes with post remediation levels less than 4 pCi/1.

     A closer look at the homes mitigated by certified firms only
reveals an even greater increase in the number of homes mitigated
below 4 pCi/1.  Figure 2 compares levels in 225 homes mitigated by
certified firms in this study with 382 homes mitigated by certified
firms in the 1988 study.  The distribution of radon levels shows
that 78% of the homes in this study had post remediation levels less
than 4 pCi/1 compared to only 48% in the 1988 study.  This shows an
increase in mitigation effectiveness of 30%.  This higher success
rate may be attributed to the industry's increased experience in
diagnosing, designing and installing proper mitigation systems.

     A comparison of each mitigation method installed by homeowner/
non-certified firms and certified firms is presented in Figure 3.
As can be seen, certified firms were significantly more effective  in
reducing indoor radon levels to below 4 pCi/1 for all methods.
Horaeowner/non-certified firms were only 27% effective using sealing
measures only while certified firms were 65% effective in reducing
radon levels.  Certified firms were even more successful in
installing effective soil depressurization systems, with 81% of the
systems reducing radon levels below 4 pCi/1.  This is a somewhat
greater performance rate when compared to all methods used by
certified firms (78%).  This is encouraging information since
certified firms are generally employed by homeowners with high radon
levels or when do-it-yourself attempts have failed.  A conclusion
that can be drawn from this information is that properly trained
certified firms have had a consistently higher success rate than
homeowner/non-certified firms for all mitigation methods in reducing
radon levels below 4 pCi/1.

     To further investigate the effectiveness of mitigation systems
to reduce lowest floor radon levels to below 4 pCi/1, a comparison
of the distribution of post remediation radon levels for each
mitigation technique is presented in Figures 4, 5, 6 and 7.  Sealing
methods only is illustrated in Figure 4.  A dramatic increase in
performance can be seen as the percentage of homes with post
remediation radon levels less than 4 pCi/1 more than doubled from
15% in the 1988 study to 37% in this study.  The continued high
failure rate of sealing methods only, 63% in this study, might be
attributed to the fact that this method is most often selected by
inexperienced homeowner/non-certified firms who may not know how to
properly seal or know when such techniques alone will be successful.

-------
     The effectiveness of air to air exchangers is shown in Figure
5.   The distribution of post remediation results indicate an
increase in performance in homes remediated from 23% in the 1988
study to 57% in this study.  One point should be emphasized in
explaining this significant increase.  In this study only 14 air to
air heat exchangers were examined while in the 1988 study 57
installations were examined.  Of the 57 installations, 23 were of
one particular design which often failed, thus accounting for 77% of
the homes having levels greater than 4 pCi/1.  This particular
system was not noted in any of the designs surveyed in 1989.

     The distribution of radon levels in homes in which soil
depressurization systems were installed is shown in Figure 6.
Although the performance of these systems has moderately increased
from 50% in the 1988 study to 67% in this study, it is still not
near the 90% effectiveness reported by firms which participated in
EPA's pilot survey of private sector firms.   One explanation for
the 33% failure rate may be in the high number of soil
depressurization attempts by inexperienced mitigators.  Forty
percent of all soil depressurization systems surveyed in this study
were installed by homeowner/non-certified persons (see Table 1).
Since these methods require diagnostic surveys and sealing measures
to maximize performance, they are more likely to be improperly
installed by inexperienced personnel.

     Figure 7 shows the distribution of post remediation radon
levels in homes utilizing "other" methods.  In the 1988 survey only
20% of the methods were effective.  In this study, 55% of the
"other" methods reported post remediation radon levels less than 4
pCi/1.  Although this may be impressive it should be emphasized that
these techniques included passive and forced ventilation which
usually cost the homeowner an additional heating expense.
Consequently, these methods are usually temporary measures and may
explain why they were only used in 18 of 473 remediations.

     The performance of mitigation systems installed by certified
firms was also evaluated from data compiled from quarterly
mitigation reports submitted to the DEP by these same firms.
Mitigation firms participating in the DEP voluntary certification
program are required to report all radon mitigation work performed
in New Jersey on a quarterly basis.  Pre and post remediation radon
levels, type cf mitigation design installed and the total cost of
mitigation system are some of the data reported.  A total of 3,197
residential remediations were reported to the DEP from April, 1988
to October, 1989.

     The percentage of homes reported as having post remediation
radon levels less than 4 pCi/1 was significantly greater than that
presented by the DEP.  Certified firms reported effective
remediations in 96% of all mitigation methods attempted compared to
only 78% reported in DEP's post remediation data.  This discrepancy
in the two data sets may be explained when considering the following
factors:  various radon testing techniques were used by firms when
performing post remediation tests; firm and DEP post remediation
testing were often performed during different periods of the year;

-------
the possibility that firms might inaccurately report results, and
that firms may have not reported all post remediation radon levels
greater than 4 pCi/1 to the DEP.


                     Average Cost of Mitigation

     The average cost of radon remediation charged to homeowners by
certified firms was also computed from the quarterly survey reports
submitted to the DEP.  Table 3 compares the average costs of each
mitigation method in this study to the 1988 study.  The average cost
of remediation for all mitigation methods was $1,100.  This is
appreciatively less than the $1,300 reported in the 1988 study where
942 remediations were evaluated.  The average cost of sealing only
methods in this study was $601 compared to $700 in the 1988 study.
Soil depressurization methods decreased by more than 10% to $1,120
in this study from $1,270 in the 1988 study.  Air to air heat
exchangers decreased by about 20%.  Their average cost in this study
was $1,620 compared to $2,000 per installation in the 1988 study.
Finally, remediation by "other" methods also cost less in this
study.  The average cost was only $695 compared to $1,900 in the
1988 study.


                             Conclusions

     The proportion of homes effectively remediated for elevated
indoor radon levels is increasing.  Of the homes surveyed in this
study compared to the 1988 study, radon remediation effectiveness by
certified firms has increased from 48% to 78%.  The proportion of
soil depressurization systems installed is about the same, however,
their effectiveness in reducing radon has increased significantly,
and this method has remained the most effective method in reducing
indoor radon levels below 4 pCi/1.  Sealing only techniques were
chosen most often by homeowners who performed their own mitigation
work.  Air to air heat exchangers were the least employed type of
mitigation performed in this study.  This may be because of the high
costs of installation and maintenance along with their actual
ineffectiveness in lowering indoor radon levels.

     All mitigation methods decreased by 15% in cost from the
previous year.  The average cost of mitigation charged to New Jersey
residents by certified firms in this study was $1,100.

     Although this survey is encouraging for both the radon industry
and New Jersey residents it also demonstrates the need for a program
of continuing education and training.  State and federal training
programs will enable the radon industry to stay abreast of the
latest scientific and technological research as well as changes in
state and federal radon legislation.  Support and participation in
such programs will enable the radon industry to become more
knowledgeable and effective in their remediation efforts and thus
offer the residents of New Jersey the quality workmanship they
deserve.

-------
     This work described in this paper was not funded by the United
States Environmental Protection Agency, therefore, the contents do
not necessarily reflect the views of the agency and no endorsement
should be inferred.


                            Bibliography

1.   USEPA, Radon Reduction Techniques for Detached Houses,
     EPA/625/5086/019, 1986

2.   USEPA, Application of Radon Reduction Methods,
     EPA/625/5-88/024, August, 1988

3.   DePierro, N. and Cahill, M. "Radon Reduction Efforts in New
     Jersey", NJDEP, October, 1988

4.   Cohen, S.A., "Results of a Pilot Survey of Radon Prevention and
     Mitigation Firms," USEPA, Contract No. 68-02-4375, November,
     1987
ndl!3089

-------
Table 1   Homeowner/Non-Certified Firm and DEP Certified Firm
          Mitigation Installations in 473 Homes Surveyed in 1989
                 Sealing         Soil          Air to Air
                  Only     Depressurization    Exchanger    Other    Total
Homeowner/
Non-Certified
Firms

DEP Certified
Firms

Total
 119
  46
165[35]
 110
 166
276^58]
                                           12
14 [3]
18[4]
                    246[52]
                    227[48]
473(100]
             Bracketed numbers represent percent of total homes.
Table 2   Homeowner/Non-Certified  Firm and  DEP Certified  Firm Mitigation
          Installations  in 716 Homes  Surveys  in  1988.
                 Sealing          Soil          Air  to Air
                  Only      Depressurization    Exchanger     Other    Total
 Homeowner/
 Non-Certified
 Firms

 DEP  Certified
 Firms

 Total
 176
  27
203[28]
 95
 326
421[59]
                                43
                             20
 14         15


57[8]     35[5]
          334[47]


          382[53]


          716[100]
              Bracketed numbers represent percent of total homes.
 ndl!3089

-------
Table 3^   Average Cost of Radon Mitigation  ($)
Mitigation
Technicrue
Sealing

Soil Depressurization

Air to Air Exchangers

Other

Overall Average
1989 Study
(3197 Homes)
   601

   1120

   1620

   695

   1100
1988 Study
(942 Homes)
   730

   1270

   2000

   1900

   1300
ndl!3089

-------
FIGURE  1

Post-Remediation  Radon  Levels  in  Homes

Mitigated by  Homeowners/Non-Certified  Firms

and  Certified Firms  for All Methods
   9OT
   SO--
   7O--
 co 6O
 CD
 e
 o
 1C 50
 
-------
FIGURE  2
Post-Remediation  Levels In  Homes
Mitigated  by  DEP  Certified  Firms
for  All  Methods
   9O-r
                                                      LEGEND
                                                          1988 Study

                                                          1989 Study
            <4     4-8    8-20   20-5O   50-20O
              Post Mitigation  Radon  Levels
                          pCi/l

-------
FIGURE  3
Effectiveness of  Mitigation  Measures
     9OT
                                                  LEGEND
                                                       Owner/Non-Certified
            Sealing  Soil Depr.
Other

-------
FIGURE  4
Post-Remediation  Radon  Levels  in  Homes
Mitigated Using  Sealing  Only
(DEP Certified and Homeowner/Non-Certified)
   9OT
   80--
   70--
 to 60--
 o>
 E
 o
 m 50+
   40--
 o
 a>
 O_
   3O--
   20--
   10--
                                                        LEGEND
1988 Study

1989 Study
            <4      4-8     8-2O    20-5O   5O-2OO
               Post Mitigation  Radon Levels
                           pCi/l

-------
FIGURE  5
Post-Remediation Radon  Levels  in Homes
Mitigated  Using  Air to  Air Heat  Exchangers
(DEP Certified and  Homeowner/Non-Certified)
                                                    LEGEND
                                                        1988 Study

                                                        1989 Study
           <4      4-8     8-2O    20-50   50-2OO
              Post Mitigation  Radon  Levels
                         pCi/l

-------
FIGURE  6
Post-Remediation Radon  Levels  in  Homes
Mitigated  Using  Soil Depressurization
(DEP Certified and  Homeowner/Non-Certified)
   90 T
                                                    LEGEND
                                                         1988 Study

                                                         1989 Study
           <4      4-8    8-2O    20-50   50-2OO
             Post Mitigation  Radon  Levels
                         pCi/l

-------
FIGURE  7
Post-Remediation  Radon  Levels in  Homes
Mitigated  Using  Other  Methods
(DEP Certified and  Homeowner/Non-Certified)
   90 T
                                                     LEGEND
                                                         1988 Study

                                                         1989 Study
                                         en
            <4     4-8     8-20    20-50   5O-2OO
              Post Mitigation Radon Levels
                         pCi/l

-------
                                                                        VII-8
        LONG-TERM PERFORMANCE AND  DURABILITY OF ACTIVE RADON MITIGATION
                    SYSTEMS IN EASTERN PENNSYLVANIA HOUSES

                       by:     A.  G.  Scott
                               A.  Robertson
                               AMERICAN ATCON,  INC.
                               Wilmington,  DE   19899
                                   ABSTRACT

    Indoor radon reduction techniques were installed in 40 houses in the Reading
Prong region of eastern Pennsylvania over the period 1985 to 1987.  Active soil
ventilation systems  were installed in 36 of  the houses; 3 had  heat recovery
ventilators (air-to-air  heat  exchangers), and  2  installations included carbon
adsorption units to remove radon from well water.   Follow-up measurements of the
post-mitigation radon concentrations in these houses were conducted with alpha-
track detectors during the winters of 1986/87, 1987/88, and 1988/89.  The systems
were inspected and  the owners interviewed during each installation and collection
visit.

    In those houses where the radon reduction system was in continuous operation
during the monitoring periods, the measured  average radon concentration in each
house compared well from year to year.  High radon concentrations were measured
if the system did not operate during a monitoring period.  Overall, no general
degradation in system performance was observed.

     One manufacturer supplied 34  fans with the same motor and  impeller assembly
which were used in  the active soil ventilation systems. There have been five fan
failures in this population to date.  The causes were  capacitor failure in four
fans, and noisy bearings  leading to shutdown  in one fan. Assuming an exponential
lifetime distribution, the mean life of these fans  (between repairs/replacements)
is calculated  as 15  years,  based upon  this experience.   There have  been no
failures among three other fans from another manufacturer.

    Of the three installed heat recovery  ventilators,  one fan  bearing failed in
early 1989.  This experience suggests a mean life between repairs of 8 years.

    The performance of one charcoal adsorption unit has deteriorated from <95%
to 65% removal over 3 years.  The second unit, which contains charcoal specially
selected for radon  removal, has maintained a  97% removal efficiency over the same
period.

    Five houses from this group of 39 houses (after 1 early dropout) have been
sold over the period of the study, suggesting a mean occupancy life of 29 years
in this study group.

    This paper  has been reviewed  in  accordance  with  the U.  S.  Environmental
Protection  Agency's peer  and administrative  review  policies,  and has  been
approved for presentation and publication.

-------
                                 INTRODUCTION

    Indoor radon reduction techniques were installed and tested in a total of 40
homes in communities on the Reading Prong in eastern Pennsylvania over the period
of  June 1985  to  June 1987  as  part of  EPA  Contract 68-02-4203.  The primary
mitigation  method in 36 of  these  houses was active soil  ventilation,  with 3
houses  receiving  heat-recovery  ventilators  (air-to-air  heat exchangers) and 1
house receiving just a radon-in-water removal unit.  This project was reported
in detail in EPA-600/8-88-002 (Reference 1).

    Follow-up alpha-track-detector (ATD) radon concentration measurements have
been made annually in 38 of the 40 demonstration homes during the heating seasons
of 1987/88  (Reference 2) and 1988/89 (Reference 3).  Two of the houses were no
longer  available  for monitoring - one house had been removed from the original
site, and  the  owner of  a  second  house has  discontinued participation  in the
project.

    This paper examines  the  results in the  38 demonstration homes  to make an
assessment of long-term system performance and equipment durability.
                             MEASUREMENT PROCEDURE

    Average radon concentrations are measured by Terradex  'Track-Etch1 detectors
exposed for 4 months during the heating season.  The forces that urge radon from
the soil into the home are highest during this time, and therefore this season
presents the greatest challenge to radon mitigation systems.

    Quality assurance and control (QA/QC) measures were instituted to ensure that
the acquired data were of high quality.   Unexposed detectors  (blanks)  were
included in  each  set sent  to Terradex for  analysis,  and  a number of detectors
from  each  batch  were  exposed to  known concentrations  in the  radon  chamber
operated by the U.S. Department of Energy's Environmental Measurements Laboratory
(EML) in New York.  All  these detectors  were given fictitious house identifying
numbers,  so  that  they  were  "blind"  samples to  the  processor.   A  "zero
correction",  equivalent to  the mean reported  exposure of  the  blanks,  was
subtracted  from  each ATD  result  reported by  Terradex.   A  "gain correction"
derived from the spiked samples (EML Exposure/Zero-Corrected Reported Terradex
Exposure) was used as a multiplier to adjust the reported house concentrations
to conformity with the EML.

    As an additional QA/QC measure, detectors  were  exposed in groups of three
(1986/87 and  1987/88) or  two (1988/89) at each measurement  location,  and the
results averaged to reduce the measurement uncertainties.

    The demonstration houses were visited during December of each year to install
Terradex "Type SF" Track-Etch detectors  in accordance with the  guidance given
in the EPA Measurements Protocol (Reference 4).  Detectors were placed in both
the main living area and the basement, hung together  in groups of three (two in
1988/89) from an interior wall or ceiling in the living area and a central joist
in the basement.  Each detector was marked with the installation date  and the

-------
house identification  code.   That  information,  plus the detector  numbers and
their locations, were  recorded  in a Track-Etch Record Book kept specifically for
that purpose.  During these  visits  the  soil ventilation fans and the air-to-air
heat exchangers were checked to see that they were operating.  The houses were
visited about 16 weeks  later,  at the end of  the heating season,  to remove the
detectors.   An  ATD retrieval rate  of 100% was  achieved.   The  homeowners were
interviewed  to  discover any  events  that  might  have  affected  the  system
performance  during  the  monitoring  period.  Retrieved detectors were placed in
pairs into the manufacturer's envelopes, which were then sealed by folding and
taping in a  low-radon atmosphere.   The envelopes were returned to Terradex in
one batch.
                                 QA/QC RESULTS

    The 1986/87 and 1987/88 blanks did not indicate a need for zero correction.
The results from the QA/QC blanks  in 1988/89 led to a zero correction of 10 Bq/m3
(0.26
    The detectors exposed at EML had known exposures which were always greater
than the  reported exposures,  and  so gain corrections were applied of 1.16 for
1986/87;  1.30 for 1987/88; and 1.21 for 1988/89.

  •  One outlier  was identified in  a  group  of three detectors  in the 1986/87
reported  results and was  discarded.   No outlier was discovered in the 1987/88
results.   In 1988/89  there were two probable  outliers;  these were of limited
significance, since the measurements were both from houses where the mitigation
system did not operate during the entire measurement period.
                                    RESULTS

RADON REDUCTION PERFORMANCE

    Table I  shows  the results by year for  the  27  houses where the mitigation
systems operated  for the entire 3-year period and were monitored each winter,
thus  providing a  fair  measure of  long-term mitigation  system performance.
Excluded  from the  table are  the  results  from  five houses where  the system
equipment failed during the 3 years,  three houses in which the system was turned
off during  the monitoring period,  three houses with measurements only in 1987
and 1988, and one house where the system was  incomplete because the owner refused
further work.

    The average radon concentration in the basement of these 27 houses  was 4.7,
4.7, and  4.9 pCi/*  in 1987,  1988,  and 1989, respectively;  in the living area,
the average  was  4.0, 3.9, and  3.9 pCi/l.   The  small differences in the group
average from year to year are without  statistical  significance.  The constancy
of the group average with time shows that there has been no general deterioration
of system effectiveness  in 3 years of  operation.

-------
Table I.
ALPHA-TRACK MONITORING RESULTS
POST-MITIGATION RADON RESULTS foCi/i)
ID
NO.
2
3
5
6
8
9
10
12
13
16
18
20
21
22
24
25
26
27
28
29
30
32
33
34
35
36
37
PREMIT
1989
(oCi/jn B*
413
350
110
60
183
533
626
11
64
395
12
210
172
24
66
122
89
21
21
61
17
6
82
470
144
300
87
Average - 172
5
3
5
3
2
12
10
1
2
4
12
8
1
10
4
7
0
5
3
2
3
0
11
5
2
0
0
4
%RDEV* - 104
.5
.0
.0
.2
.9
.4
.4
.6
.7
.8
.7
.3
.9
.8
.3
.2
.6
.6
.6
.1
.9
.5
.2
.1
.4
.8
.9
.9
76
LA*
8.7
1.9
4.4
2.7
1.0
17.1
8.9
2.1
2.8
1.1
5.1
9.3
2.7
4.0
3.7
5.3
1.1
2.1
5.1
2.3
2.1
3.2
0.7
5.5
1.0
0.7
0.5
3.9
94
B
4.8
3.5
5.0
4.1
3.5
10.4
15.2
2.2
2.6
5.7
13.5
6.5
2.0
8.6
3.6
7.7
1.1
4.0
4.1
1.6
4.0
1.2
3.5
5.4
1.0
1.1
1.2
4.7
78
B=Basement ; LA-Living Area ;

Average ;
%RDEV-Relative
Conversion factor:
Bq/m3 -
1988
LA
6.7
2.3
4.4
3.2
1.5
12.9
9.9
2.2
3.9
2.5
3.4
10.0
2.7
4.4
3.8
6.0
1.6
2.2
4.4
2.0
1.6
4.4
1.2
5.5
0.9
1.0
0.7
3.9
78
1987
B
3.0
4.1
5.0
3.8
4.5
13.5
10.4
4.3
2.7
6.3
10.2
6.7
3.6
8.8
5.0
6.3
2.4
4.4
2.8
2.2
3.5
1.2
2.6
6.4
0.9
1.9
0.7
4.7
66
BAVE-Basement
Standard Deviation
37 x value in
pd/Ji
LA
6.0
2.4
5.0
5.7
2.1
16.8
7.5
2.9
2.3
2.0
2.4
11.5
3.0
3.1
5.3
3.5
1.7
2.6
6.1
1.6
1.5
3.7
1.3
4.3
0.8
0.8
2.0
4.0
87
Average;
(in %).
•
SAVE*
4.4
3.5
5.0
3.7
3.6
12.1
12.0
2.7
2.7
5.6
12.1
7.2
2.5
9.4
4.3
7.1
1.4
4.7
3.5
2.0
3.8
1.0
5.8
5.6
1.4
1.3
0.9
4.8
70
LAVE'
7
2
4
3
1
15
8
2
3
1
3
10
2
3
4
4
1
2
5
2
1
3
1
5
0
0
1
3

.1
.2
.6
.9
.5
.6
.8
.4
.0
.9
.6
.3
.8
.8
.3
.9
.5
.3
.2
.0
.7
.8
.1
.1
.9
.8
.1
.9
84
LAVE-Living Area






    Table I shows the average radon concentrations in the basement and the  living
 area of  the  individual houses.   In  general,  the  average concentration does not
 vary greatly from year to year  in a given house,  but House 33  is an exception.
 The  radon concentration in 1989 in the  basement increased by over  300% while
 concentrations  in the living area remained  similar to previous years.

    A number of  houses (2, 9, 20, 28, and 32) have higher radon concentrations
 upstairs than in the basement.   This is more clearly seen in Table  II, where the
 ratios  of radon concentration  in the living area to  that in the basement are

-------
Table II.   RATIO  OF RADON  CONCENTRATIONS  IN LIVING AREAS TO BASEMENTS
ID
No.
2
3
5
6
8
9
10
12
13
16
18
20
21
22
24
25
26
27
28
29
30
32
33
34
35
36
37
Average-
%RDEV*=
* %RDEV = Relative
LIVING AREA/BASEMENT RADON CONCENTRATION
1987
2.00
0.59
1.00
1.50
0.47
1.24
0.72
0.67
0.85
0.32
0.24
1.72
0.83
0.35
1.06
0.56
0.71
0.59
2.18
0.73
0.43
3.08
0.50
0.67
0.89
0.42
2.86
1.01
75
standard
1988
1.40
0.66
0.88
0.78
0.43
1.24
0.65
1.00
1.50
0.44
0.25
1.54
1.35
0.51
1.06
0.78
1.45
0.55
1.07
1.25
0.40
3.67
0.34
1.02
0.90
0.91
0.58
0.99
67
deviation (in %).
1989
1.58
0.63
0.88
0.84
0.34
1.38
0.86
1.31
1.04
0.23
0.40
1.12
1.42
0.37
0.86
0.74
1.83
0.38
1.42
1.10
0.54
6.40
0.06
1.08
0.42
0.88
0.56
1.06
109

RATIOS
Averaee
1.66
0.63
0.92
1.04
0.41
1.29
0.74
1.00
1.13
0.33
0.30
1.46
1.20
0.41
0.99
0.69
1.33
0.51
1.56
1.02
0.46
4.38
0.30
0.92
0.74
0.74
1.33
1.02
77

 shown for individual houses, over a 3 year period.   The  ratios are consistently
 above unity in those five houses.  Houses 9,  20,  and 32  are known to have radon
 levels   in  well  water  above  20,000 pCi/X,  which  could  be  contributing
 preferentially,  through water-usage pattern,  to higher concentrations upstairs.
 The concentration of radon in water at House 28, which has a well, is not known.
 In  addition,  Houses 2, 9, and  28  have  block fireplace  structures  which could
 provide  a direct soil-gas entry route upstairs without entering the basement.
 Houses  2 and 9  also have block wall pressurization systems  which could force
 radon-laden air  upstairs.  House 20 has  an adjoining paved crawl  space which may
 not be fully treated.   Although these are perhaps special cases, there are enough
 of  them  to  indicate   that  basement  measurements  do  not  always provide  a
 conservative  estimate  of  the  radon exposure to the  occupants.

-------
INOPERATIVE SYSTEMS

    Some mitigation systems were  inoperative during the monitoring periods, and
are listed in Table III, where the post-mitigation radon results obtained during
all monitoring periods  are  provided along with  brief comments as to why the
systems were not functioning.   House  1, which was demolished and moved from its
site before post-mitigation monitoring could begin, and House 11, where the owner
refused further participation, have been included for completeness.

Table III.  SYSTEMS  INOPERATIVE DURING ONE OF THE ATD MONITORING PERIODS
POST -MITIGATION RADON RESULTS foCi/ii
ID
NO.
i
2
4
7
11
14
15
17
31
39
PREMIT.
(oCi/i)
U6
413
25
402
49
36
18
9
485
111
1989 1988 1987 ADDITIONAL
B** LA** B LA B
House moved from site
5.5 8.7 4.8* 6.7* 3.0
1.2 1.0 7.3* 3.1* 0.8
103.1* 24.1* 4.9 3.8 4.8
Owner no longer participating
10.8* 8.0* 1.1 1.4 0.6
1.3 1.3 19.7* 11.0* 1.3
8.5* 4.9* 8.2 6.4 8.8
252.7* 93.3* 2.8 8.3 2.1
7.5 1.8 46.1* 17.5*
* Asterisk marks monitoring period system was
B - Basement; LA - Living Area. *" -- -
LA COMMENTS

6.0 1. Capacitor failure;
repaired prior to
monitoring.
2 . Fan turned off by
homeowner.
0.9 Capacitor failure.
3.2 Capacitor failure.

0.8 Fan turned off.
1.2 Bearing failure.
4.8 HRV bearing failure .
6.6 Capacitor failure.
..*** Fan turned off.
inoperative .
No measurement during 1987.
    In three houses, the soil ventilation fan was turned off by the homeowner,
either  deliberately or inadvertently.   Reportedly,  the fans  at  House 2 were
turned off for 12 days during the 1988 measurement period while the family was
away  on vacation.    However  no  increase in  radon  concentration was observed,
indicating that the concentrations did not return promptly to the very high pre-
mitigation levels.   The results were  similar to 1989 when the  owners took care
to have  the fans running constantly throughout the monitoring  period.  Perhaps

-------
only one of the two fans involved in this system was turned off.   Likewise, at
House 39, the system fan was turned off  for an extended period while the owners
were away.  At House 14, the owner unintentionally disconnected the fan from the
electrical outlet for some time during the 1989 monitoring period.

    Besides the above three examples of  soil ventilation systems being rendered
inoperative due to owner  intervention,  there have been five instances of soil
ventilation systems not working due to mechanical  failure  of  the  system fans
- Houses 2, 4,  7,  15,  and 31.   The first failure occurred at House 2, and was
corrected before the start of the 1987 monitoring season,  the second two during
the 1988 monitoring period (Houses 4 and 15), and  two more during the 1989 period
(Houses  7  and  31).   In  addition to these five mechanical failures  in soil
ventilation systems, there was one  failure of a HRV fan (House 17).

    When the fans  did not  operate during monitoring periods in these houses, the
average radon concentrations were high,  as would be expected.  Levels at Houses
2, 4 and 15  returned to previous post-mitigation values after system repairs.
The systems have been repaired at Houses 7 and 31, and await  further monitoring.

FAILURE EXPERIENCE WITH SOIL VENTILATION FANS

    Three different types of soil ventilation fans, from two different manufac-
turers,  were used  in  the  active soil ventilation systems installed  in the
demonstration houses  in Pennsylvania.   One type,  from the  first manufacturer,
was installed in three  of the houses; there have been no failures to date  among
these  three fans.

    Two  types of fans,  from  the second manufacturer, use  the  same motor and
impeller  assembly  in  different housings.   A  total  of  34  fans  from  this
manufacturer  have been installed  in  31 houses  as part of the Pennsylvania
project.   These 34 fans have suffered 5 failures  (Houses 2, 4, 7, 15, and 31)
since  installation 2 to 4 years ago.   The  number of fans  is greater than the
number of houses because three  houses each had two fans in their soil ventilation
systems.

    Four of the fan  failures were  due to  failure of the electrolytic  capacitor
used in the split-phase electric motor.  When the capacitor fails, the fan cannot
be  started.   If  the  fan is operating when the  capacitor fails,  the fan will
continue  to operate for  some  time, but at significantly reduced performance.

    One  fan,  at House  15, developed noisy  bearings and was turned off by the
owner  while  awaiting replacement.

    In all  instances  the fans  were repaired or  replaced under warranty  and are
functioning properly once more.

    If we  assume  an  exponential probability distribution  for fan  lifetimes,  we
can write  the probability for a failure after  time, t,  as

                             P(t)dt - A exp(-At)dt

-------
where A is the average failure  rate.  The mean lifetime between events requiring
fan repair or  replacement  is given by T-l/A.  If we take experience to date as
extending from the time of installation until the end  of the  latest monitoring
period (April 1989),  then the accumulated experience  is  897 fan-months or 75 fan-
years.  This sets the average failure rate, A, at 5 fan failures in 75 fan-years,
and therefore indicates a mean  lifetime, T, between repairs/replacements of 75/5
= 15 years.

    This lifetime can be  used to estimate the number of fan  failures  to be
expected year  by year, by integrating  the probability above  over  time.   The
expectation value for the  number  of  fans, N,  failed by the end of time,  t, is
given by

                           = N(0)[l-exp(-t/T)],

where T is 15 years and N(0) stands for the number of fans  installed, 34 in this
instance.  The expected number of failures can then be calculated as 2.2 after
1 year, 4.2 after 2 years,  6.2  after  3 years, and  8.0 after 4  years.  These are
in keeping with our  findings to date.

HRV EXPERIENCE

    The system breakdown  at  House  17  is the  first  failure  of an HRV  unit
installed at  three of the demonstration  houses.  The  owner  reported that the
bearings seized on the motor which drives  both the intake and  exhaust fans, and
the rotary heat exchange wheel,  in this HRV design.   The homeowner suspects that
the  fan shaft may  have   been bent by  the  contractor  during  the  original
installation,  potentially  causing  the eventual  deterioration of the bearings.
The fan has since been replaced by the vendor,  and the unit is now operating.

    To date, then, our experience with HRV units indicates 1  failure requiring
repair  in  100 unit-months of  operation.  Although   this failure  apparently
resulted from installation problems rather than  from normal long-term wear, the
installation problem is felt  to  be  representative  of the factors  which  can
influence  system  lifetime;  thus,   this  failure  can be  considered  in  the
calculation of equipment lifetime between repairs.  If one assumes  an exponential
distribution  of   lifetimes  to   failure,  the mean  lifetime,   between  required
repairs, can be estimated as 100/1-100 months or 8 years.   The expected number
of failures after 3 years of operation would be 0.94 and after  9 years that would
have grown to 2.0 failures.

WELL WATER TREATMENT  BY CHARCOAL ADSORPTION

    Adsorption on activated charcoal  was chosen  for  removal of radon from well
water at House 2, where it supplemented an active soil-ventilation mitigation
system, and at  House  30, where it was  the sole means of reducing household radon.
Two different suppliers for charcoal  units were used.   The first unit came from
a local water  purification supplier  and  contained charcoal of unknown  origin,
probably selected for organics  removal.   This was installed  in House 2,  where
radon concentrations  in the water were only moderately elevated at >200,000 Bq/m3
(>50,000 pCi/i).   The initial radon removal efficiency was about 95%.

-------
    This system was checked in January 1988, and removal efficiency was  79%.  By
December  1988,  the removal efficiency was 65%.   This  steady deterioration in
performance  is not  altogether  surprising,  since the  charcoal used  was  not
specially selected for radon  removal from water.

    The  second unit  came from  a firm  in Maine  which specialized  in water
treatment  systems for radon  removal.    This unit contained  coconut charcoal
specifically  selected to remove radon.   This was  installed  in House 30 where
radon concentrations  in  water were high  -  about  800,000 Bq/m3  (200,000 pCi/i).
The initial removal efficiency ranged from 95% to 99%.

    This second system was checked in 1987, and the removal efficiency was 99%
at  that time.   In December  1988,  the  removal efficiency was 97%.   This is
consistent  with a constant  removal  efficiency.    There is no degradation in
removal efficiency evident with this unit despite the high input concentration
and the high water usage associated with a family with  two  young children.
Perhaps this  is attributable  to the specially selected charcoal.

TURNOVER OF HOUSING STOCK

    During  the  period of this  study, starting from the  initial  selection of
houses  in April  1985  through the latest  monitoring  period  in April  1989,  5
homeowners out  of total  of  39 have moved.   (The  40th homeowner dropped out of
the program within the first year, and is  thus not included in this calculation.)
Since the total occupancy to date  is 146 owner-years,  the average rate of moving
during  the  study  is  5/146 per year.  Assuming an exponential distribution of
occupancy times,  the mean occupancy life, T, is 146/5 - 29 years.
                                  CONCLUSIONS

    For the 27 houses in which the mitigation systems operated continually during
all three monitoring periods from 1986 to  1989, the post-mitigation ATD results
did  not change  greatly from  year  to year.   Only  at House  33 was  there  a
significant  increase  in  radon concentration in  1989, and  then only  in the
basement.   There does  not appear to be any  general  degradation in mitigation
system performance with time alone.

    There are 34  fans from one manufacturer used in the soil ventilation systems,
and 5 failures  (requiring  repair or fan replacement) have occurred over the past
2 to 4 years.  Four of  these failures have  been due to capacitor failure, and
one due to bearing noise.   Based upon  this experience,  these fans are estimated
to have a mean lifetime of 15 years before repair/replacement.

    There are three HRV units  installed,  and one  of  these  units had a bearing
seizure requiring repair after 3 years.  The  estimated mean HRV lifetime between
such repairs, based upon this experience,  is 8 years.

    There are two charcoal adsorption units for removal of radon from well water.
The removal  efficiency  of one has fallen  from 95% to  65% over  a period of 28

-------
months from installation.   The removal efficiency of the other unit, containing
specially chosen charcoal, has remained constant at 97% over the same period.

    There are  39 homeowners in the study  group, and 5 of them have moved over the
period.  A mean occupancy period of 29 years is estimated.
                                  REFERENCES

1.  Scott A.G., Robertson A.,  and Findlay W.O., "Installation  and  Testing of
    Indoor Radon Reduction Techniques in 40 Eastern Pennsylvania Houses," report
    prepared  for  U.S.  Environmental  Protection Agency by  American  ATCON,
    EPA-600/8-88-002 (NTIS  PB88-156617),  Research  Triangle Park,  NC,  January
    1988.

2.  Scott A.G.  and Robertson A. , "Follow-up Alpha-Track Monitoring in 40 Eastern
    Pennsylvania Houses with Indoor Radon Reduction Systems (Winter 1987-88),"
    report prepared for U.S. Environmental Protection Agency by American ATCON,
    .EPA-600/8-88-098 (NTIS PB89-110035), Research Triangle Park, NC, September
    1988.

3.  Scott A.G.  and Robertson A. , "Follow-up Alpha-Track Monitoring in 40 Eastern
    Pennsylvania Houses with Indoor Radon Reduction Systems (Winter 1988-89),"
    report prepared for U. S. Environmental Protection  Agency by American ATCON,
    EPA-600/8-89-083 (in press), Research Triangle  Park, NC,  October 1989.

4.  U.S.  Environmental Protection Agency,  "Interim Indoor Radon and Radon Decay
    Product  Measurement   Protocols,"  EPA-520/1-86-04   (NTIS   PB86-215258),
    Washington, D.C., February 1986.

-------
          Session D-VII:
Radon Reduction Methods—POSTERS

-------
                                                       D-VII-1
ENERGY  PENALTIES   ASSOCIATED  WITH   THE  USE  OF   A  SUB-SLAB
DEPRESSURIZATION SYSTEM
                    by : Mike Clarkin
                         Terry Brennan
                         Camroden Associates, Inc.
                         Oriskany, NY 13424

                         Michael c. Osborne
                         USEPA/AEERL
                         Research Triangle Park, NC 27711
                             ABSTRACT

     One of the primary radon mitigation techniques used to reduce
indoor   radon   concentrations   in   houses   is   a   sub-slab
depressurization system. In this type of system, a  fan removes soil
gases containing radon from beneath the floor slab  and exhausts the
gases to the outdoors  by creating a pressure field  beneath the slab
that is negative relative to  the basement air pressure. Because of
this negative pressure, indoor conditioned air can  be drawn through
the  floor  penetrations  and  exhausted  outdoors.  In  order  to
determine  the amount  of  conditioned air  that is being  lost,  a
series of  experiments utilizing  tracer  gases were performed in
three houses. This paper presents the results of these experiments.


                           INTRODUCTION

     One of the most accurate ways of determining  a building's air
exchange rate is through the  use of a tracer gas. A gas is injected
into the  building's  interior. Specialized eguipment  is  used to
monitor the concentration of the tracer gas. The rate of decay of
the tracer  gas  can  then be used  to  determine  the building's air
exchange rate. This  procedure is commonly utilized  to determine the
building's  overall   air   exchange   rate.   For the  purpose  of
determining the amount of conditioned air being lost due to the
     This paper  has been reviewed  in  accordance with the  U.  S.
     Environmental Protection Agency peer and administrative review
     policies and approved for presentation and publication.

-------
sub-slab  depressurization  system,  the methods normally used were
modified  because the method described above would  include  the  air
lost  from the basement to  the superstructure of the house  through
thermal bypasses and  other penetrations, and the air lost to  the
outdoors  through  above  grade  basement  penetrations.   What   was
required  was  a  method that would disclose the amount of air that
the sub-slab  depressurization  system was actually drawing out of
the  basement through  the  below grade  penetrations  due  to  the
negative   pressure  field   being   developed  by   the  sub-slab
depressurization system.
GENERAL PROCEDURES

     A  series of tests  utilizing  sulfur hexafluoride  (SF6) as  a
tracer  gas and Miran  101  specific gas  analyzers  to monitor the
tracer  gas  concentrations  were performed  in  each  house.  One
.analyzer  would monitor  the basement  SF(  concentration,  and one
^analyzer  would monitor the SF6 concentration within the  sub-slab
depressurization  system exhaust pipe.   Figure  1 is a  simplified
schematic of  the  equipment setup. An  amount of  tracer gas was
injected  into the basement. The gas  and basement air were mixed
using several small fans.  The tracer gas was continuously  seeded
until the basement and the exhaust pipe concentrations were stable.
System airflow rates were  measured, and the ratio of basement to
exhaust pipe  SF6 concentrations at  that airflow  rate  was then used
to  determine  the  amount of indoor conditioned air that was being
lost due  to the sub-slab depressurization  system.

-------
                     Some of the air being exhausted by the sub-slab
                     depressurization system is being drawn through
                     penetrations by the negative pressure field being
                     developed. The rest of the exhausted air is coming
                     from the soil or bedrock, the amount of air drawn from
                     each source depends on how easily air can be drawn
                     through the foundation penetrations and through the
                     soil.
One  analyzer  monitors  basement  SF6  concentrations and  the other
monitors the exhaust pipe concentration.  A constant concentration
is maintained in the basement.  The ratio of the two concentrations
reveals  the amount of basement air being  exhausted by the sub-slab
depressurization system.
Figure  l.  Equipment  layout of tracer gas experiments.

-------
      Figure 2 illustrates the data gathered during a typical test.
 For this experiment, the  sub-slab system was in the  active mode.
 Data were collected for 5 minutes prior to injecting the SF, as a
 check on the stability of the analyzers. At the end  of  the first
 5 minute period,  SF4 was seeded and the injection rate was adjusted
 until a stable concentration  in the basement and exhaust pipe was
 obtained.
      50
  —  40-,
      30-
      20 H
      10 -I
     Basement concentration
     Pipe concentration
                    started tracer gas seed
10
                         —I-
                         20
   30       40
ELAPSED TIME (min.)
                                   50
                             60
                                                    70
 Figure  2.  Tracer  gas  experiment  performed  in  House  VA2  to
determine the energy penalty associated with  the use of  a sub-slab
depressurization  system.

     Figure  3  illustrates  a  test  performed  at  house  VA2  to
determine  the  effect that  penetrations  through  the slab have  on
the sub-slab system. For  this  test the sub-slab system was in the
active  mode.  SF,  was  continuously  injected  into  the basement
throughout the entire monitoring period.  For the first 45 minutes
of the test, the  floor  slab was left  intact.   At time 45, a  1/2-
in.* diameter  (0.2 sq in.)  hole was drilled through the
* Readers more  familiar with metric units  may use the conversion
factors at the end of this  paper.

-------
floor  slab,  and the  SFt  concentration  in  the  exhaust pipe  and
basement  air allowed to stabilize.  At time 85,  another 1/2 - in.
diameter   hole  was  drilled  in  the  floor   slab  and  the  SFC
concentrations again allowed  to stabilize. At  time 110 a  3.1  sq
in.  hole  was opened  in  the  sump  cover,  and  the concentration
stabilized.  For the final test  hole, at  time 150, the  3.1  sg in.
hole in the  sump cover was reduced to 1 sg in., and again, the SFC
concentration allowed to stabilize.
  a
  z
  o
  H
  W


  8
   «
  CO
                                                            I 4 sq in.
                                                            slab leakage
                                                            area
3 5 sq in
slab leakage
area
0 4 sq in.
slab leakage
area
                            60         90
                              ELAPSED TIME (min.)
                                                120
          150
                                                                    160
Figure 3. Tracer gas experiment performed in House VA2 to determine
the effect slab penetrations have on the sub-slab depressurization
system.

     The same type of experiment was performed in House PA1. SF, was
injected into the basement  . A number of penetrations were made  in
the  slab.  The  ratio   of   the  basement  to  exhaust  pipe  SF6
concentration, along with the volume of air being exhausted by the
sub-slab system, can  be  used to determine the additional  energy

-------
 costs incurred  by the use of  a sub-slab depressurization system.
 The data collected during this experiment are illustrated in Figure
 4.
                                     0 2 sq in.
                                     slab
                                     leakage
                                     area
                      0 3 sq in
                      slab
                      leakage
                      area
01 sq in.
slab
leakage
area
                                                      Basement concentration

                                                      Pipe concentration
                30
60
   90      120      150
   ELAPSED TIME (min.)
Figure 4. Tracer gas tests performed  in House PA1 to determine the
effect  slab penetrations have on  the performance of  a  sub-slab
depressurization system.
RESULTS
          The results of the tests for the energy losses performed
in the three  houses are presented in Table l. The estimated energy
costs were  determined  by using  the degree  day method.  A furnace
efficiency of 90% was used for Houses VA1 and VA2 and 75% for House
PA1.

-------
TABLE 1. APPROXIMATE ADDITIONAL ENERGY COSTS DUE TO LOSS OF CONDITIONED AIR
THROUGH THE USE OF A SUB-SLAB DEPRESSURIZATION SYSTEM
HOUSE
NO.

VA1
VA2
PA1
% CONDITIONED
AIR BEING EXHAUSTED
BY SYSTEM
43
47
21
AIR VOLUME
LOSS (cfm)

34
11
3
HEATING
DEGREE
DAYS
5010
5010
5827
EST. ADDITIONAL
ENERGY USAGE
(10s Btu)
4.9
1.6
0.6
     The results  of the  tests performed to  determine the  effect barrier
penetrations have on the performance of  the sub-slab system in House VA2 are
presented in Table 2.

TABLE 2.  SYSTEM PERFORMANCE AS EFFECTED BY FLOOR PENETRATIONS IN HOUSE VA2.
LEAKAGE AREA
(sq in.)
AIR VOLUME AT
EXHAUST (cfm)
PRESSURE DIFFERENTIAL
PIPE TO BASEMENT (in. H,0)
SLAB INTACT
0.8
1.6
2.6
4.7
24
25
27
30
30
0.212  ±
0.194  ±
0.180  ±
0.115  ±
0.071  ±
0.005
0.004
0.005
0.004
0.005

-------
Equation 1 is used to determine the fraction of basement air being
exhausted by the sub-slab depressurization system:
Cs = f(Cb
                                   +  fCa  x
(1)
                                Qt
where:
     Cs   = SF6  concentration  in  the  exhaust  pipe stack
     Cb   = SF6  concentration  in  the  basement
     Qbs  = Volume of air flowing from the basement to the
            stack
     Ca   = SF«  concentration  in  the  outdoor  air getting back
            inside the stack
     Qas  = Volume of outdoor air getting back inside the
            stack
     Qt   = Total volume of air being exhausted by the system
Factoring this equation results in Equation 2 where:
                Qt x Cs = fCb x Obs) + (Ca x Oasl
                 Cb                Cb
                                          (2)
Assuming that  Ca is zero, which is not quite true but close enough
considering the miniscule amount of SF6 being exhausted compared to
the volume of outdoor air, and factoring again results in Equation
3:
                              ££
                              Cb
                                          (3)
                  Qt
Equation 3 translates to:  the ratio of the SF« concentration  in the
stack to the concentration in the basement  equals the  ratio  of the
volume of air flowing from the basement into the stack  to the total
airflow out of the stack. Therefore, if the  SF« concentration  in the
stack is 31 ppm and the SF, concentration in the basement is 72 ppm,
applying Equation 3  results  in  a value of  0.43 or 43% of the air
being exhausted by the sub-slab depressurization  system is  coming
from the basement.

Equation  4,  used for determining the additional  cost of  energy
associated with the use of the sub-slab depressurization  system,

-------
is known as the degree day method where:


              Ht = V X 60  X 24 X 0.018 X HDD X Eff         (4)

where:

               HL =      Heat loss due  to  exfiltration of air by
                         sub-slab system for the heating season
               V =       Volume of air  being exhausted by system
                         in cubic feet per minute
               60 =      Number of minutes in an hour
               24 =      Number of hours in a day
               0.018 =   Heat capacity  of  air in British thermal
                         units   per   cubic    foot  per   degree
                         fahrenheit.  This means  the  number  of
                         British  thermal  units  required  to heat
                         each 1 cu ft of air 1°F.
               HDD =     Heating degree days. This is the average
                         daily  outdoor  temperature  in  degrees
                         fahrenheit,  subtracted from  the indoor
                         temperature setpoint of 65 °F. This daily
                         temperature difference is totalled for the
                         entire  heating season.  HDDs  should  be
                         thought  of  as  the  number  of  degrees
                         fahrenheit for the heating season that
                         outdoor air entering the house must be
                         raised to reach the indoor temperature
                         of 65 °F.
               Eff =     Furnace efficiency.  In this case 90%

Therefore, if the sub-slab depressurization  system is exhausting
basement air at  a volume of 34 cu ft/min.,  as in House VA1,  the
volume of air being exhausted each day is 48,960 cu ft.  To  find the
number of British thermal units required to heat the exhausted air
1°F,  multiply the volume  of air  by the heat capacity of  1  cu ft of
air. This results in a heat capacity of  881.3 Btu for each 1°F that
the air must be heated.  Finally,  by multiplying the heat  capacity
of the exhausted  air  by  the number  of HDDs per heating season at
the  location  of  the  house  (for   this  house,   5010  HDD)  and
considering  the  efficiency of  the  furnace,  the  added amount of
energy required  to heat the air being  exhausted by the  sub-slab
depressurization  system  is  4.9 million  Btu.

-------
     Table 3 presents the amount of energy lost in House VA2 with the various
size holes in the floor slab. Again, a heating degree day value of 5010 HDD
and a furnace efficiency of 90% are used.

TABLE 3. APPROXIMATE ADDITIONAL ENERGY COSTS DUE TO LOSS OF CONDITIONED AIR
IN HOUSE VA2 WITH VARIOUS SIZE FLOOR SLAB PENETRATIONS.
LEAKAGE
AREA
(sq in.)
SLAB INTACT
0.8
1.6
2.6
4.7
% CONDITIONED
AIR BEING EXHAUSTED
BY SYSTEM
47
52
57
75
77
AIR VOLUME
LOSS (Cfm)

11
13
15
23
23
EST. ADDITIONAL
ENERGY USAGE
(10* Btu)
1.6
1.9
2.1
3.3
3.3
     The  results  of the  tests performed to  determine the  effect barrier
penetrations have on the performance of the  sub-slab system in House PA1 are
presented in Table 4.

TABLE 4.  SUB-SLAB DEPRESSURIZATION  SYSTEM PERFORMANCE AS AFFECTED BY FLOOR
PENETRATIONS IN HOUSE PA1
LEAKAGE AREA
(sq in.)
SLAB INTACT
0.4
0.8
1.2
AIR VOLUME AT
EXHAUST (cfm)
15
16
19
20
PRESSURE DIFFERENTIAL
PIPE TO BASEMENT (in. H,0)
0.80 ±
0.81 ±
0.78 ±
0.77 ±
0.01
0.01
0.02
0.01

-------
     Table 5 presents the amount of energy lost in House PA1 with the various
size holes in the floor slab. Again, a heating degree day value of 5827 HDD
and a furnace efficiency of 75% are used

TABLE 5. APPROXIMATE ADDITIONAL ENERGY COSTS DUE TO LOSS OF CONDITIONED AIR
IN HOUSE PA1 WITH VARIOUS SIZE FLOOR SLAB PENETRATIONS
LEAKAGE
AREA
(sq in.)
SLAB INTACT
0.4
0.8
1.2
% CONDITIONED
AIR BEING EXHAUSTED
BY SYSTEM
21
29
36
41
AIR VOLUME
LOSS (Cfm)

3
5
7
8
EST. ADDITIONAL
ENERGY USAGE
(10* Btu)
0.6
0.9
1.3
1.5
CONCLUSIONS

     Although a large percentage of the air being exhausted by the sub-slab
depressurization  system  can  come  from  inside  the  building  and  not  from
beneath the  slab, the amount of air  that  this percentage constitutes  is
relatively small. The estimated annual additional costs associated with the
use of the sub-slab depressurization  system  for the three houses tested (not
including electrical costs to run the fan) are illustrated in Figure 5. For
the dollar costs, a value of $0.65 per therm (10* Btu)  of gas  is  used.
     The effect on the performance  of the sub-slab system of even such small
holes in the floor slab that were tested here seems to be a greater concern
than the  additional  costs  of energy. For example, at  House VA2, a leakage
area of 4.7 sq in.  resulted in a drop in the  negative  pressure  at the exhaust
pipe of 67%.  A loss of this much pressure  could have  serious effects  on a
marginally operating system.

METRIC CONVERSION FACTORS

     Readers more familiar with metric units may use the following to convert
to that system:

     Nonmetric           Multiply by              Yields Metric

     Btu                      1.055                    kJ
     °F                       5/9 (°F-32)              *C
     ft1                       28.3                     L
     in.                      2.54                     cm
     in. H,0                   6.45                     cm'
     in. H,0                   2.5                      Pa
     therm                    105,506                  kJ
11

-------
      40
      30 -
  !     1
  s   »
  _;     H
                          $31.85
                           VA1
   VA2

HOUSE NO.
                                                PA1
Figure 5. Estimated additional annual energy costs associated with
the  use   of   a  sub-slab  depressurization   system.  Electrical
consumption of  the fan  in  each  system is not included.
12

-------
                                                         D-VII-2
EXHAUST OF CONDITIONED AIR BY RADON MITIGATION SYSTEMS IN HOMES

          by:   R. J.  Saultz
                Oak Ridge National Labs
                Oak Ridge, Tennessee
                      WITHDRAWN BY AUTHOR

-------
                                                               D-VII-3
    LONG TERM DURABILITY AND PERFORMANCE OF RADON MITIGATION SUBS1AB
                        DEPRESSURIZATION SYSTEMS

                 David T. Harrje and Kenneth J. Gadsby
              Center for Energy and Environmental Studies
               Princeton University, Princeton, NJ  08544

    David C.  Sanchez, Air and Energy Engineering Research Laboratory
                  U.S.  Environmental Protection Agency
                   Research Triangle Park, NC  27711

                                ABSTRACT

   Subslab depressurization  (SSD) systems have become the most popular
radon mitigation technique used by professional mitigators.  Readily
available supplies and equipment from other trades are commonly utilized in
the installation of these systems.  What is not currently known is the long-
term performance of some of  the caulks or sealants and the mechanical
devices  such as fans or controls.  The operating environment for the fans is
often different from that for which they were designed.  Specifically, this
involves studying whether the fans are still operating as installed,  whether
there has been an increase or decrease in the radon levels after several
years of operation, and whether the house occupants are operating the
systems  the way the systems were originally designed.

   To answer these questions, we have tested eight of the Piedmont Study
houses as veil as houses in which post-mitigation measurements by the New
Jersey Department of Environmental Protection (NJDEP) have shown readings
that were above the U.S. EPA guideline of 4 pCi/L.  Quarterly visits to each
house involved measurement of flows and pressures in the mitigation system,
taking radon grab samples in the basement and of the mitigation system
exhaust, interviews with the occupants, retrieving and installing track-etch
detectors, and a visual inspection of the mitigation system and house
substructure.  This paper presents the findings of this research and
discusses some of the durability aspects of SSD systems and their
interaction with the house substructure.

   This paper has been reviewed in accordance with the U.S. EPA peer and
administrative review policies and approved for presentation and publica-
tion.  Mention of trade names does not constitute endorsement by the Agency.

-------
INTRODUCTION AND BACKGROUND

    There is increasing evidence that the health risks in those houses
with significant levels of radon gas (above the EPA guideline of 4 pCi/L)
may constitute the most serious indoor air quality problem in the United
States.  Radon gas intrusion is often pictured as a seasonal phenomenon,
with stack effect and other pressure-driven factors influencing soil gas
entry to building substructures.  A number of solutions have been proposed.
These approaches involve energy use as well as indoor air quality concerns.
The proposed solutions must be tailored for the specific nature of the radon
source.  If the radon enters the house via veil water, one approach is
necessary; radon in building materials may suggest other strategies.  In
this report our attention will be focused on radon entry with soil gas
through the building substructure, and what that mechanism implies for radon
mitigation.

   Even limiting our scope to soil gas entry, many solutions exist to
reducing radon gas concentrations in the house.  Radon removal is one method
and is often the preferred approach.  Local exhaust, after radon entry into
the living space, is another strategy.  Ventilation used to dilute the radon
gas concentrations is still another option.  Each approach must be matched
to a given radon condition in the individual building.  This report will
consider only the method of radon removal from under the slab in basements
or crawl spaces.  Known as subslab depressurization, SSD, this mitigation
approach has been proven to be very effective, often removing enough soil
gas containing radon so that substructure measurements of radon
concentrations drop by 90% or more following mitigation.

   1.  Theory of Operation of SSD Systems for Radon Mitigation

   The theory of operation for  the SSD system is that, by penetrating the
concrete floor slab with an exhaust pipe, one gains access to the area
beneath the slab.  The area, often a gravel bed, serves as a collection site
for the soil gas containing radon.  The negative pressure provided by the
exhaust pipe causes the soil gas to be removed from the gravel bed  and, if
communication exists between the subslab volume and the walls of the
building,  soil gas will simultaneously be exhausted from the walls.  The
exhaust mechanism can be passive which implies that suction pressures
beneath the slab will vary on a seasonal basis, with  the greatest suction
occurring  during the coldest weather due to  increased buoyancy of the air in
the vertical exhaust stack.  In the systems  tested in this research,  exhaust
fans were  used.  These  "active  systems" were  shown to maintain near constant
suction pressures under the slabs all year long.

   The key point to remember in the merits of year round radon removal  is
that there is no guarantee that radon problems will not be present  even in
the summer months.  The radon  levels  found in individual houses  are a
complex result of radon source  strength, soil transport, weather, and the
way the house  is operated.  To  be certain  of maintaining low radon  levels in
the house  normally requires that  a  SSD mitigation system work properly  24
hours  per  day, 365 days per year.   It is for this  reason that durability and
system performance are  such important considerations.  Performance  level
goals  are  for  100% on-time operation  for the life  of  the building.  Thus we
require excellent durability of system components  and a reliable means  for

-------
determining whether the system is fully operational at all times.

   The lack of long-term data on SSD systems is a major stumbling block in
determining whether or not the SSD systems perform adequately.  This project
has been directed toward gathering such data from eight research houses that
were part of the Piedmont Study(l), as well as houses tested by the NJDEP as
a follow-up to mitigation activities.

   2.  Operational Environment

   The question of durability of the mitigation system arises not only
from the need for a lifetime operation in the house, but concerns about the
environment to which the SSD system is subjected(2).  Soil gas is often very
moist, causing condensation problems in the piping and at the fan of the
mitigation system.  Also, particles can be drawn from the gravel bed; they
in turn may line the pipes and deposit on the fan or possibly interfere with
the fan bearings.

   The moisture removed from the subslab can be very substantial, and
could amount to many gallons of water per day(2).  Unless the piping design
allows for that water to drain back into the soil, the water could clog the
piping or interfere with the fan operation.  Evidence of the moisture and
other debris has also been found in the staining of roofs near the exhaust
pipes of the SSD systems.  Drying soil could result in slab cracking.

   The amount of sand and other particles sucked from the soil must be
viewed as a possible cause of bearing failure or of the generation of
bearing noise (such effects can also be caused by the moisture).  Noise can
directly influence the occupant to shut down the SSD system.

   Another environmental effect that should not be overlooked is the
amount of airflow through the fan.  To remain at an appropriate operating
temperature requires sufficient air to remove fan motor heat.  Operating the
fan in a way that the blades are stalled can result in high bearing loads
and invites early fan failure.

OBJECTIVES

   The following are our objectives for this research:

   1.  Our first objective has been to document the ability of the SSD
radon mitigation system to maintain houses at radon concentration levels
below the EPA guideline of 4 pCi/L.  In these measurements we hoped to
observe the influence of such parameters as seasonal factors.  Local weather
effects (e.g., rain storms) are covered in more detailed radon
monitoring(2).

   2.  A second objective was to observe the long-term characteristics of
radon levels in the SSD system exhaust.  Source strength and transport
properties of the soil may be determined from these measurements.

   3.  A third objective was to evaluate the long-term influences of the
SSD system operation on the house substructure.  Since our objective is
long-tern) operation,  we need to know of any associated consequences.

-------
4.  A fourth objective of the study is to determine critical parameters that
can degrade SSD performance and recommend ways to minimize such degradation.

THE TWO-PRONGED APPROACH FOR DURABILITY TESTING

   Our approach to evaluation of durability is based upon our own
experience as to what might happen over time as well as on the experiences
of others; e.g., New York State Energy Research and Development Authority's
efforts to quantify durability(3), Lawrence Berkeley Laboratory research(4),
and Swedish studies which were able to look at houses after 5 years of
operation(S).  Three short data sheets have been developed.  One emphasizes
the house and mitigation system as observed by the house occupants; the
other data sheets involve a series of diagnostic tests to determine if the
mitigation system is achieving the necessary radon mitigation goals.

   1.  The house occupant questionnaire

   Reviewing the data sheet, Radon Durability Diagnostics - I. the
emphasis is initially placed on whether the system has been running
steadily.  Swedish studies have pointed to this problem as an explanation of
increasing radon concentrations(5).  Our own experience is that occupants
hate to admit shutting off the system although system noise, radio
interference, and conserving electricity during the summer have been offered
as reasons to turn off the fan.

   The second question concerns noise perceived by the house occupant.  If
the system is becoming noisy, the occupant fears that the fan may be "on its
last leg," or that any noise may prompt an occupant to shut down the system.

   The third question involves moisture.  Ve are seeking to gain insight
into condensation, collection of water in the mitigation piping, or moisture
related events taking place at the roof exhaust.  Water in the piping can
directly influence the amount of exhaust airflow possible.  Condensation can
be another cause for occupants to turn off the mitigation system.

   The fourth question is aimed at finding out about possible power
outages, construction in the house, or other events that could account for
higher than expected radon levels.

   Question five involves the house occupants' perception of the system
and whether or not they have any questions about the way it functions.

   2.  Diagnostics and measurements

   The second data sheet. Radon Durability Diagnostics - II. emphasizes
diagnostic procedures employed by the visiting inspection team.  The
observations are initially visual: the house type, special features,
location of the SSD system, etc.  On subsequent visits the visual
observations concentrate on location of cracks, sealed areas, etc.  For
example, are there new cracks or places through which radon might enter the
house substructure?  The inspection also concentrates on what sealing was
done previously and how well it has held up.

   The second item involves noise generation.   A stethoscope is used in

-------
order  Co better detect early signs of bearing squeal which is an indication
that fan lifetime may be short.

   Item three  is a diagnostic check of the airflow in the mitigation
system piping.  A heated-wire anemometer inserted through an inspection hole
to the center  of the plastic pipe is used to measure the velocity.  (In
houses where durability tests are in progress, openings in the piping are
sealed with duct tape to provide easy access during the study.)  Care must
be taken that  the airflow probe is sealed where it enters the pipe in order
to prevent erroneous airflow readings.

   Item four makes use of those same openings in the piping to evaluate
pressure differentials.  In this case, a digital readout micromanometer is
used.  For this instrument, using the scale "inches of water," the readout
is the most sensitive (0.004 in. of water - 1 Pa).

   Item five deals with measurement of radon levels in the exhaust of the
mitigation system.  Over the long term, questions arise as to whether the
radon  is being depleted or whether the soil is drying out; the latter could
lead to the system exhausting more distant radon gas.  The radon levels in
the system exhaust would be an indication of such changes.  Both pumped grab
samples and evacuated grab samples with Lucas cells are used in this
testing.

   The final item is general observations.  This is an opportunity to note
what a general inspection has revealed, and what may have influenced or
interfered with the radon mitigation process.

   We  also use the data sheet, Radon Durability Diagnostics - III.  We
have been noting on this data sheet the serial numbers of alpha track
detectors that were in place as well as the replacement detectors.  Normally
detectors have been placed in the basement/crawl space areas as well as on
the first floor.

RESULTS FROM THE NEW JERSEY PIEDMONT HOUSES

   As  previously discussed, there are four objectives of the study on
which  we must  focus attention.  Perhaps the easiest way to review the
results is to plot the radon levels measured over time for each of the
Piedmont houses.  Typical plots are shown in Figures 1-2 and summarized in
Table  1.

   1.  Occupant Effects

   Immediately evident (in looking at Table 1) is that, while the majority
of the houses  show more-or-less constant radon levels over time, two houses
(Houses 3 and  5 in Figure 1) show major variations in radon levels.  These
variations include radon concentrations above the EPA guidelines as well as
House  3's return to premitigation levels.

   On  checking the occupant questionnaire it was noted that the occupant
of House 5 had turned off the SSD system because of radio interference and
because the occupant felt that, under summer conditions with open basement
windows, it was wasteful in terms of energy to operate the SSD system(6).

-------
   The occupant of House 3 gave no hint on the questionnaire that the SSD
system wasn't working 100% of the time.  Only when we informed the occupant
of the return to pre-mitigation radon values did the occupant remember that
the system had been turned off during a dinner party (room adjacent to the
fan) when mitigation fan noise was annoying and they had forgotten to turn
it back on for an extended period.  We pursued this point further and found
that the noise was the result of vibration between the SSD fan motor and the
wood-framing in the attic above the garage.  A small modification in the
mounting eliminated the problem.  Similar vibrational problems have also
been experienced in one of our current research houses.  The annoyance of
the vibration had resulted in the system being turned off.  The importance
of avoiding such problems should be emphasized with the mitigator.

   One point was very clear from even this very limited number of test
houses.  The SSD system should not be turned off even for short periods of
time without immediately impacting the radon level.  One occupant explained
that the system was only turned off for radio weather broadcasts to avoid
the static.  The lesson is that the static should not be present through the
use of a higher quality speed controller, and careful checking of the wiring
arrangement would avoid interaction with sensitive electrical equipment.

   2. Seasonal Variability

   In order to look for such effects as seasonal variability, we must
focus our attention on houses where the occupants have allowed the SSD
systems to operate 100% of the time.  Figure 2 (also see Table 1) indicates
this type of operation.  A very simple evaluation of events over the
measurement period using Table 1 data (the basis for the figures) will
demonstrate the stability of radon concentrations over time.  In this
exercise we have averaged readings for the first two periods and compared
them to the last two measurement periods.  Altogether 10 measurements can be
compared if all basement and crawl space values are averaged equally.  Thus
comparing October 1987 through May 1988 to November 1988 through June 1989
(a similar weather period) we find that in 7 out of 10 measurements radon
levels have dropped an average of 0.6 pCi/L.  In 2 out of 10 houses, radon
levels have increased by an average of 0.1 pCi/L while one value has stayed
the same.  Since the alpha track method of measuring radon levels has an
error band that would include these variations, our conclusion would be that
there is no significant change over the year.

   Because of these same arguments of measurement error it is even more
difficult to look for seasonal effects.  However, any significant seasonal
influences should be evident in these data.  House 2, for example, indicaces
that winter readings are slightly higher than summer readings.  But again
using October-February and November-March basement values we find a 1.6
pCi/L average versus 1.2 pCi/L for May-November combined with March-June.
This 0.4 pCi/L difference is well within the error band.

   House 4 shows a general decline in radon concentrations with time and
no sign of seasonal fluctuations.

   House 6 shows only a slight increase in the November-June reading and a
noticeable drop in radon level after the first October-February period.

-------
   House 7 also shows a very small increase in the November-June period in
basement and crawl space radon concentrations.  First floor concentrations
are basically background levels.  Note that the substructure concentration
change is from approximately 0.5 to 1.0 pCi/L, again well within the error
band of the alpha track radon sensors.

   Only in Houses 8 and 10 (Figure 2) do we see variations in the
November-March period that could be viewed as seasonal influences.  In the
case of House B similar increased concentrations were measured during the
October-February period, the year prior.  Averaging basement and first floor
values for House 8 (from Table 1) for the "winter periods," we obtain an
average value of 4.3 pCi/L, a value in excess of the EPA guideline.  If we
average "summer periods" the value is only 1.8 pCi/L.  This house
illustrates how seasonal influences can clearly affect SSD performance.

   House 10 indicates only a slight increase in radon concentration in the
October-February period; i.e., 2.2 pCi/L average versus 1.5 pCi/L for the
"summer periods."  However, the increase is more substantial in the
November-March period where values of 2.7 pCi/L are observed.  This house
would also appear to exhibit seasonal effects that increase radon
concentrations by approximately 1 pCi/L.

   3.  Radon Levels in the Mitigation Exhaust System

   One measurement of durability testing that is of interest is the
concentration of radon in the exhaust pipe from the subslab mitigation
system.  When this information is combined with the velocity (or flow
measurement) at the same pipe location, we are in a position to calculate
the total flow of radon gas from the mitigation system(7).

   One question to be resolved is:  Can we compare the amount of radon
exhausted from any given house and understand fully the role played by the
mitigation system?  One such comparison involves the natural flow of radon
through the same house.  The physical meaning would be a radon flow rate
equal to the quantity of radon exhausting through a "perfect mitigation
pipe" through which the airflow rate is equal to air infiltration rate of
the house. To make the calculation of the natural flow requires a knowledge
of the average air infiltration rate for the house, the radon concentrations
upstairs and in the basement/crawl space, and the volumes of those zones.
The calculation is:

                              (Cu x Vu) + (Cb/c x Vb/c)
              R - AI x 103 x   	-	-
                                    Vu + Vb/c

   where R   - radon flow, pCi/h
           AI  - air infiltration, m3/h
           C   - radon concentration, pCi/L
           V   - volume, or
           u   - upstairs
         b/c   - basement/crawl space
   Comparative results for five Piedmont houses are shown in Table 2.  The
ratio of the radon gas exhausted from each house via the mitigation system
is compared to the radon natural flow value through the house.  Ratios vary

-------
 from 1  to  9.   The  question that  comes  to mind  is:  "What is special about
 House 4 which has  the  same flow  of  radon gas both  through the mitigation
 system  or  by  natural means?"  One answer could be, it is the house with the
 least porous  soil  -- basically pure wet clay.  It  is a house where high
 ventilation rates  (such  as using a  blower door) depress the radon levels,
 and then it takes  many hours  for the house to  return to the previous
 elevated levels.   Such behavior  has been interpreted as evidence of a
 limited radon entry rate.   However, just comparing the natural radon entry
 rate with  those of the other  houses would not  single out this house.  In
 fact, looking at natural radon entry rates, it is  only House 3 that stands
 out with a rate far above  the others;  i.e., 35 x 106 pCi/h versus 7.5 + 3 x
 10° pCi/h  for all  the  rest.   House  3 has a soil condition of high porolity,
 (i.e.,  stone  flour roughly 1/8-in.  (0.3 cm) diameter and a good gravel bed),
 Just the opposite  conditions  of  House  4.

    The  total  amount of radon  gas mechanically  exhausted from the soil
 varies  by  a factor of  7  in these houses.  The  lowest value is for House 4
 with clay  soil.  The highest  value  of  radon is for House 3 with very porous
 soil.   The two highest houses (Houses  3 and 5) were those with recurring
 periods of above-guideline  levels of radon.  However, House 7 has clay soil
 mixed with shale and ranks  third.

    The  method used to  analyze the radon concentrations from the mitigation
 system  exhaust involves  Lucas cells.  Analysis of  the Lucas cells takes into
 account the time elapsed from the time the sample was taken to the time it
 was analyzed,  the  background  level  of  the cell, and the efficiency of the
 measurement equipment.   The relationship used  is:


                           (CPMread  - CPMbgd) x e«>-00803 x At)
         cactual   ~        	
                                       
-------
period and House 3 in which real levels have increased because of airflow
variation.  This will be discussed later.

   In the second and third testing periods, when we measured exhaust
concentrations of radon (3/89 through 6/89, and 6/89 through 11/89), the
concentrations vary with the individual houses.  Houses 3 and 10 show a
decreasing trend, while Houses 2, 4, 5, 7, and 8 indicate increasing levels.
Some seasonal effects may be present in that the lowest readings for the
majority of the houses were in months 15 and 18 which is the warmer time of
the year [the plots start at month 11 (November 1988); therefore, 15 is
March 1989 and 18 is June 1989].  Throughout these discussions, radon
concentration has been used since, hopefully, it provides the more physical
meaning, and because airflow rates (necessary to provide total radon flow as
shown in Table 2) tend to be constant in the houses.  One exception to
constant flow is House 3 where, after the fan had been turned off by the
occupant and after radon levels had increased in the living space, the fan
speed was increased.  Duct velocities changed from 2.65 m/s in November 1988
to 8.2 m/s in November 1989 (intermediate readings were 7.2 m/s).  The trend
of falling concentration levels over time for House 3 does apply for the
last three periods, with a peak at month 15.  Because of the lower fan
airflow rates present during month 11, radon flow rates are less than half
the month 15 peak.  The profile of House 3 is opposite to the general trend
experienced in the majority of the test houses.

   4.  Substructure Changes

   Based upon diagnostic team observation, only House 3 showed evidence of
physical changes.  Two cracks appeared in the basement slab near the slab
edge extending toward the center of the room.  Length of the cracks was
approximately 6-ft (1.8 m) and the width exceeded 1/16-in. (0.2 cm) at some
locations.

   A noteworthy observation of what was happening in the basements or
beneath the slab was that conditions were noticeably drier.  Several
occupants have stated that the need for summer dehumidification was
eliminated in their houses.  Where observations were possible, water in
gravel beds was no longer visible.  No quantitative measurements of relative
humidity have been made.

PRELIMINARY RESULTS FROM NJDEP HOUSES (MITIGATED BY PROFESSIONALS)

   The concern for durability and performance of radon mitigation systems
was pointed out by Depierro and Cahill of the NJDEP(8).  Based upon their
findings, 64% of the houses mitigated by occupants and professional
mitigators were not achieving the 4 pCi/L guideline.  Even when only
professionally mitigated houses were assessed, the percentage of houses
failing to meet guidelines still exceeded 50%.

   With this information as background, we undertook a program of
upgrading the radon information.  Our test houses were taken from the larger
list of NJDEP-tested houses.  From that list our criteria of selection were
houses less than half an hour's drive from Princeton, houses with
predominantly SSD systems installed, and houses with the highest post-
mitigation radon levels.  Our approach was to question the occupant on the

-------
radon history in the house, to briefly inspect the mitigation system
installation, and to leave a charcoal canister in the house which 'after 3
days the occupant would mail in for analysis.  Two examples are given.

   Test house A.  This house has a SSD system which uses three slab
penetrations routed to a fan mounted on a cinder block wall.  Noise from the
fan caused the occupant to build an insulated box around the fan.  The  house
in early 1987 had measured radon levels from 40 to 120 pCi/L in the basement
and 13 to IS pCi/L upstairs.  The values disagree between the state and the
private mitigator.  Measurements 1 year later, after mitigation, indicated
levels reduced to less than the EPA guideline value.  Our measurements  in
October 1989 confirmed that the guideline was being met.

   Test house G.  The SSD system in this house has two floor penetrations
as well as a penetration at the wall under the crawl space.  Early readings
in September 1987 were 143 pCi/L in the basement and 75 pCi/L in the living
room.  Levels dropped to only 6-7 pCi/L in the basement and living room
until an improved fan, Kanalflakt K-6, replaced the original fan and dropped
levels to the 0.8 pCi/L level.  Our tests in August 1989 indicated minimum
detectable concentrations of 0.54 pCi/L.

   To briefly review these data, we found no cases where the substructure
radon levels were above the 4 pCi/L guideline.  This was true even though we
chose the highest radon houses on the list supplied by NJDEP.  One reason
for this was that occupants continued to have their systems checked and
improved with better fans and higher flow rates.  Fan lifetime appears to be
short in several of the fans used, while the Kanalflakt fans continued to
perform satisfactorily.  Similar to the Piedmont houses and our other test
houses, noisy fans can be a result of fan mounting, necessitating good
installation procedures if this critical problem is to be avoided.
CONCLUSIONS AND RECOMMENDATIONS

   1.  House occupant interaction significantly affected the operation of
the mitigation systems, based upon our survey.

   One point that is very evident from the durability diagnostics program
is that increases in radon levels can often be traced back to occupant
intervention with the SSD radon mitigation system.  In the majority of
cases, noise from the system caused annoyance and the system was shut off,
often for prolonged periods.  Another instance was an occupant seeking to
limit operation of the system in the summer period when he erroneously
thought that open basement windows would suffice, and therefore the
electricity to operate the SSD system could be saved.  Even in this
instance, radio interference --a different kind of noise -- instituted the
habit of this occupant to shut off the fan.

   2.  Installation and Building Anomalies

   Every house is different from a radon standpoint.  Even in the same
area, where the weather is the same, the soil porosity and radon content
often varies noticeably across an individual basement floor, wall systems
below grade are highly variable, equipment used to space condition the house

-------
can drastically alter interior pressures and that condition can vary from
room-to-room, the house design can have a major impact,  etc.   Given this
list of variables it is no wonder that any mitigation system must take these
factors into account.  The installation must also consider moisture in the
soil gas.  One of the Piedmont houses experienced major  problems with water
in the mitigation piping.  The problem was lack of attention to the slope of
the piping, causing condensation water to collect and not drain to the
gravel bed.

   Anomalies often can be associated with the unknown conditions that
prevail beneath the basement and crawl space slabs.   The amount of gravel,
hidden pipes, or partial concrete pours that separate the total floor area
into several zones requires that more than one slab penetration must be
used.  Ve observed other houses with wall suction at the wall adjacent to
slab-on-grade construction.  Here extra radon entry was  expected and
basically taken care of by the wall penetration.  In another house, very
similar when viewed casually, the SSD system may perform beautifully with
one penetration and remove radon from the walls at the same time.

   Space conditioning cannot only circulate radon in the house but can
also directly influence the amount present in the house.  Ve were aware that
inadequate returns in warm air systems resulted in under-pressurization of
the basement and more soil gas being drawn into the house.

   3.  Recommendations

   Our recommendations to the installer of mitigation systems is to use
long-service-life components.  The fan should be designed  to operate  for
long periods; the fewer replacements and  the more trouble  free,  the safer
the occupants.  Since wide variations in  fan flow rates were found between
different  regions of the country, the choice of fan must take this  into
account  to avoid fan overheating and/or bearing failure.   The same  criteria
of quality apply to  sealants.  This  is a  long-term installation and not  the
place  to use cheap caulk.  Installations  at the sump require careful  design
so that  the water-removal  function of the sump pump does not compromise  the
SSD  system and vice  versa,   when done properly we observe  little or no
degradation  over time of these installations.

   A clear perception of the characteristic of  soil gas  requires that the
mitigation system installer  must properly slope all piping to drain any
condensing moisture  into the gravel  bed.

   Electrical hookup of  the  SSD system should not generate electrical
noise  in sensitive household equipment such as  radios.   The  installer should
check  with the occupant  at the time  of  installation of  the system  that such
problems do  not exist.  Mechanical noise  and vibration  should also be elimi-
nated  to prevent annoyance to the occupant and make certain noise  is  not a
cause  for  system shutdown.   Several  of  the Piedmont and research house in-
stallations  suffered from  fan vibration  interacting with nearby structures,
or with  mitigation piping.   These were easily corrected.   Checking for such
noise  problems at the time of installation is highly recommended.

   The occupant not  only is  sensitive to  SSD system noise  but also is
normally very interested in  maintaining  the lowest possible radon  levels.

-------
We recommend that the occupant regularly retest for radon.  In the NJDE?
houses, retesting led directly to system improvements and system operation
better suited to the long-term (improved fans,  sound suppression, etc.).
These houses lowered radon levels to values below EPA guidelines.
Installation of a pressure switch-activated indicator light to inform the
occupant when the system is not operating properly is highly recommended.

ACKNOWLEDGEMENTS

   This work was funded by the U.S. Environmental Protection Agency under
Cooperative Agreement No. CR-814673 with Princeton University.
REFERENCES

1.  Dudney, C.S., et al., Investigation of Radon Entry and Effectiveness of
Mitigation Measures in Seven Houses in New Jersey, Oak Ridge National
Laboratory Rpt. ORNL-6487, August 1989.

2.  Harrje, D.T. and Hubbard, L.M., Proceedings  of the Radon Diagnoseics
Workshop. April 13-14. 1987.  EPA-600/9-B9-057 (NITS PB 89-207898), June
1989, also see Harrje, D.T., Hubbard, L.M., and Sanchez, D.C.. "Diagnostic
Approaches to Better Solutions of Radon IAQ Problems," Healthy Buildines '88
- Planning. Physics and Climate Technology for Healthier Buildings. Vol. 2,
Swedish Council for Building Research, Stockholm, Sweden, 020:1988 (ISBN* 91-
540-4933-4), pp. 143-152.

3.  Nitschke, I., et al., "Preliminary Results from the New York State
Radon-Reduction Demonstration Program," Proceedings of the 1988  Symposium
on Radon and Radon Reduction Technology. Volume 1, EPA-600/9-89-006a (NIIS
PB89-167480), March 1989.

4.  Prill R.J., Fisk, V.J.. and Turk. B.H., "Monitoring and Evaluation of
Radon Mitigation Systems Over a Two-Year Period," Ibid.

5.  Nilsson, I. and Sandberg, P.I., "Radon in Residential Buildings -
Examples of Different Types of Structural Counter-Measures," Healthy
Buildings '88 - Planning. Phvslcs. and Climate Technology for Healthier
Buildings. Vol. 2, Swedish Council for Building Research, Stockholm,  Sweden
020:1988 (ISBN 91-540-4933-4), pp. 163-172.

6.  Harrje, D.T., et al., "The Effect of Radon Mitigation Systems on
Ventilation in Buildings." ASHRAE Transactions 1989, Vol. 95, Ft. 1.

7.  Harrje, D.T. and Cadsby, K.J., "Airflow Measurement Techniques Applied
to Radon Mitigation Problems."  Proceedings of the 10th AIVC Conference -
Progress and Trends in Air Infiltration and Ventilation Research. AIVC,
Coventry, UK, 1989.

8.  Depierro, N. and Cahill, M., "Radon Reduction Efforts in New Jersey,"
Proceedings of the 1988 Symposium on Radon and Radon Reduction Technology.
Volume 1, EPA-600/9-89-006a (NIIS PB89-167480), March 1989.

-------
                     DURABILITY-HOUSE  3
<

-------
                     DURABILITY-HOUSE  8
 u
 o.
                                IUY-NOV
                           HUE (1987-1989}

                       C  HAS     	A
                                       HO V-MAR
KA3. ..'UH
o
o.
o
B
7




6





a





4-




s -
      1 -
                    DURABILITY-HOUSE  10
                     c—
     OCT-rt3
                  m-iur
                               MAY-MOV
                                       HOV-KAR.
                           TIME (t9B7-19B«3

                      •0  BAS    —ft 1-nMR
KAft-AJM
   Figure 2  Radon levels in the  interior spaces of Houses 8 and 10.

-------
    1500
H2,Crci
       II
                                                                    H3.C«Tr
                                                                    H3,Pum?
                                                                    H4,Purr,p
                                                                    HS.Grob
                                                                    H5.Pump
Figure 3   Radon levels  In the SSD system exhausts of Houses 2 through  5.

-------
                                 TABLE 1
                 RADON' CONCENTRATIONS  IN EIOHT HOUSES
                   Occ '87
                  11 Feb '88
11 Feb
20 Kay
'88
'88
23 Kay
16 Kov
'88
'88
16 Kov '£=
30 Kar '65
HOUSE LOCATION
33 Mar '£9
25 Juni '£9
2


3


4


5


6


7


8

10

Basesenc
Base=enC
Dining Rooa
Basemen:
Basesenc
Living Re OB
Base=enc
Living Rooa
Breezevay
Basesene
Kiddle BR
Crawl Space
Base=en:
Basesene
Living Rooa
Bas eseat
Crawl Space
Living Reea
Basesenc
Living ROOD
Basesenc
Living Rooa
2.1
2.6

6.S
9.9
4.8
3.1/3.0
2.8
•
11.6
8.4
6.8
4.8
5.1
2.6
1.0
0.8/0.6
0.6
5.6
3.6
2.4
1.9
1.91
1.3

1.1
1.2
0.6
2.3/2.6
3.1
-
0.7
0.8
2.8
1.9
2.5
1.6
0.3
0.6/0.3
0.3
1.9
1.7
2.2
1.9
1.4
1.0
0.6
7.6
8.4
4.7
2.8
2.7
•
9.8
6.0
2.2
1.7
2.3
1.3
0.5
0.2/0.4
0.3
0.9
1.3
1.8
1.7
1.8
2.0
1.5
53.2
40.6
27.0
2.6
1.8
2.1
12.9
12.4
-
2.7
2.2
1.8
1.2
1.1/0.8
0.3
3.6
4.6
2.6
2.7
1.3
1.1
0.4
2.5
3.6
0.8
•>
1.0
1.3
0.4
0.6
-
3.1
2.6
1.3
0.5
0.3
0.3
2.2
0.9
1.5
1.3
                               TABLE 2

          COMPARISONS 07 RADON QUANTITIES EXHAUSTED BY MITIGATION
             SYSTEMS AND BY NATURAL  MEANS BASED ON FIVE HOUSES
Ho. IMIMIHJ vp*uir* Dumtnt | Uptuln
I
J
'
S
7
II
170
19
to
33
1}
TO
Si
IS
It
219
».
Ill
JT1
119
»t
161
•99
39*
391
i.i/si bft.utcKi/u issfa fgjsr"? 'aw "«/.:
19B 1S«
131 9«t
«3 M
135 <1S
IOJ SOI
J.JO
I.6S
I.M
«.SS
».»]
1S.TJ1.OM
TJ, 1*7, 000
10,901.000
ST,7t7,000
JJ.SBT.Oflfl
«, 971,000
n.sas.ooo
10.1TI.000
l,J»3,ooo
1. 180.000
t.Ii
l.ll
1.01
9.01
1.10

-------
                      TABLE 3

                DURABILITY RADON  SAMPLES


House       Date     Month     Grab       Pump
                                                   H2,Grab
                                                   H2,Pump

                                                   H3,Grab
                                                   H3.Pump
                                                   H4,Grap
                                                   H4,Pump

                                                   HS.Grab
                                                   H5,Pump
                                                   H6,Grab
                                                   H7,Grab
                                                   H7,Pump
                                                   H8,Grab
                                                   H8,Pump
  10
                                                   HlO.Grab
                                                   HLO,Pump
11/88
3/89
6/89
11/88
3/89
6/89
11/89
11/88
3/89
6/89
11/88
3/89
6/89
11/89
11/88
3/89
6/89
11/88
3/89
6/89
11/89
11/88
3/89
6/89
11/89
11/88
3/89
6/89
11
15
18
11
15
18
23
11
15
18
11
15
18
23
11
15
18
11
15
18
23
11
15
18
23
11
15
18
146
65
224
808
810
601
299
28
50
84
360
70
99
136
N/A
N/A
458
382
116
179
342
N/A
N/A
485
978
N/A
542
305
163
65
268
1086
864
621
133
60
61
119
510
75
134
170
N/A
N/A
300
625
166
244
384
N/A
N/A
566
1050
N/A
614
360

-------
                                                                  D-VII-4
                     RADON ABATEMENT SYSTEM ANCILLARY ITEM
          by:   Ronald F. Simon, President
               RFSimon Company, Inc.
               RD#2 Box 481
               Barto, PA  19504
                                    ABSTRACT

     The task of the radon researcher was to develop approaches to reduce
indoor radon concentrations.  As techniques were developed and tested, the
mitigation industry developed around these approaches.  In most cases, the
research techniques were applied directly to the private sector.  As the
number of installations grew and the technology was understood, the focus
changed from installation methods to ancillary items.  These ancillary
items pertain to system installation techniques as they pertain to standing
building and fire codes.  This paper deals with maintaining fire protection in
a soil depressurization system.

-------
                     RADON ABATEMENT SYSTEM ANCILLARY ITEM
     The primary task of the early radon professional was to identify methods
to reduce the indoor radon concentrations. In some cases, he was not sure that
reductions could be made to the accepted 4 pCi/1 level.  The task at hand was
to develop mitigation strategies that would control indoor radon concentrations.
Through research programs and field installations a variety of techniques
became known and generally accepted by an emerging mitigation industry.  Soil
pressure manipulation soon became the work horse of the industry.  In particular,
soil depressurization became the system of choice.  As more was learned about
soil depressurization, installation techniques and material system choices were
modified in order to align with the level of effort required.  The Intent of
this alignment was to make system installation less labor intense while still
controlling indoor radon.

     Mitigation system installation standards have eluded the industry as a
whole.  One of the contributing factors allowing this to take place is that the
industry has sprinted ahead with little regard for standardized installation
practices, unaware of possible unintended consequences.  Typically the mitigation
installation focuses on a high profit/low cost installation.  Standards also are
difficult to initiate and maintain possibly because of the installers limited
capabilities based on background, knowledge and skills.  EFA's recommendation
for above eave exhaust location is a basic practice accepted in hazardous
material discharge, yet today, some installers still do not follow this guideline.

     Within the frame work of the mitigation system evolution, little attention
has been placed on the appropriateness of the system with respect to the applicable
building codes.  This paper is written to address the code issue and in particular
the fire barrier penetrations in both residential and commercial structures with
respect to radon soil pressure manipulation systems.  Emerging EPA guidelines
recommend the installation of soil depressurization systems utilizing above
eave exhaust locations.  To accomplish this, pipe systems typically penetrate
in place fire barriers within the structure both floor to floor in commercial
construction and garage to occupied space in residential construction.   When
this takes place, the occupied quarters are placed at risk.  The potential for
loss of life because of an installed abatement system and less important, the
liability issue, directed RFSimon Co., Inc., to identify installation techniques
and practices that would eliminate this risk.  In 1988, 4,955 lives were lost due
to residential fires.

-------
     The BOCA (Building Officials and Code Administrators) for commercial
and the CABO (Council of American Building Officials) for residential construction
addressed this issue with the following quote from the BOCA National Fire
Prevention Code: "The BOCA National Building Code was initially prepared and
has been maintained on the premise that all matters pertaining to the construction
of the building and built into it, either in its initial construction or through
subsequent alterations, repair or extension, should be covered by the building
code.  This includes fire protection of the building elements as well as fire
separation walls or other precautions required for protection against specific
hazards of the particular use of the building."

     F-105.1 General:  Whenever the code official or the code official's designated
representative shall find in any structure or upon any premises dangerous or
hazardous conditions or materials as follows, the code official shall order such
dangerous conditions or materials to be removed or remedied in accordance with
the provisions of this code:

     1.  Dangerous conditions which are liable to cause or contribute to the
         spread of fire in or on said premises, building or structure or endanger
         the occupants thereof.

     2.  Hazardous conditions arising from defective or improperly used or
         Installed electrical wiring, equipment or appliances.

     To understand how a typical soil depressurization system would perform in
a fire situation, a full size model was fabricated to simulate a typical Radon
Vent Duct (RVD) routing.  The model consisted of a floor deck utilizing 2x10 nom.
joist and collar or band joist, 5/8" plywood decking and a 2x4 nom. wall system
with wall insulation.  The wall system and band joist were covered with fire
code dry wall.  This model represents a typical wall/floor construction detail
in houses with attached garages.  The building code  typically requires the
garage floor be 8" min. below the house floor in order to contain gasoline
engine fume emissions.  However, because of this required step, the garage/house
common wall represents a direct exit from the basement to the garage area.
Typical RVD systems are routed horizontally from the basement to the garage
area, then vertically to the attic and terminated outside of the building shell.
As part of the model, a 4" PVC pipe system was installed as described above.
See Sketch SK-1.  A Fan-Tech Model R-150 fan was installed approximately 10'
above the horizontal pipe run.  The pipe was insulated using a polyethylene
covered fiberglass duct insulation.  The effluent end was covered with a
SIMON CAP while the inlet end was dampered to simulate expected operating
static pressures within the system.

     A fire was set at the base of the vertical stack just at the transitional
90° el through the dry wall/band joist area.  The model was built to confirm
what was thought would take place.  The thought was  that  the PVC pipe would
melt open.  The open pipe would draw the fire and smoke up the pipe as a result
of the force exerted by the fan thus spreading the fire and smoke to other
areas of the structure.

-------
The damper box is secured to the structure according to the manufacturer's
recommendations.  The box is fastened so that it will remain in position and
function during a fire.  Within the damper box rests a curtain type blade
closure mechanism.  A fusible link holds the blade closure mechanism in the
open position.  The blade closure mechanism is activated when the fusible
link is destroyed through exposure to elevated temperatures.  Typically
the link will soften and release at 160° F.  Vertical mount units depend
on gravitational force to close the curtain within the housing.  Horizontal
mount units are fitted with a closing spring that draws the blades across the
opening.  One concern when incorporating these units in the Radon Vent Duct
or RVD is the potential build up and continuing exposure to condensation.
These units are fabricated utilizing galvanized metal enclosure and curtain
dampers and the spring in the horizontal unit is stainless steel.  However,
long term exposure to typical system conditions could perhaps render the
closing mechanism ineffective due to corrosion build-up, particularly in
the horizontal mount units.

     Some structural work may be necessary to meet manufacuturer installation
requirements.  The opening for the damper housing must be larger than the
assembly, typically 1/8" per ft. of damper housing plus 1" for the maximum
and a minimum of 1/4".  These clearances are critical to allow for housing
expansion and free curtain damper mobility in elevated temperatures.
Mechanical fastening of the unit to the structure requires a metal sleeve
and angle supports mechanically fastened to the damper housing.  These
supports are positioned on both sides of the penetration captivating the
structure while allowing the damper housing to expand freely.  Sealing
between the RVD and the extended collar should be completed to maintain a
closed RVD system.  The cost of these units vary depending on size and mounting
configuration.  A 4" extended collar, horizontal mount unit cost is $40 and
the vertical mount equivalent is slightly more.  Ancillary materials, i.e.,
sleeves, mounting angle, screws and caulk will be in the $10-$12 range.
Structural preparation will vary from site to site.  The installation techniques
required for proper installation add significantly to the installed cost.
This type of installation would be utilized more in commercial RVD systems,
particularly in through floor to floor applications.  The question of durability
and function will, however, require further study.

     Another approach that was evaluated is the use of an intumescent fire stop
system.  This technique incorporates some of the elements required for the
damper systems identified above.  The installation requires 5 key elements
consisting of the intumescent wrap, foil faced tape, galvanized retaining
collar, adjustable stainless steel band and intumescent fire caulk.  The
installation is completed at the fire wall in the unprotected area.  The
procedure requires the application of the foil faced tape over the FVC pipe
system to protect the pipe from softening due to volatiles within the
intumescent.  The preformed strips of intumescent are 2" x 20" x 1/4" thick
and are cut to length to fit snuggly around the RVD at the fire wall penetration.
Tape is used to temporarily secure the joint and hold the strip in place.
Additional layers are installed with the joints staggered.  The number of
intumescent wraps depends on the pipe diameter; a 4" diameter RVD will require
3 bands of material.

-------
When the system was ignited the polyethylene vapor barrier was consumed
immediately in the area of the flame and heat.  The duct insulation delayed
the impact of the flame on the PVC pipe because of its inherent properties.
Gradually the pipe was exposed to increased heat as indicated by the
increasing density of effluent smoke.  As the flames and heat exposure to
the PVC pipe and elbow increased, the pipe system softened.  The softening
of the pipe continued until the structural integrity of the pipe system was
below the level required to remain open.  It began to collapse inward
responding to the system induced negative pressure.

     The PVC pipe continued to close in on itself until it was completely
closed, causing the fan side of the system to be isolated from the fire source
and from further damage.  The system was heat softened and sealed, subsequently,
operating under full negative pressure capacity of the fan.  This was evidenced
by the gradual and eventual complete termination of smoke from the stack.  This
event was in complete disagreement with anticipated results.

     At this point, however, the flames had an open access through the in place
fire wall and into the basement.  The PVC pipe system burned and created a
pathway for the fire to follow.  The flames spread along the RVD into the
occupied side of the fire wall.  Within minutes the band joist and plywood
sub-floor were ignited.  Eventually, the model was consumed in flames.

     Armed with this information the task ahead was to find an installation
process that would eliminate this from taking place.  Several approaches were
investigated that would allow a mitigation system to overcome this potentially
lethal installation flaw.

     An obvious approach would be to completely enclose the pipe system in a
fire code dry wall chase.  There is, however, concern that this type of remedy
is not acceptable in that a chase is created.  The chase would still have the
ability to transfer flames within itself and perhaps require special sealing
at the pipe penetrations within the chase.  From an installation standpoint
this approach is labor intense and requires structural materials to complete.
Also in some installations this approach would be impractical or impossible
to implement.

     Another approach would be to use a non-combustible pipe system through
the non-fire protected areas, i.e., garage.  This could be accomplished using
a copper Drain, Waste, Vent or DWV pipe.  The pipe would be connected to the
standard PVC RVD system using flexible couplings to make the transition to
PVC.  A special intumescent sealant must be applied to stop flame spread
through the fire wall/pipe penetration.  This will serve to inhibit flame
spread around the pipe through an irregular opening.  Although this technique
is feasible, the expense of the copper pipe, its weight and pipe support
implications would make a significant contribution to the installation cost.

     A mitigation system can be isolated from the occupied spaces utilizing
an in-line fire damper.  This device is a sealed metal box with 4" diameter
extended collars at opposite ends.

-------
When this is completed, a pre-stamped galvanized metal retaining collar is
cut to length and folded to capture the intumescent and form attachment ears.
The retaining collar is secured around the intumescent with a stainless
steel adjustable band.  The unit is then secured with screws through the
attachment ears to the structure.  The final step is to seal the area
between the fire inhibiting wall or ceiling and the retaining collar.  This
requires an intumescent caulk material.  The intumescent material and caulk
system is designed to expand when exposed to elevated temperatures.  This
expansion is directed toward the PVC RVD pipe system reacting against the
retaining collar and the stainless steel retainer band.  The PVC RVD is
softened as the temperature increases facilitating its forced closure.  The
intumescent forms a hard char material which, by design, remains intact
throughout the flame exposure.  The materials cost for this type of
installation is $45.  The number of installations will vary depending on
the number of fire barrier penetrations.

     The issue of risk prevention is an important element within the frame
work of the mitigation installation.  The approaches presented here have the
ability to significantly reduce the potential loss of life due to system
interaction with fire.  The fire protection system chosen, as with many other
mitigation options, is based on the site specific influences.  In some cases
the damper system may be appropriate while the intumescent system may be
applicable in others.  The cost associated with the installation is minimal
when compared with the potential for loss of life and property.  Professional
design and mitigation companies should develop skills in the installation of
fire protection systems recognizing this is not an option but a mandatory
ingredient for code compliance.

      The work described in this paper was not funded by the U. S.
Environmental Protection Agency and therefore the contents do not necessarily
reflect the views of the Agency and no official endorsement should be
inferred.

-------
                   5K-1
GARAGE
SSV  PIPING
CONCRETE SLAB
                         STUD WALL
FIRE CODE DRY WALL
  SUB-SLAB AREA
                                 BASEMENT

-------
                                         Paper D-VII-5

          LABORATORY STUDIES OF "BETWEEN THE ROOMS"
             RADON DECAY PRODUCT REMOVAL UNITS	

               by:  Dade W. Moeller, and
                    Xiaowei Yan
                    Harvard School of Public Health
                    Boston, MA 02115
                          ABSTRACT

     A temporary wall was constructed to divide a large
laboratory chamber into two sections.  Mounted in the wall
were two fan-positive ion generator units, each blowing in
opposite directions so as to maintain a balance in the
airflow between the two sections.  Data showed that the
units provided PAEC reductions in the two sections of the
chamber in excess of 90%.  Estimates of the overall
reductions in the lung doses were 86% (using the Jacobi-
Eisfeld model), 88% (using the Harley-Pasternak model), and
92% (using the James-Birchall model).
INTRODUCTION

     A temporary, 4-m wide wall was constructed in the
Harvard Radon Chamber (1) to simulate two adjacent rooms in
a residence.  The left section of the chamber had a volume
of 35.3 m3 and the right section had a volume of 43.2 m3.
Two radon decay product  (RDP) removal units (fan and
positive ion generators) were installed 1.5 m apart in this
wall at a height of 2.0 m above the floor (shown
schematically in Figure 1).  One fan directed an airflow of
approximately 50 L/s from one section of the chamber into
the other.  The other fan directed an equal airflow in the
opposite direction so that pressures in the two sections
were equal.  Positive air ions were dispersed into the
respective airflows.

LABORATORY TESTS AND RESULTS

     The sequence of tests consisted of an initial time
period when the RDP removal unit ion generators were off
(only the fans were operating) and a second time period when
both the fans and ion generators were operating, with
measurements being made in one section of the chamber.  This

-------
sequence was then repeated with measurements being made in
the other section of the chamber.  Radon concentrations in
the laboratory chamber, as well as graphs of the potential
alpha energy concentrations with the fans on, and with the
fans and ion generators on, are shown in Figure 2.  As may
be noted, the radon concentrations (upper curve) were
essentially constant throughout the studies, whereas the
fans plus ion generators reduced the potential alpha energy
concentrations by greater than 85%.  Due to limitations on
the availability of monitoring equipment, the PAECs were
monitored in the left section of the chamber on the first
two days, and in the right section on the subsequent two
days.
 Figure 1.  Schematic representation of "between the rooms"
            fan-ion generator airborne radon decay product
            removal units.

-------
              500

           S
           M
           2  400


           II

           If

           £l 200
           ^s
           M

           £  100
     Left Section
     of Chanber
-•j-Fans Onn
        Right Section
        of Chaaber
 Fans + Ion
Generators On
                                  SO
                                  10
                               Tim (DAYS)
Figure 2. Radon and potential  alpha energy concentrations in
          left and right sections of the laboratory chamber
          with either the fans or the fan-ion generator
          units in operation.   Only the fans were in
          operation on the  first and fourth days;  both the
          fans and the ion  generators were in operation
          during the second and third days.

-------
     Grab samples of radon and its decay products were
analyzed by standard methods (1).  Single-screen samples
were analyzed for unattached decay products (2).  These data
were used to calculate reductions in total PAECs, and
unattached PAECs which are summarized in Tables 1 and 2.
Reductions in total PAECs averaged 87% and reductions in
unattached PAECs averaged 56%.  Estimates of the reductions
in lung dose to the basal cells of the tracheobronchial
region are summarized in Tables 3, 4 and 5.  The percentage
reductions given in Tables 1-5 are based on the use of the
two "between rooms" RDP removal units (i.e., fans and ion
generators) as compared to the use of the identical devices
with the ion generator turned off (i.e., only the fans
operating).  If the reductions had been compared to "no
treatment conditions"  (i.e., both the fans and  ion
generators off), the calculated percentage reductions would
have been greater.  The range of probable reductions in the
lung doses attributable to the fans and ion generators,
based on linear forms of three dosimetric lung  moels  (3,4),
and their corresponding limits are shown below.  Estimates
of the reductions, as compared to "no treatment conditions,"
were calculated on the assumption that the fans, alone,
would have reduced both the total and unattached PAECs by
50%  (5).

     Dosimetric Lung Models          Percentage  Percentage
     and Model Limits	          Reduction    Remaining

     Minimum  (i.e., effect on
           unattached PAEC)              79%         21%

     Lung  Dose
           Jacobi-Eisfeld Model  (6)      86%         14%
           Harley-Pasternak Model  (7)    88%         12%

           James-Birchall Model  (8)      92%           8%
     Maximum  Effect  (i.e., effect on    93%           7%
           total  PAEC)


     The work described  in this  paper was  not funded  by the
U.S. Environmental  Protection Agency and  therefore  the
contents do not  necessarily  reflect the views of the  Agency
and no official  endorsement  should  be  inferred.

Acknowledgments:

     Assistance  in  the collection of  the  data reported in
these  studies was provided by Captain  Carl A. Curling, U.S.
Army,  formerly a Doctoral  Student,  Department of
 Environmental Science and Physiology,  Harvard School  of
 Public Health.

-------
 TABLE  1.   EFFECTIVENESS OF FAN-ION GENERATORS IN REDUCING TOTAL  PAEC*

Experimen
Number
1
2
3
Average
TABLE 2.


Experimen
Number
1
2
3
Average
Left Section
PAEC (mWL)
of Chamber

t Fan Fan + Ion Percent
On Generator
180 19.1
219 26.8
201 22.6
200 23
EFFECTIVENESS OF FAN
Left Section
PAEC (mWL)
t Fan Fan + Ion
On Generator
8.8 3.6
10.8 4.5
7.3 4.6
9.0 4.2
On Reduction
90
88
89
89
-ION GENERATORS
of Chamber

Percent
On Reduction
59
58
37
51


Fan
On
191
237
201
210
Right Section of
PAEC (mWLJ
Fan + Ion
Generator On
22.5
30.5
27.7
26.9
IN REDUCING UNATTACHED


Fan
On
12.4
13.5
9.4
11.8
Right Section of
PAEC (mWL)
Fan + Ion
Generator On
4.0
5.2
5.6
4.9
Chamber
Percent
Reduction
88
87
86
87
PAEC*
Chamber

Percent
Reduction
68
61
40
56
*External air exchange rate was 0.2/h.

-------
TABLE 3.  EFFECTIVENESS  OF  FAN-ION  GENERATORS  IN  REDUCING MEAN BRONCHIAL LUNG
                   DOSE  EQUIVALENT  RATE  (James-Birchal1  Model)(8)

E




xperiment
Number
1
2
3
Average

1
(

1
1
1
1

Left Section
of
Chamber
reatment Conditions
Dose Rates in Sv/y)
Fan
On
.24
.50
.19
.31
Fan + Ion
Generator On
0.319
0.409
0.402
0.377




Percent
Reduction
75
73
67
72

1
(

1
1
1
1
Ri
ght Section of
Chamber

reatment Conditions
Dose Rates in Sv/y)
Fan
On
.53
.76
.35
.55
Fan + Ion
Generator On
0.360
0.470
0.489
0.439
Percen
Reducti
77
73
64
71
t
on




Notes :
1
2
The external
Percent
air exchange
Dose Rate (F
rate was 0.2/h.
an) - Dose Rate
(F
an +
Ion Generator)
( innm .

 3.
Reduction
James-Birchal1  Model
   Dose Rate (Fan)
Conversion Factor:
92 + 180 fn (mSv/WLM).

-------
TABLE 4.  EFFECTIVENESS OF FAN-ION GENERATORS IN REDUCING MEAN BRONCHIAL  LUNG
                   DOSE EQUIVALENT RATE (Har1ey-Pasternak Model)(7)
Left Section of
Chamber
Treatment Conditions

Experiment
Number
1
2
3
Average
(Dose
Fan
On
1.01
1.23
1.02
1.09
Rates in Sv/y)
Fan + Ion
Generator On
0.216
0.281
0.270
0.256

Percent
Reduction
79
77
74
77
R
ight Section of
Chamber
Treatment Conditions
(Dose
Fan
On
1.20
1.40
1.11
1.24
Rates in Sv/y)
Fan + Ion
Generator On
0.245
0.322
0.329
0.299

Percent
Reduction
80
77
70
76
Notes:
1.  The external  air exchange rate was 0.2/h.
2.  Percent   =    Dose Rate (Fan) -  Dose Rate (Fan  + Ion Generator)
    Reduction                 Dose Rate (Fan)
3.  Harley-Pasternak Model  Conversion Factor:   70  +  795 f  (mSv/WLM).
(100%)

-------
TABLE 5.  EFFECTIVENESS OF  FAN-ION GENERATORS  IN  REDUCING MEAN BRONCHIAL LUNG
                   DOSE EQUIVALENT RATE  (Jacobi-Eisfeld Model)(6)


Left Section
Treatment Conditi
(Dose Rates in Sv
Experiment
Number
1
2
3
Average
Fan
On
0.935
1.14
1.02
1.03
Notes:
1. The external
2. Percent
of Chamber
ons
A/1
Fan + Ion Percent
Generator On Reduction
0.124
0.169
0.150
0.148
air exchange
Dose Rate (
87
85
85
86
rate was 0.2/h.
Fan) - Dose Rate
R
iqht Section of
Treatment Conditions
(Dose Rates in Sv/y)
Fan
On
1.02
1.25
1.04
1.10
(Fan
Fan + Ion
Generator On
0.144
0.193
0.183
0.173
+ Ion Generator
Chamber

Percent
Reduction
86
85
82
84
.1 (1005M


3.
Reduction                Dose  Rate  (Fan)
Jacobi-Eisfeld Model  Conversion  Factor:  66 + 1370
f  (mSv/WLM).

-------
REFERENCES

1.   Rudnick, S. N., Hinds, W. C., Maher, E. F., and First,
M.W., "Effect of plateout, air motion, and dust removal on
radon decay product concentration in a simulated residence,"
Health Physics. 45: 463, 1983.

2.   George, Andreas C., "Measurement of the uncombined
fraction of radon daughters with wire screens," Health
Physics. 23: 390, 1972.

3.   Maher, E. F., Rudnick, S. N.. and Moeller, D. W.,
"Effective removal of airborne 222Rn decay products inside
buildings," Health Physics. 51: 356, 1987.

4.   Maher, Edward F., "The control and characterization of
radon decay products in residences," Doctoral Thesis, School
of Public Health, Harvard University, Boston, MA, 1985.

5.   Rudnick, S. N. and Maher, E. F., "Surface deposition of
222Rn decay products with and without enhanced air motion,"
Health Physics. 51: 283, 1986.

6.   Jacob! W. and Eisfeld K., "Dose to tissues and
effective dose equivalent by inhalation of radon-222, radon
220, and their short-lived daughters", Report Gesellschaft
fur Strahlen- und Umweltforachung GSF-S0626, Munich,
Germany, 1980.

7.   Harley, N. H. and Pasternack, B. S., 1982, "Environ-
mental radon daughter alpha dose factors in a five-lobed
human lung", Health Physics. 42: 789, 1982.

8.   James, A. C. "Dosimetric approaches to risk assessment
for indoor exposure to radon daughters", Radiation
Protection Dosimetry. 7: 353, 1984.

-------
                                                                              D-V1I-6
               RADON MITIGATION TECHNIQUES FOR NORWEGIAN HOUSES
                            l            i                  2
                   Bjom Lind , Terje Strand  and J0m T.Brunsell
                1 National Institute of Radiation Hygiene, Osteras, Norway
                2 Norwegian Building Research Institute, Oslo, Norway
                              ABSTRACT
      Results of an experimental study of different techniques to reduce the radon levels in
Norwegian houses are reported. A total sample of 30 detached houses, representing a broad range
of different types of Norwegian houses accordring to construction techniques and architecture, were
selected for the study. In this sample of houses, earlier measurements had shown average radon
concentrations ranging from about 500 Bq/m> to 15 000 Bo/m>. Passive measurements of radon
were performed at different stages in the process. In a few houses more extensive measurements
were carried out This included contineous measurements of radon, radon daughters and air
exchange rate, and grab sampling measurements of unattached fraction of radon daughters. Of the
different solutions studied, the methods based on sub-slab ventilation/depressurisation were found
to have the best reduction effect on the radon level in indoor air.
                              INTRODUCTION

      In 1988, an extensive study on mitigation techniques for Norwegian houses was started. The
study was organised as a collaboration between the Norwegian Building Research Institute and the
National Institute of Radiation Hygiene. The aim of the study was to evaluate between different
possible mitigation methods/techniques to reduce the radon level for different types of Norwegian
houses and to give recommendations.

      So far, most of the research work on mitigation methods/techniques in the Nordic countries
have been carried out in Sweden (1,2). Most of the recommendations in Norway have been based
on experience from there. However, the buidling stocks are quite different The most important
difference, as far as radon is concerned, is the fact that most detached houses in Sweden do not
have a basement, while probably more than 80% of detached houses in Norway do have.

-------
                           MATERIAL AND METHODS

SELECTION OF HOUSES
     In Norway there is about 1.6 million dwellings. According to census data from the Central
Bureau of Statistics (3), about 85% of the Norwegian housing stock is detached and undetached
houses. In the most rural municipalities the percentage is nearly 100%. However, in Oslo, the
largest town in Norway, about 65% of the dwellings are  in blocks of flats. Influx of radon from the
ground is found to be the most important source of indoor radon in detached and undetached
houses. In houses where the radon concentration is found to be considerably higher than normal,
the contribution from other sources can usually be neglected. For dwellings in blocks, on the other
hand, building materials are usually the main source of indoor radon. However, in most cases the
radon concentration is very low compared to detached and undetached houses. Therefore, blocks of
flats were not included in the experimental sample of our study.

     Most of the houses were selected from earlier and recent measurements from the National
Institute of Radiation Hygiene. It was important to get a, as far as possible, representative sample
of typical houses according to construction technique, architecture, age of the house and geology. It
is assumed that the final sample  of 30 houses meets these requirements fairly good. In table 1, the
houses is classified into three categories according to type of house and radon concentration. As
shown, most of the houses have basements.
    TABLE 1. THE SAMPLE OF DETACHED HOUSES.
                             NUMBER OF HOUSES
                                                   1)
                        Radon concentration (Bq/m»)
                       < 1000   1000-5000    > 5000
     With a basement
     floor below           14        11
     ground level

     Concrete slab
     directly on ground      2

     On crawl space                    2
    ^Representative averages of radon concentration for each house.


THE MEASUREMENTS
     Passive measurements of radon were performed at different stages in he mitigation process in
all of the 30 houses. The ETB-method, which is a combination of activated charcoal and TLD (4),
were used in the measurements. The integration time in the measurements was between  7 and 10
days. In order to minimize the influence on the measurements from short term variations in the
radon level, it was essensial that the they were performed at about equal meteorological/-
climatical conditions. The uncertainties in the individual measurements were estimated to be about
15% at the 95% confidence level.

-------
     In a few selected houses, more extensive measurement programs were performed. This
included contineous measurements of radon and radon daughters and measurements of the air
exchange rate. At the diagnostic stage in the process, grab sampling measurements of radon were
found to be especially valuable in characterising the source. The method and equipment has been
described in an earlier paper by Stranden and Berteig (5).
                              RESULTS AND DISCUSSION
     Generally, mitigation techniques may be divided into four groups:

          (1) removal of radon sources
          (2) methods to reduce the radon influx from the ground
          (3) methods to increase the air exchange rate
          (4) removal of radon daughters by electric fields/electrostatic
             filter

     Methods involving a removal or exchange of source material from the building site and/or
near the house may have a considerable reduction effect on the radon level in future houses if it is
done at the site before the buidling work  is  started. For existing houses this kind of solution will
usually be too expensive and too disruptive. However, for houses with a basement floor below
ground level, where alum shale or other materials with  a very high exhalation rate of radon has
been used as filler on the outside of the foundation wall, removal or exchange of the source may
be a successful solution.

     If the air exchange rate is at a normal level, the relative reduction effect on  the indoor level
of radon by increasing the air exchange rate will  usually be very low. Energy saving aspects,
climatic conditions and indoor comfort limits the  possibility to increase the air exchange rate in
most situations. Due to an increase in the energy costs,  and the campaign for saving energy in the
70*s and 80"s, a large proportion of old houses have been retrofitted in the last decade. At the same
time period, modem houses have been made energy efficient. According to studies in Sweden  (6),
this may have reduced the air exchange rate in a significant part of the housing stock. In tight
houses with no mechanical ventilation systems, average  air exhange rates down to 1/10 of the
normal level have been reported. In such  houses,  the effect of an increase in the air exchange rate
to normal values, may lead to a reduction in the  radon  concentration by an order  of magnitude. It
is, however, important that the ventilation systems do not increase the underpressure inside. If so,
then the influx of radon from the ground  may increase.

     For most houses in our study, diagnostic measurements showed  that influx of radon from the
ground was the most important source of radon.  In 28 out of the 30 houses, different methods to
reduce the influx from the ground were chosen. One of the main objectives of our study was to
quantify the effect of different low  cost solutions. This of course, did have an impact on the
individual choice of solution. In several houses the mitigation process had to be carried out in
different steps, starting with  the cheapest possible one. In table 2, a general view of the different
solutions is shown. In the following, a few examples from our study will be discussed in more
detail.

-------
    TABLE 2. CLASSIFICATION OF THE DIFFERENT MITIGATION TECHNIQUES
            IN OUR STUDY
     Category of mitigation solution       Number of houses
     Sealing of cracks and openings
     in the foundation wall and the slab            3

     Changing the pressure differences
     between indoor air and the ground   1)       20

     Changing the ventilation conditions            2

     A combination  of
     different solutions 2)                        5
    1) Including four different methods: pressurize basement and/or crawl
      space (5 houss) sub-slab ventilation (13 houses), outdoor well (2 houses)
      and wall ventilation (1 house).
    2) Combinations of sealing and sub-slab ventilation.
     In figure 1, the reduction effect of a step-wise solution is illustrated in a house with a
basement. The first step was based simply on sealing visable cracks and openings in the concrete
slab and the foundation wall. In one of the rooms in the cellar, the average radon level was
reduced from about 4,000 Bq/m' to 800 Bq/mV In this room a higly permeable wall of light
expanded clay aggregate blocks, with direct contact below the slab, were assumed to be an
important entry route. Most of the reduction effect can probably be explained by the plastering of
this wall.  However, the radon level was still too high and further actions were found necessary. As
a second step, a sub-slab ventilation system (7) was  installed. Due to the small base area of the
house (about 80 m3) and considerations about the permeability of the ground, one central exhaust
pipe was  assumed to give a sufficient suction. The pipe was connected to a fan on the loft. The
flow rate was measured to 50-60 mVh. These type of solutions reduces the influx of soil gas by
reducing the pressure differential between the basement and the ground and by reducing the
concentration of radon below the slab by  ventilation. After this second step, the radon
concentration was reduced to about 150 Bo/m» in the basement and to about 100 Bq/m* on the
first floor. On the average, in the four rooms where passive measurements were performed, the
radon concentration was reduced by more than 90%. The actions were concluded to be very
successful.

-------
         Bq/m
           4000 -i
           3000
/I
^| Before action
| ';. | After action  step 1
•1 After action  step 2
           2000-
           1000
                  Room 1       Room 1      Room 2      Room 3

                FIRST FLOOR           BASEMENT

Figure 1. A two-step solution in a house with basement. Step 1: Sealing of cracks and openings in
        the concrete slab and the foundation wall. Step 2: Sub-slab ventilation (one exhaust pipe).
     Figure 2 illustrates the effect of a pure solution based on sub-slab ventilation/-
depressurisation. In this example, two exhaust pipes were connected to the fan on the loft The fan
is identical to the one in the previous case. For one of the rooms in the basement, the radon level
was reduced from about 2,700 Bq/m3 to below 100 Bq/m3. This is a reduction of nearly 97%. The
average radon concentration in the main living room on the first floor was reduced from
840 Bq/m' to 60 Bq/m3. The solution was concluded to be very successful.


            Bq/m
            3000
            2000
            1000-
                    Room 1

                 FIRST  FLOOR
                 Room 1
 Room 2

BASEMENT
Room 3
Figure 2. The effect of sub-slab ventilation/depressurisation (two exhaust pipes)
         in a house with basement.

-------
     It is important that the capacity of the fan is as low as possible. Not only because of energy
saving, but also due to problems in the cold winter months if large quantities of cold air are
beeing sucked beneath the basement floor. Outdoor temperatures below -30^2 in the coldest winter
months are not unusual  in several parts of Norway. Our results have shown that fans with a very
low capacity (about 30 Watt/60 mVh at Ap about 100 Pa) may be sufficient to solve the radon
problem in many situations.

     In one of the houses, the radon level was found to be very high. From contineous
measurements, short time concentrations of radon up to 190 000 Bq/m3 were found in the
basement. Grab sampling measurements  of radon and radon daughters confirmed these results.
From passive measurements, the radon level in the basement was found to be between 10 000 and
15 000 Bq/m». Grab sampling measurements showed that an old and dry floor drain (without a
water trap) in the basement floor probably was one of the main entry points. This floor drain was
directly connected to alum shale rich soil. Since the floor drain was not in use any more, the first
step was then to glog it  and try to seal all visable cracks and openings in the concrete slab and the
foundation wall. The results of this first step is illustrated in figure 3. In one of the rooms in the
basement the radon level was reduced by almost 80%. In the other room in basement, sealing of
the floor was found very difficult due to  presence of  numerios cracks and openings in the slab. In
this room the average radon concentration was found to increase by about 15%. In addition to
further sealing of the floor and the walls, an almost identical sub-slab ventilation/-
depressurisation system as the example in figure 2, was installed. As illustrated, the radon
concentration was reduced by nearly 95% in the basement and by about 90% on the first floor.
However, the radon level is still too high and further work is needed. One of the possibilities is to
increase the flow rate capacity of the fan.
           Bq/m
            20000 A
            16000
            12000
             8000
             4000
Before action
After action step 1
After action step 2
                     Room 1         Room 1         Room 2

                 FIRST  FLOOR        BASEMENT
Figure 3. A two-step solution in a house with a basement. Step 1: glogging of a floor drain
         in the basement floor and sealing of different cracks and openings in  the concrete
         slab and the foundation wall.  Step 2: Sub-slab ventilation (two pipes).

-------
     The effect of sub-slab ventilation in the other houses where this type of technique as studied,
was generally found to be successful. However, in a couple of houses were the ground most
probably has a very low permeability, and the sealing of different cracks and openings in the slab
were found to be very difficult, there was only a very slight reduction in the radon level.

     For two of the houses in our study, the effect of a soil gas ventilation well on the outside  of
the foundation wall, were studied. The effect for one of these houses is ullustrated in figure 4. The
well was at 3m depth and 1m in diameter. The exhaust  pipe (diameter and lenght of O.lm and 3m
respectively), which was placed in the centre of the well reaching below basement floor level, was
connected to a small fan (28 Watt) on the top. The flow rate was measured to about 50 m»/h. In
one of the rooms closest to the well in the basement, the radon concentration was reduced from
1300 Bq/ms to slightly below 350 Bq/ms. On the first floor where  the living room, bedrooms and
kitchen were located, the radon level was reduced from 420 to 240 Bq/m».
                 Bq/rrT
                  1500
                  1200-
                           Room 1
                         FIRST FLOOR
Room 1         Room 2
       BASEMENT
 Figure 4. The effect of an outdoor soil gas ventilation well in a house with a basement.
      In five houses, different techniques to pressurize the basement and/or the crawl space by
 supply of fresh air from the outside, were investigated. Due to the climatic condtions it was found
 necessary to include a heater unit in the fans, which of cource increases the energy costs.
 Preliminary results from these exsperiments seems to show that these types of solutions may have a
 relatively slight reduction effect on the radon level. In figure 5, the effect of pressurizing  the
 basement and the crawl space in a house on alum-shale ground is illustrated. The fan (99 Watt)
 was installed in the foundation wall at the basement level. The flow rate was measured to
 310 m'/h. The radon concentration in the basement and the first floor were reduced by  50% and
 30%, respectively. One of the main explanation of this low reduction effect may be leakage
 between the basement and the first floor.

-------
              Bq/nrT
                       First floor
Basement
Crawl space
 Figure 5. An example of a solution based on pressurizing the crawl space and the basement.
                                  CONCLUSIONS
From the results and discussion of our study the following conclusions may be drawn:

1. Sealing of viseable cracks and openings in the slab and the foundation wall is recommended as
   a first step in any mitigation process. However, the results of our study shows that sealing
   alone in most situations will not be sufficient to solve the entire radon problem. This is
   especially true in cases where the radon level is very high.

2. Sub-slab ventilation by exhaust pipes connected to a small fan were generally found to be the
   most successful technique in our study. In most cases it is sufficient with only one or two
   exhaust pipes and a very small fan (30 Watt/60 m'/h).

3. Preliminary results of our exprimental studies of techniques based on pressurizing crawl
   space and/or basement by supply of outdoor air have shown to have a relatively low reduction
   effect on the radon level.

The final results of our study will be presented later.

-------
                             ACKNOWLEDGEMENTS

     This work was supported by the Royal Norwegian Council of Scientific and Industrial
Research (NTNF), the Ministry of Local Government and Labour, Ministry of Petroleum and Energy
and Ministry of Health and Social Affairs in Norway. The authours wish to thank Miss A.K.Kolstad
at the National Institute of Radiation Hygiene for her assistance during the measurements.
     The work described in this paper was not funded by U.S. Environmental Protection Agency
and therefore the contents do not necessaryli reflect the views of the Agency and no official
endorsement should be inferred.
                              REFERENCES

1. National Institute fo Radiation Protection, National Board of Health and Welfare, National
  Board of Physicsl Planning and Building.  Radon in dwellings, interim report 1987. Report
  SSI 87-17, National Institute of Radiation Protection, Sweden (in Swedish).

2. Ericson, S-0, Schmied, H. and Clavensjo, B.  Modified technology in new constructions
  and cost effective remedial action in existing structures  to prevent infiltration of
  soil gas carrying radon. Radiat.Prot.Dosim. 7(1/4): 223-226, 1984.

3. Central Bureau of Statistics, Census data, Personnal Communications 1986.

4. Stranden, E., Kolstad,  A.K. and Lind, B.  The ETB dosemeter, a passive integrating
  radon dosemeter combining activated charcoal and TLD. Radiat.Prot.Dosim. 5(4): 241-245,
  1983.

5. Stranden, E. and Berteig, L.  Radon daughter equilibrium and unattached fraction in
  mine atmospheres. Health Phys. 42: 479-487, 1982.

6. Swedjemark, GA  The equilibrium factor F.  Health Phys. 45(2): 453-462, 1983.

7. Sanchez, D.C. and Henschel, D.B.  Radon reduction techniques for detached houses -
  technical guidance. U.S. Environmental Protection Agency Report EPA/625/5-86/0019,
  Center for Environmental Research Information, Cincinnati, OH, 1986.

-------
               Session D-IX:
Radon in Schools and Large Buildings—POSTERS

-------
                                                         D-IX-1
      PREDICTION OF MAXIMUM RADON CONCENTRATIONS IN SCHOOLS
                 USING PARTIAL SAMPLING METHODS

          by:  William Belanger,  P.E.
               U. S. Environmental Protection Agency, Region III
               Philadelphia, Pennsylvania 19107

               Michael Pyles
               Pennsylvania Bureau of Radiation Protection
               Harrisburg, Pennsylvania 17120
                            ABSTRACT

     The Environmental Protection Agency published interim
guidance for radon measurements in schools in April of 1989.
That guidance recommends testing of radon in all rooms that are
frequently used.  This paper describes a partial sampling
technique which might help identify schools with a potential
radon problem.  Its use is not endorsed by EPA.  It is an
analytical tool which might be used as an initial step in
facilitating implementation of the 1989 guidance.

     Radon measurements were obtained for 198 schools in
Pennsylvania and Virginia.  Schools in Pennsylvania were
generally located in the Reading Prong and were surveyed by the
Pennsylvania Department of Environmental Resources.  In each
school, multiple rooms were measured using either charcoal
adsorption, alpha track or grab working level.  Virginia schools
were surveyed by the Fairfax county, VA Public School System
using charcoal samplers.  These data were analyzed using a Monte
Carlo method to determine the ability of partial sampling and
statistical methods to predict the highest room in the school.  A
method was developed which yields reasonable success in detecting
the probable presence of elevated rooms (above 4 pCi/1) when ten
percent (or a minimum of ten)  rooms are sampled.  The confidence
in the prediction (type a and type b errors) is quantified as a
function of the parameters of the analysis.  The implications of
potential errors induced by partial sampling are discussed.

     This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review
policies and approved for presentation and publication.

-------
      PREDICTION OF MAXIMUM RADON CONCENTRATIONS IN SCHOOLS
                  USING  PARTIAL SAMPLING  METHODS

                           INTRODUCTION

     On April 20, 1989, the Environmental Protection Agency (EPA)
published preliminary guidance for monitoring of schools in the
United States (1).  This guidance recommends measurement of radon
in all rooms that are frequently used and are on or below ground
level.  Frequently used rooms include classrooms, offices,
cafeterias, libraries and gymnasiums.  Rooms used for storage,
broom closets, etc. would not be monitored.  This recommendation
is based on observations in schools in an EPA pilot study(2).

     While EPA clearly prefers measurement of every room, the
Agency recognizes that budgetary and other constraints may
prevent this, at least in the immediate future.  School officials
who cannot measure all frequently used rooms are then encouraged
to test those rooms with the highest potential for elevated
radon.  These include (1) basement classrooms, (2) occupied rooms
that are isolated from the central ventilation system or on
systems which only recirculate room air,  (3) rooms on or near
structural joints such as adjacent slabs, (4) rooms with a large
floor/wall joint perimeter and  (5) rooms that have floor slabs
with significant cracks.  The intent of this guidance is to
maximize the probability of detecting rooms with elevated radon
concentrations.

     This paper reports a technique which may be useful in
finding schools having rooms with elevated radon concentration
when it has been determined it is not possible to test every
room.  The technique is not endorsed by the Environmental
Protection Agency.  The method described here makes use of the
properties of the distribution of radon concentrations measured
in schools.  This distribution is used to predict the
concentration in the room with maximum radon based on a partial
sample.

     The method could be applied using a sample of rooms chosen
on a purely random basis, or it could be used to evaluate the
remainder of the rooms in a school after the rooms with highest
radon potential have been selected as described above.  It could
also be used as a screening tool to decide the order in which
several schools should be tested more completely if it is not
possible to test all at once.  The method yields an approximation
of the highest radon concentration in the school based on a
sampling of ten percent, or a minimum of ten rooms.

-------
                    DESCRIPTION OF THE METHOD

     In radon screening procedures, one would prefer that there
be a low probability of failing to detect rooms with elevated
radon (false negatives) because potential health problems may go
uncorrected.  Incorrectly classifying a school as having elevated
radon based on a preliminary screening procedure (false positive)
may be more acceptable for it provides a more conservative
approach in protecting public health.

     The basis of the partial sampling technique is that the
properties of a population (for example the radon concentration
in the rooms of a school) can be estimated from a sample drawn
from the population.  Given a sample of radon measurements taken
in some randomly selected rooms of a school, it should be
possible to devise a method to estimate the distribution of radon
in all the rooms within certain limits of uncertainty.  It may
also be possible to arrive at an estimate of the maximum radon
level in the school based on empirical data.  That is exactly
what is attempted here.

     For populations where the property being measured is
normally distributed and the sample  is chosen randomly, the mean
and standard deviation of a sample yield an unbiased estimate of
the mean and standard of the population.  If the parameter is not
normally distributed, the mean and standard deviation of the
sample are not unbiased estimators of the population.  Therefore,
in order to utilize the mean and standard deviation as
estimators, the population must be normally distributed.  For the
radon in schools data used here, the radon concentrations have
been found not to be normally distributed.  Most of these schools
show an acceptable  (p=.05) fit to a  lognormal distribution, so
lognormal statistics must be used.   The most straightforward way
to do this is to transform the radon data by taking the log
(either natural or base 10) of the radon concentrations(3).  The
resulting data yield an acceptable fit to a normal distribution.

     When the distribution of radon  concentrations is known, it
is straightforward to  estimate the radon concentration at any
percentile of the distribution.  Conversely, if a percentile is
chosen the radon concentration at this percentile can be
estimated.  The confidence limits of these estimates are a
function of the degree of certainty  with which the parameters of
the distribution have  been estimated, and how well the
distribution conforms  to the model (in this case lognormal).  The
allowable confidence limits can also be used to determine the
necessary sample size.

     However for the purpose of this paper it was desired to
estimate the radon concentration in  the maximum room  in each
school, not the radon  at a fixed percentile.  Since the number of
rooms is variable, there is not a  fixed percentile of the
distribution that can  be used.  Instead it was decided to use an
empirical approach to  relate the maximum observed radon

-------
concentration to the mean and standard deviation of the measured
radon concentrations in each of a large number of schools.  A
similar empirical approach, described below, was used to
demonstrate the resulting confidence limits.

     Similarly, the number of rooms in the sample was not derived
from sampling theory.  It was based on the authors' estimate of
what is practical.  It was felt that a minimum of ten rooms, or a
minimum of ten percent of the total number of basement and first
floor rooms, might yield an estimate of the mean and standard
deviation sufficient for the purpose at hand, an estimate of the
radon concentration in the maximum room in each school.  This
approach is tested empirically using a Monte Carlo method and the
resulting confidence limits are shown for each school
individually.  The number of rooms in the sample was varied and
found not to be an important parameter in the schools with the
"worst" predictions.  The overview of all schools for which this
calculation was done demonstrates the strengths and weaknesses of
the method.

     The mathematical form of the test is

           (1)       Xn + K sn < C     pass

           and (2)   Xn + K sn > C     fail

           where     Xn is the  mean of the sample.

                     sn is the  standard deviation of the sample.

                     K is a multiplier (number of standard
                       deviations) determined empirically to
                       yield the desired sensitivity.

     The method proposed here makes use of the technique outlined
above, and has been structured to be of practical benefit.  Many
choices in this candidate method are purely arbitrary and based
on practical considerations.  For example a method requiring 90%
sampling is of little practical use since any cost savings would
be more than offset by the expense of confirmatory monitoring if
a problem is detected.  A sampling rate of 10% was chosen to
yield a tangible benefit, and a 25% rate was also investigated
and found to yield little improvement over a 10% rate.

     The environment within which the method would be applied is
as follows.   If the estimate of the maximum radon is below 4
pCi/1 school officials may test further if they so decide, but
EPA believes it may be costly at this point to reduce the radon
below these levels.  If additional monitoring were performed it
would be a relatively low priority compared to schools yielding a
higher estimate.   An estimate of the maximum room above 4 pCi/1
would trigger confirmatory monitoring of the remaining rooms.   It
would not be the basis for mitigation.   There would be two ways
in which a radon problem could  be "found".  It could be directly

-------
measured if one of the radon tests came out above 4 pCi/1 or
could be inferred from the statistical calculation of the highest
room.  This yields two ways in which problems can be found and
increases the power of the test.  If rooms with a high radon
potential are also preselected for separate monitoring, the power
of the method is further improved.

     The first condition that must be satisfied is normality of
the data distribution.  Radon data in the schools here was found
to be approximately log-normal, and so can be normalized by
simply taking the logarithm.  All data used in this analysis has
been transformed in this way.  This allows the use of the sample
mean and standard deviation to estimate these parameters for the
population.  It is unimportant whether natural or base 10
logarithms are used here since the difference is only a scaling
factor which cancels out in the later mathematics.

     The second and more tenuous part of the analysis is to
estimate the radon concentration in the worst room.  To do this
an empirical approach was selected.  The logarithms of radon
concentrations were calculated from a subset of the schools
described below.  The mean and standard deviation of the
logarithms were calculated for each school.  The logarithm of the
maximum radon concentration was also calculated for each school.
The number of standard deviations from the log mean to the log of
the maximum observed radon concentration was then calculated for
each school.  For convenience this resulting statistic will be
called K1 because of its use in selecting a value of K used in
estimating the maximum radon as described above.  The values of
K1 were compiled and graphed separately for each method of
measurement used in Pennsylvania, for the Virginia charcoal data
and for data from the EPA phase I study of schools nationwide.
There was a remarkable similarity in the results in spite of
sampling methods and sample durations and radon levels which were
quite dissimilar.

     The derived values of K1 are plotted on Figures 1 through 5.
It can be seen that there is a distribution of this statistic
with a maximum at K1 = 2 to 2.5 standard deviations from the
mean.  This relationship appeared to hold well across the data
examined here. This is the basis for the choice of the value of K
in the predictive applications of the method.  Note that the
maximum value of the statistic K1 is 4.5 standard deviations from
the mean and several schools yielded K' of 4 standard deviations.
This did not appear to be a function of the number of rooms in
the school, but was related to the presence of rooms with unique
characteristics that caused high radon.  Given this distribution
of K1 the choice of K at 2.0 or 2.5 in the method above would
yield a central value in the prediction of the highest radon in
the school.  The presence of schools with K1 as high as 4.5
guarantees that the method cannot predict all rooms above 4 pCi/1
all the time unless K is set at or above 4.5.  The presence of
schools with K1 close to 1 will result in overprediction part of

-------
the  time.  The  probability  if  overprediction and underprediction
at various values  of  K is the  principal  result of this paper.
                       NAT I ON A L S C HOOL S
                              EPA SURVEY



                                      
                     1    1.3   2    2.9   3    3.3
                     NUMBER OF LOG STD DEVIATIONS TO MAX
Figure l
         e
         e
         7
         6
         9
         4
         3
         2
                      VIRGINIA SCHOOLS
                              CHARCOAL


          I
           O.SO   1 DO   1.SO   2 OO
                     NUMBER OP LOG
2 SO
sro
   3 00   3. SO
DEVIATIONS TO MAX
                                                    SO   S.OO
Figure 2

-------
                            PENNSYLVANIA  SCHOOLS
           20
           19 -
           18 -
           17 -
           1G -
           15 -

           13 -
           12 -
           11 -
           10 -
            9 -
            a -
            7 -
            B -
            5 -
            4 -
            3 -
            2 -
            1 -
                                      GRAB WORKING LEVEL
               0 500  1 DOO  1 SOO  2 000  2 500  3 000  3 500  4 000   4 500   5  000
                              NUMBER OF  LOG STD DEVIATIONS TO MAX
Figure  3
                            PENNSYLVANIA  SCHOOLS
                                           CHARCOAL
            3 -
            2 -
            1 -
                0 50    1 00    1  50    2 00    2 50    3 00    3 50
                              NUMBER OF LOG STD DEVIATIONS TO MAX
                                                                 4 00
                                                                        4 50
                                                                               5 00
Figure  4

-------
                    PENNSYLVANIA  SCHOOLS
                             ALPHA TRACK
        5 -
        3 -
        2 -
           0 50   1 00   1 50   2 00   2 50   3 00   3 SO   4 00   4 50   5 00

                     NUMBER OP LOG STD DEVIATIONS TO MAX
Figure 5
                    DATA USED FOR THE ANALYSIS

     Radon measurements were obtained for eighty-seven schools  in
Virginia which were measured by the Fairfax County Virginia
Public School System using bag-type charcoal detectors.  The
Environmental Protection Agency Phase I school data(4) for nine
other areas of the country was also obtained.  The radon in this
dataset was measured with EPA charcoal canisters.  Radon data was
also obtained from the State of Pennsylvania for schools measured
by the State as a part of its Reading Prong investigation.
Seventy-nine schools were measured using three month alpha-track
detectors, thirty-two schools were measured using charcoal
detectors and fifty-five schools were measures with grab working
level samples.  At the time when these measurements were made,
the grab working level was the principal tool for doing rapid
radon assessments.  In schools where elevated working levels were
found in "high probability" rooms as defined by Pennsylvania
(this may differ from the later EPA definition), every room in
the school was tested.  Where no elevated working levels were
found, only a few rooms were tested.  This data sometimes
included boiler rooms which are infreguently inhabited.  School
grab samples were also followed up with alpha track detectors,  so
many of the working level and alpha track schools are common
between the datasets.  Charcoal tests were used later in the
Pennsylvania program, and were made in a separate group of
schools.  For this analysis the grab samples and alpha track

-------
results are treated as if they were not from the same schools
because it was desired to test the method with both measurements.
This did not affect the outcome of the analysis since the number
of rooms tested with alpha track detectors was below the lower
limit included in the Monte Carlo analysis but allowed the value
of K1 to be investigated for this measurement method.

     A preliminary culling of data was performed to eliminate all
schools where less than ten rooms were sampled.  This was done
because the first part of the analysis examined the mean and
standard deviation of the logs of the concentrations.  It was
felt that the statistics would only be meaningful where there
were at least ten items of data.  This also served to eliminate
schools where very few rooms were sampled.  An additional
restriction placed on the data was that the maximum radon
concentration be at least 2.7 pCi/1 or .014 WL.  This was done to
assure that the radon measurements were well above the lower
limit of detection so that there would be some confidence in the
radon measurements reported.  After this first cut there were
twenty-nine schools in the Virginia charcoal dataset, eleven
Pennsylvania schools measured with alpha-track detectors, seven
Pennsylvania schools measured with charcoal samplers and thirty-
three Pennsylvania schools measured using a working level grab
sample.  This also left 110 schools in the Phase I national
dataset which were measured with EPA charcoal canisters.  Further
culling was done for the Monte Carlo analysis below.

                       MONTE CARLO  ANALYSIS

     Monte Carlo analysis is a technique that allows analysis of
data and distributions which are not regular or which are not
easily manipulated using theoretical techniques(5).  At this
point it is important to recognize that the processes which
determine the radon concentration in a school room are, for the
most part, not random.  Factors which determine the radon level
include the radon concentration in the soil gas, the ease with
which radon can move into the room, and the air pressure that
drives the radon.  Because the dominant factors governing the
radon concentration are not random, there is no good reason to
expect that the radon concentrations will exactly fit any of the
commonly used statistical distributions.   The observed fit to the
lognormal distribution allows the data to be characterized using
the geometric mean and geometric standard deviation, but does not
allow an exact analysis of the real-world situation.  In this
case, Monte Carlo techniques allowed assessment of the
performance of the proposed radon analysis method using real
radon distributions rather than an idealized model.

     The Monte Carlo simulation discussed below was restricted to
schools where more than fifty rooms were sampled.  This was done
to improve the validity of the results, and because the
utilization of the method proposed here will most likely occur
where there are large buildings to measure.   This selection of
data also removed the schools from the dataset where all rooms in

-------
the school were not sampled.  This selection of schools resulted
in a Monte Carlo analysis of twenty-nine Virginia schools
measured with charcoal and four Pennsylvania schools measured for
grab working levels.  No schools in the Pennsylvania dataset had
50 rooms measured with charcoal or alpha track detectors.
Because the Pennsylvania schools with all rooms measured were
selected because of high working level measurements, this allowed
examination of schools with very high radon concentrations as
well as those with only moderate levels.

     The selected schools had radon measurements in at least 50
rooms.  From these populations, samples of ten rooms each were
selected at random without replacement  (no room could be selected
twice in the same sample).  One hundred of these random samples
were generated for each school.  This is as if the school had
actually been sampled 100 times, each time with a different
selection of rooms.

     For each sample, the mean and standard deviation of the
logarithms of the radon concentrations were calculated.  For each
sample, the mean was added to the standard deviation multiplied
by 0, 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 3.5, 4.0, 4.5, and 5.0.  This
is in accord with the form of the test described above, and the
multipliers are the values of K to be tested.  The antilog of the
result was taken for each sample and compared to 4 pCi/1 or .02
WL.  As the number of standard deviations increases, there is a
greater probability that a result above 4 will be calculated.

     Since there are 100 different samples for each value of K,
the number of times the result is above 4 pCi/1 or .02 WL
represents the probability, p, (in percent) of a predicted
maximum in excess of the selected limit at that school (a "hit").
This result is specific for that value of K.  This analysis
utilizes the real observed radon distribution in the school, and
the result varies from school to school.  For schools with no
actual concentrations above 4 pCi/1 or .02 WL, the number of
"hits" represents the probability of a false positive at that
school.  For schools with measured rooms above 4 pCi/1 or .02 WL,
the complement of the probability of detection (1-p) represents
the probability of a false negative.  This data is presented in
graphic form for a number of schools is in figures 6 through 8.

     Figure 6 (school 4151) is typical of a school with no rooms
above 4 pCi/1, but selected to have at least one room above 2.7
pCi/1.  Because this school had no rooms above 4 pCi/1, the chart
shows the probability of a false positive.   Note that for K =
1.5 and below, none of the 100 trials resulted in a predicted
maximum greater than 4 pCi/1.  For K = 2.5 there were 34 hits,
giving a 34% probability of a false positive for this school for
this value of K.  As K is increased to 3.0, there is a 67%
probability of a false positive.   Several similar schools yielded
similar results.  This suggests that K be chosen below 3.0 if the
false positive rate is to be kept below 50% for schools with the
highest room just under 4 pCi/1.

-------
   s
   u.
   =
   i
   ft
   e
11O

1OO -

 so -

 eo -

 70 -

 oo -

 so -

 •40 -

 ao -

 2G -

 1O -

 O
               SCHOOL  MONTE-CARLO ANALYSIS
                         SCHOOL -»131 - NO ROOMS »• •*
            O    O    O    O




           O OO  O SO  1 OO  1 SO  a OO  2 SO  3 OO

                    NUMBER OP LOC STO DEVIATIONS - LOC
                                        3 30

                                        MEAN
                                                     i
                                                     SO  9 00
Figure 6
               SCHOOL  MONTE- CARLO ANALYSIS
£
^
i
11D


1OO

 SO

 SO


 70


 OO


 SO




 aa


 3D


 10


 O
                        SCHOOL SO BO - 3M OP ROOMS «-
            ooo





           o.oo  o.so  i oo  t.so  a.oo  a so  3.00  3.90

                    NUMBER OP LOG STO DEVIATIONS » LOG MEAN
                                                        I
                                                     .30  9.OO
Figure 7
     Figure 7  (school  2086)  is typical of a school with only a
few rooms slightly  above  4  pCi/1.   Only 3% if the rooms in this
school are above  4  pCi/1.   Because this school actually has rooms
above the critical  radon  concentration, the chart shows the
probability of correctly  classifying the school.  Subtracting
these probabilities from  one yields the probability of a false
negative.  In this  case,  a  K of 2.0 gave a probability of
detection of 28%, or a probability of a false negative of 62%.
For K = 2.5 the probability of a false negative is 40%.  For K =
3.0 it is 20%.  Several similar schools yielded similar results.
This suggests K be  selected above  2.0 if marginal cases like this

-------
are to be detected with a reasonable degree of certainty.
Whether 40% probability of a false negative is acceptable in this
case is a policy decision beyond the scope of this paper.  The
highest radon measured in this school was 5.4 pCi/1.
SCHOOL

FBKBIT PROBABILITY GF POSITIVE FIIOIIG
j
-lUU&UOOCDCDC
IOOOODOODOC
1 	 Ill
a o
0 . OO 0.30
1O
^
MONTE- CARLO ANALYSIS
SCHOOL 332-7 - 13* OF ROOMS *• •*

03
I

Bf
1


I

1QO
I

1OO 1OO 1OQ 1OO
I







1.00 1.SO Z.OO 2 SO 3 OO 3 SO 1 OO 1 . 3O 3 . OO
NUMBER OP LOG STD DEVIATIONS » LOG MEAN
Figure 8
     Figure  8  (school  3327)  is clearly elevated in radon with 13%
 of  rooms  above  4 pCi/1.  The highest radon concentration is 6.9
 pCi/1.  In this case,  K =  2.5 gives a 97% probability of
 detecting this  problem.  There was another similar school which
 yielded similar results.   This suggests the choice of about 2.5
 for K yields a  good probability of detecting schools with similar
 radon concentrations.  Several Pennsylvania schools with much
 higher radon concentrations  were  included in the analysis.  For
 these schools,  the probability of detection is virtually 100% for
 K = 2.5.

     There were four schools in the dataset which gave misleading
 results.  One  school with  no rooms above 4 pCi/1 yielded a
 probability  of  a false positive of 93% for K = 2.5.  Decreasing  K
 below 2.0 made  only a  slight improvement.  This problem appears
 to  be due to a  wide spread in radon concentrations which gave a
 large standard  deviation.  The maximum radon was 3.7 pCi/1.  It
 is  possible  that measurement at a different time would result in
 a room above 4  pCi/1.  Three schools with 6%, 2% and 2.4% rooms
 above 4 pCi/1  yielded  probabilities of a false negative of  53%,
 82% and 69%  for K = 2.5.   For these three schools, increasing K
 above 3 did  not result in  a  great improvement.  This appeared to
 be  due to a  large number of  rooms with a similar radon
 concentration  which yielded  a small standard deviation.  In each
 case there was  a room  which  did not appear to fit the
 distribution.   For all four  of these schools, increasing the
 sampling  rate  to 25% of rooms with a minimum of 25 did not  yield
 any improvement.  None of  these rooms exceeded 10 pCi/1.  It is

-------
the opinion of the authors that these rooms could not be detected
except by measuring them.  No system based on the distribution of
the other rooms will detect them.  Preselection of rooms based on
high potential for radon entry gives the best chance of finding
these rooms short of measuring every room, but depends heavily on
the ability to identify such rooms based on known physical
features.   It was not possible to examine the effect of
preselection in this study because the physical characteristics
of the rooms which would allow preselection were not reported in
the data used.

                             RESULTS

     Examining a number of schools allows the pattern of false
negatives and false positives to be determined at each multiplier
of the standard deviation.  If acceptable levels of error can be
chosen, then a constant can be selected to yield that level of
error.  Figures 9, 10, and 11 are a summary of all schools
analyzed with the Monte Carlo method.







i
s
UJ
H
8
$
t
£
S
tt








90 -

ao -
70-!

60 -
50 -

40 -
30 -

20 -1
I
10 -j
.i
c
METHOD SUMMARY - MEAN + 2 STD DEV
GREATER OF 10 ROOMS OR 101 OF ROOMS

O

D

O
I



§


0 a
i
1 % a
i
1 a Q
2 4 G B 10 12 14 1
PERCENT OF ROOMS ABOVE 4 PCI/L



















B
Figure 9
     Figure 9 shows the performance of the method with K = 2.0.
Each data point represents an individual school.  The abscissa is
the percentage of rooms above 4 pCi/1 actually measured in the
school.  For the purpose of this analysis these radon
concentrations are taken as truth, though in fact there are
errors associated with the measurement methods and due to
temporal variations.  The ordinate is the probability of a "hit"

-------
with K = 2.0.  For schools on the Y axis, there were no rooms
above 4 pCi/1, so the Y value is the probability of a false
positive.  For all other schools, there were rooms above 4, so
the Y value represents the probability of correctly classifying
the school.   (The point at 0,72 is one of the exceptions
mentioned above.)
METHOD SUMMARY - MEAN + 25 STD DEV
100 -
30 -
BO -
B 70-
* 60-
9 '
a so .
ft
GREATER OF 10 ROOMS OR 10* OF ROOMS
0
I
a

S
a
l
D
I D °
I
^ 9
| 3Q 7 ° °
10 -
D -
i°° "






] 2 4 6 8 10 12 -M 16
PERCENT OF ROOMS ABOVE 4 PCI/L
Figure 10
     Figure 10 shows the results for K = 2.5.  Note the increase
in the number of false positives and the increase in confidence
in detecting schools with elevated radon.  The school at 6,47 is
one of the exceptions discussed above.  Except for this one
school, there is a high confidence in detecting schools with more
than 4% of the rooms elevated.

     Figure 11 shows the results for K = 3.0.  Note the still
higher occurrence of false positives, and the further increase in
confidence in detecting elevated rooms.  The school at 6,50 is
the same exceptional school.  The probability of false positives
becomes high enough to be cause for concern at this value of K,
but it must be kept in mind that the lowest school included had
radon above 2.7 pCi/1 in at least one room.  The schools with
very low radon have all been excluded from the dataset.
Inclusion of these schools would yield a smaller proportion of
false positives.

-------
METHOD SUMMARY - MEAN + 3 STD DEV
100 -|
90 -

3 "'
Ul
= 70 1
> '
1 «,-
fe 30-1
t
=i -,0-j
i
20 -
in -
c
GREATER OF 10 ROOMS OR 10K OF ROOMS
a
a

D
a
a
a
D a
-a
a
a
a










' 2 ' 4 ' B ' B ' 10 12 1" IB
PERCENT OF ROOMS ABOVE •* PCI/L
Figure 11
                    APPLICATION OF THE METHOD

     The method might be applied as follows:

1.   Choose rooms with high potential for high radon
concentration and measure them.  This step is optional but very
desirable.  It will identify obvious problems but will not
identify rooms which are "high" in radon for hidden causes.

2.   Select a random sample from the remaining rooms of the
school.  Measure ten percent of the rooms or ten, whichever is
greater.  Ten percent was chosen because it would result in a
substantial savings over complete testing.  The minimum of ten
rooms avoids very small samples.  Choices other than ten percent
or ten samples are possible but have not been investigated here
to quantify the uncertainties.

3.   Measure the radon in the selected rooms.  Examine the radon
results for unacceptable concentrations which clearly indicate a
problem.   If a radon problem is found in this way there is no
need for the statistical analysis below.

4.   Calculate the logarithm  (natural or base 10) of each
concentration in the random sample.  Do not include the rooms
selected for high radon potential.  Calculate the mean and
standard deviation of the log radon concentrations of this

-------
sample.  Add "K" standard deviations to the mean.   Calculate the
antilog of this value.  This is the expected radon concentration
of the highest room in the population from which the sample was
drawn.  The choice of "K", the number of standard deviations,
determines the sensitivity if the test.  K = 2.0 yields fewer
false positives at the risk of greater chance of a false
negative.  K = 3.0 yields greater sensitivity at the expense of
more false positives.  A value of 2.5 for K seemed to give a good
balance, but choice of K should be made based on the error
probabilities which are acceptable to the decision-maker.

                            DISCUSSION

     It can be seen from the results above that partial sampling
does not give one hundred percent certainty of identifying
schools with rooms above 4 pCi/1.  Such a result is to be
expected.  Even when every room is sampled, the errors inherent
in measurements and the variability of radon in time prevent
positive determination in marginal cases.  When measurements are
made,  uncertainties cannot be avoided.  The method investigated
here can identify most schools with rooms above 4 pCi/1 with
partial sampling.  The errors of commission and omission are
quantified here and are available for use in deciding what choice
of rooms to be sampled is appropriate.  It is the task of school
administrators and policy-makers to decide what constitutes an
acceptable uncertainties.  The uncertainties associated with this
method are similar in magnitude to the uncertainties in short-
term measurements.

     The uncertainties associated with application of the method
can be reduced by preselecting rooms with a potential for high
radon  concentrations.  In this case, the rooms not preselected
could  be sampled and  analyzed using this technique.  Since the
EPA guidance allows monitoring of only the high radon potential
rooms  as an acceptable but  less preferred option, use of this
method on the remaining rooms would be expected to yield an
improved result and  represents a middle ground.   Its use is  not
precluded in the guidance nor  is it endorsed by EPA.  It is  also
possible that school  districts may not have the funds to measure
all rooms in a  number of  large schools at the same time.  Using
the method outlined  here, it  is possible to conduct partial
sampling on all schools with  the object of prioritizing  the  more
complete measurements.  Schools with  suspected problems  might be
measured  first  and others would be deferred until  funds  are
available for measuring every  room.   Partial sampling can
identify schools at  which we  suspect  the presence  of rooms above
4 pCi/1, but the only way to  identify specific rooms is  to test
them  individually.

      The choice of 4  pCi/1  was made within the context of EPA
guidance at the time.  The  method  can be applied  equally well  for
a cut point at  another radon  concentration, or expressed in
another unit.   Note  that  in this case the method  did not need  to
be  changed to  accommodate a 4  pCi/1  limit  and a  .02 WL limit.

-------
     To the extent that schools represent other large buildings,
this analysis may be transferable.  There would have to be some
investigation of the proper choice of K, and the tolerable errors
may be somewhat different.

                         ACKNOWLEDGEMENTS

     The authors wish to thank Harry Chmelynski of the Washington
Consulting Group for development of much of the theory upon which
this work is based and for supplying the Fairfax Schools data and
Thomas Peake and Maureen Clifford of the Environmental Protection
agency for review of the manuscript.

                            REFERENCES


1.   Radon Measurements in Schools, EPA 520/1-89-010, U.S.
Environmental Protection Agency, Washington, DC 20460, 1989.

2.   MacWaters, J. et al, S. Cohen & Associates, Inc., Me Lean,
VA and Chmelynski, H. et al, The Washington Consulting Group,
Inc., "Radon Measurement in Schools", Contract 68-02-4375 for
U. S. Environmental Protection Agency,  July, 1988.

3.   Gilbert, R. 0., Statistical Methods for Environmental
Pollution Monitoring. Van Nostrand Reinhold Company, Inc., New
York, NY, 1987.

4.   MacWaters, J. et al, S. Cohen & Associates, Inc., Me Lean,
VA and Chmelynski, H. et al, The Washington Consulting Group,
Inc., "Final Report - School Radon Protocol Development Study",
Contract 68-02-4375, for U. S. Environmental Protection Agency,
September, 1989.

5.   Hammersiey, J. M. and Handscomb, D. C., Monte Carlo Methods.
John Wiley & Sons, Inc, New York, NY, 1964

-------
                                                        D-IX-2
      COMMERCIAL MITIGATION TECHNIQUES USED IN REMEDIATING

                  A 2200 pCi/L PUBLIC BUILDING
         by:   James G. Davidson
               Radon Detection Services, Inc.
               Ringoes, New Jersey  08551
                            ABSTRACT

     In March of 1989 EPA and Pa. DER officials were amazed to
discover a school in Pennsylvania with levels in its library of
2200 pCi/L.  The library was a 30 year old, three story slab
on-grade structure more like a commercial building than a typical
school structure.  It had three separate and complex HVAC systems.
Initial diagnostics indicated radon levels under the slab at over
80,000 pCi/L.  Further investigations revealed major entry routes
and a HVAC system terribly out of balance.

     Remediation consisted of installing a complex sub-slab
depressurization system with an exterior commercical fan unit,
major entry route sealing, and working closely with a
mechanical contractor to bring the HVAC systems back into balance.
Initial post remediation testing showed a 99% drop in radon levels,
Refinements to the system are still in progress.

-------
     As the EPA and various states begin urging the testing of all
schools nationwide; and, as laws are passed mandating the testing
of commercial buildings, facility directors will ultimately have to
grapple with radon remediation in complex buildings.  Unlike
residential remediation where mitigation techniques are becoming
standardized, little work has been done in schools with complex
HVAC systems or in commercial buildings with hundreds of thousands
of square feet of office space.  In dealing with existing commercial
buildings or with commercial buildings under construction,  new
mitigation techniques are still evolving, as is the retrofitting of
residential techniques for commercial applications.  The following
presentation is the result of a recent school mitigation in
Pennsylvania in which this building had the highest radon levels
found to date in a school or commercial building.

     In March of 1988, a student attending Westminster Theological
Seminary in Philadelphia, Pennsylvania placed a charcoal canister
on his desk on the ground floor of the school's library.  To his
amazement, the levels came back at over 2,000 pCi/L.  He alerted
school officials who conducted theri own test with similar results.
School officials immediately contacted the Pennsylvania Department
of Environmental Resources who in turn contacted the EPA.  Very
quickly, EPA officials from the regional office, Washington and
the research division in North Carolina converged on the scene to
perform diagnostics and confirmatory testing using an AB-5 Pylon
Radon Monitor.  Their testing showed levels over 2200 pCi/L.  Gas
samples from under the slab, peaked out at over 80,000 pCi/L.
Students were evacuated from the ground floor of the building and as
word of the problem leaked, newspaper and T.V. crews arrived at the
school.  The story hit the nightly news and was in every local paper
the next day.

                          THE BUILDING
     The library is a three (3) story 30,000 square feet slab on
grade building constructed about 30 years ago.  The first floor
is partially built into a bank so entry is on both the first and second
floors, most of the building is open as is typical of libraries.  Each
floor has its own HVAC system.  The HVAC unit for the first floor is
on the first floor while units for the second and third floors are
located in a penthouse on the roof.  There is a large exhaust fan in
the penthouse which is ducted to all of the bathrooms for a bathroom
exhaust.  This fan is on a timer with a two (2) hour on/off cycle.

                       DIAGNOSTIC TESTING
     While residential testing has on occassion revealed radon
levels in the 2,000 - 3,000 pCi/L range, commercial buildings were
thought to pose less of a problem.  Because of the large expanse
of the building area, the complexities of the buildings HVAC

-------
systems, and the problem of finding and sealing radon entry routes,
this mitigation, as with other commercial buildings,  required more
building investigation, more diagnostic testing and,  consequently
more sophisticated solutions then any residential dwelling.

    While EPA testing data showed ambient first floor radon
levels in the library fluctuating between 1500-2200 pCi/L,
second floor measurements were about 100 pCi/L.  Sub-slab
measurements of 80,000 pCi/L indicated we were dealing with an
extremely hot slab.  Numerous floor cracks were observed as was a
large expansion pour joint in the first floor storage room.  When
a section of the rubberoid baseboard was removed a 3/4" separation
between the wall and expansion joint was observed.  Communication
testing1 revealed fair to marginal (.oo2" H^O) between the main
storage and mechanical room, meaning we had either little aggregate
under the slab or a very strong negative pressure.  Building plans,
while very detailed, nowhere indicated the amount nor the size of
aggregate under the slab.  Drilling at the communication sites
suggested 4" - 6" of coarse aggregate base.  Micro-roentgen gamma
readings inside and outside the building while elevated, did not
indicate any unusual source.

    The HVAC system was of considerable concern because the air
handling units were way out of balance with no fresh air coming
into the building.  Over a 30 year period there was never a
maintainanee contract on the system but instead, whenever there
was a problem, a different mechanical contractor was hired.  The
problem was obvious:  a strong source under the slab, numerous
entry routes and a HVAC system totally out of balance exerting
a negative pressure on the building.

                      REMEDIATION PLAN
    Any remediation plan had to take into account a major sealing
effort, a sub-slab depressurization system (SSD) to remove the high
radon concentrations under the slab, and a balanced HVAC system.

    To achieve suitable depressurization under the slab, all
wall/floor joints in the lower level had to be sealed, as well as
any floor cracks and entry routes around electrical conduits, etc. on
the first floor.  All exposed floor cracks were ground out and filled
with a flowable urethane.  All 4" rubberoid baseboards were also
removed and the wall/slab expansion joint filled with a flowable
urethane, as well.  Great care was taken to seal as many entry routes
as possible to improve the operation of the depressurization system
as well as limit the influx of the gas.
 Communication Tests - Communication tests are used to determine
the ease or difficulty with which gas can move through the soil,
aggregate under the slab or within a cavity such as a block wall.
Typically two holes are drilled 20' apart and suction applied to one
hole while micromanometer measurements are made at the other.

-------
     The principal of a sub-slab depressurization system (SSD) is to
create enough negative pressure under the slab to pull radon out
before the negative pressure of the building can pull the gas into
the structure.  The keys to this system functioning properly are
1) an aggregate base that allows for the establishment of an adequate
pressure field and 2) a fan that will have sufficient CFM to draw
the gas from all sub-slab areas, up three stories and out of the
building.

     The depressurization system designed for the building had four
suction points.  The vertical piping from these holes was 4" PVC
manifolded into 6" PVC horizontal pipe located along the first floor
ceiling.  The 6" PVC piping was then routed to a 4' x 6' pipe chase
behind the elevator shaft, and then up to the roof penthouse.  Before
a fan was chosen, the 4" holes were cut in the slab to determine
the type of thickness of the aggregate.  The building had 6" of
IV crushed stone which was a plus for the SSD.  With this information,
a wall mounted fan (Penn Ventilator WAQ-24L-800 CFM/1.00: S.P.)
was installed on the north wall of the penthouse.  By using a wall
mounted unit, the EDPM roofing did not have to be penetrated and its
guarantee was not breached.

     The HVAC system, due for scheduled repairs, was only partially
corrected.  The correction involved opening the fresh air intakes
and rough balancing the system by a new mechanical contractor.  Fine
tuning of the system is still to be accomplished.

                   POST REMEDIATION TESTING
     Installation of the system took 3 men/3 days.  Much of this
time was spent answering questions of media personnel.  Once the
system was operational we immediately retested all areas of the
building and found the first floor levels had dropped from 2200
to 10 pCi/L and the second floor from 100+ to 5 pCi/L.  Subsequent
diagnostic measurements revealed good communication with the
under slab and the ground floor ambient air with pressure differntial
ranging from 4-28 pascals.  Static pressure measurements between
the interior riser pipes ranged from 50-83 pascals with 325 pascals
at the fan.  Fine tuning of the HVAC system should drop these still
slightly elevated levels below the EPA guideline.

                          IN CLOSING
     The remediation techniques used in this mitigation were designed
and based on detailed information supplied from sophisticated diagnostic
measurements.  While the sealing and SSD systems were somewhat
similar to residential remediation techniques the HVAC system was the
curve ball.  As more commercial buildings and schools with complex HVAC
systems are shown to have radon problems, the radon mitigator in
partnership with the mechanical contractor will be required to be a
detective and problem solver all in one.  Commercial mitigation is surely
going to be the new frontier of radon.

-------
     The work described in this paper was not funded by the U.S.
Environmental Protection Agency and therefore the contents do not
necessarily reflect the views of the Agency and no official
endorsement should be inferred.

-------
                                                        D-IX-3
                 EPA'S PROTOCOL DEVELOPMENT STUDY
                       FOR RADON  IN  SCHOOLS
                           Anita Schmidt

               U.  S.  Environmental Protection Agency


                             ABSTRACT
      A year long radon measurement study is being conducted  in 21
 schools in  7  states geographically  dispersed  across the United
 states.   These schools were selected nonrandomly from 130 schools
 previously screened  for  radon during  the  Winter of  1989 in  the
 first Phase of the study.  The purpose of this study  is to gather
 seasonal  radon data and compare measurement methods to update  and
 refine EPA's measurement guidance for schools.
Gomr>r.oha«,fJ^Sei  ZI  °£  ^he  studv'  initiated   in  June  1989,
comprehensive long- and short-term measurements are being made for
£.?^r«UA"!?-«i Tiet£ °f, devices-  Long-term  measurements are
being made  with alpha track detectors and electret-ion chambers.
snort-term  measurements, using charcoal  canisters,  electret-ion
chambers, and continuous monitors for radon and radon progeny, are
being made under different ventilation conditions.  The results of
these different measurement  methods will  be compared  for their
applicability in schools.   The impact of school  construction type
and  ventilation  systems   on   radon measurement  will  also  be
     This paper has been reviewed in accordance with the U. S.  Environmental
Protection Agency's peer and administrative review policies and  approved for
presentation and publication.

-------
                                                     D-IX-4



   RADON LEVELS IN NON-RESIDENTIAL BUILDINGS IN NEW JERSEY


                        Karen Tuccillo

      New Jersey Department of Environmental Protection


                           ABSTRACT
     Radon testing in various structures has been conducted
throughout the State of New Jersey including non-residential
buildings.  Non-residential buildings differ significantly in
construction from private homes, therefore radon levels seen in
non-residential buildings may not reflect radon levels seen in
surrounding residential structures.  Testing has demonstrated
that radon/radon decay products can build up to unacceptable
levels in non-residential as well as in residential structures.
The concentration of radon was measured in over 120
non-residential buildings, including schools, hospitals, and
municipal buildings in New Jersey.  All structures tested had
initial radon levels in excess of the USEPA's 4.0
picocuries/liter continuous exposure guideline for residential
structures.

     The distribution of radon levels for various non-residential
structures is examined and the efficiency of remedial measures
taken described.  Data indicate that radon levels measured in
non-residential buildings are generally lower than those found in
residential structures.  Correlation is seen between geographic
areas and elevated radon levels.

-------
                                                                   D-IX-5
                           ELECTRET ION CHAMBERS FOR
                     RADON MEASUREMENTS IN SCHOOLS DURING
                        OCCUPIED AND UNOCCUPIED PERIODS

                  by:   Kenneth D. Niggers, Tom D. Bullets, and
                        Paul A. Zoske
                        American Radon Services, Ltd.
                        ISIS Center
                        Iowa State University Research Park
                        Ames, Iowa   50010

                        Kelly W. Leovic
                        U.S. Environmental Protection Agency
                        Air and Energy Engineering Research Laboratory
                        Research Triangle Park, North Carolina  27711

                        David W. Saum
                        Infiltec
                        P.O. Box 1533
                        Falls Church, Virginia  22041
                                   ABSTRACT

      Continuous  radon  measurements  made in some  school  buildings have shown
large diurnal variations due to factors such as occupancy patterns, HVAC system
design and operation, and weather.  Since many radon screening measurements are
made  with relatively  inexpensive detectors  that integrate  the  daytime  and
nighttime exposures, the  test results may not reflect  the  actual exposure of
students and teachers during occupancy hours.

      This paper  discusses  the results of  radon measurements  made in schools
with electret  ion chambers  (EICs)  and continuous  radon monitors (CRMs) during
both occupied and unoccupied periods.  Ten school buildings in Iowa, Maryland,
and North Carolina were  tested in this study.  The results show that radon levels
in these schools during occupied and unoccupied periods may vary significantly
depending on the  specific building (i.e.,  whether  the building is pressurized
or depressurized  during the given  testing  period)  giving both  false positives
and false negatives.


      This paper  has been reviewed in accordance  with  the  U.S.  Environmental
Protection Agency's  peer  and administrative review policies and approved for
presentation and publication.

-------
                                 INTRODUCTION

      In April  1989,  the  Environmental Protection  Agency (EPA)  released an
interim report  for radon testing in schools^ ?   Two types  of radon screening
measurement devices and two respective methods of radon testing during the winter
months are recommended:  (1) 2-day exposure open-faced charcoal canister exposed
over a 2-day weekend with the  school's  mechanical  system operated in the same
manner as  it  is operated when school  is  in session;  and  (2) an alpha track
detector exposed for 3 months.   It  is recommended that frequently used rooms on
and below ground-level be tested.

      As an alternative  to  conducting radon screening  measurements in schools
with charcoal canisters  or  alpha track detectors, radon measurements  in this
study were made in 10 schools using electret ion chambers (EICs).  Seven of the
school buildings are in Iowa, two are in Maryland,  and one is in North Carolina.
The measurements in the Maryland and North  Carolina schools  were part of an EPA
research project  on radon reduction methods in  schools.   A  continuous radon
monitor (CRM) was used  for comparison with the EICs in the  measurements conducted
in the Maryland schools.

      EICs were used for the radon screening measurements in this study since
they provide one method of measuring radon levels during occupied and unoccupied
periods without investing in an expensive CRM.  Because the  EIC measurements in
the three states were  conducted independently, they are  discussed separately.

                              BACKGROUND ON EICs

      EICs require no power and function as true  integrating detectors, measuring
the  average concentration  during  the  measurement  period.   EICs contain  a
permanently charged electret  which collects  ions formed  in the  chamber by
radiation emitted from radon decay  products. When  the device  is exposed, radon
diffuses into the chamber through filtered openings.  Ions which are generated
continuously by the  decay of radon and radon decay  products  are  drawn to the
surface of the electret and reduce its surface  voltage.  The amount of voltage
reduction is directly related to the average radon  concentration present during
the exposure period. Both short-term (2 to 7 day)  and long-term (1 to 12 month)
EICs are currently marketed.  The thickness of the electret affects the usable
measurement period.  Short-term EICs were used in this study.

      The electret must  be  removed from the canister and the electret voltage
must be measured with a special  surface voltmeter both before and after exposure.
The difference  between the  initial and final  voltage is divided  first  by  a
calibration factor  and then by the number of  exposure  days  to  determine the
average radon concentration during  the  exposure period.   For a 7-day exposure
period using  a short-term  EIC, the  lower  level  of  detection  is about 0.3
picocuries per liter, pCi/L  (1 pCi/L - 37 becquerels per cubic meter,  Bq/m3). (2)

-------
                          MEASUREMENTS  IN  IOWA  SCHOOLS

METHODOLOGY AND RESULTS

Phase 1.   EICs were used for all radon  testing  in the Iowa  schools.   Every
classroom  in  the six building Urbandale  Community School system  and all the
offices in the Administration building were tested for radon over two intervals:
(1) from  a Friday after  classes  to  late  Sunday afternoon; and  (2)  from late
Sunday afternoon to  Wednesday after classes.  The  EIC  electret  voltages were
measured on-site late Sunday  afternoon; the EICs were left in place and picked
up Wednesday  after classes and returned  to  the laboratory for final electret
voltage reading.  Testing was conducted while school was  in session in December
1988.

      The  first  step in the analysis of the Urbandale radon concentration data
was to extract from the database values greater than 3.0 pCi/L.  These  data were
sorted into two  groups:   (1) weekday radon concentration greater  than weekend
radon concentration, and  (2)  weekend radon concentration greater  than weekday
radon concentration.

      Figures  1  and 2 show the values greater than 3.0 pCi/L in bar graph form.
A building specificity for weekday radon concentration being greater than weekend
radon concentration  or the converse  is apparent in both figures.  A mechanism
seems to function at the building  level to determine the radon concentration for
the respective time periods.

Phase 2

      The  radon  concentration was measured during  the occupied time and during
the  unoccupied  time (approximately  daytime  and nighttime) in several of the
school  buildings.    Buildings  were  selected  that  represented both  radon
concentrations  greater   on  the   weekday  than  on  the  weekend and  radon
concentrations greater on the weekend than the weekday.   Each frequently used
room of the Administration and Olmsted buildings was  tested with EICs  during the
occupied  and  the unoccupied time  of the  5-day workweek.   Two  EICs were placed
side-by-side  with a log  sheet for recording opening and closure  of  the EICs.
One  EIC was marked "day"  (occupied time)  and the other  EIC was marked "night"
 (unoccupied time).  School personnel opened and closed the EICs and kept records
on a log  sheet indicating  dates and times of openings  and  closures of the
respective EICs.  This part of Phase  2  was conducted  in February and March 1989.
The  testing in  the Administration building,  replicated in March  1989,  showed
the  pattern seen in Figure 3.  The Olmstead building radon concentration data
are  shown in  Figure  4.

Phase  3

      All rooms in the Ballard School System's Kelley  Elementary School were
measured   for  radon  concentration as  described in  Phase  2.    These  data are
presented in Figure  5.

-------
DISCUSSION

      A total  of 196 classrooms were  tested in the Urbandale  School System.
Figure 1 shows data from 13 rooms with radon concentrations greater than 4 pCi/L
during  the  weekdays but less  than 4 pCi/L  during  the weekend.  The weekend
operation of  the mechanical system was  the  same  as during the  week,  and the
classrooms  in  the schools are  generally unoccupied during the  weekend.   The
weekday testing period includes night and day (24 hour weekdays) in this plot.
Students and teachers are in school during the weekdays; therefore, the weekday
testing period may better represent estimated exposure to radon concentration.
If the  weekday testing period is more predictive of  occupational  exposure to
radon  than  the  weekend  testing period,  16 false  negatives  would have  been
indicated by weekend testing only.   Likewise, 6  false positives would have been
indicated by testing on the  weekend only (see  Figure  2).   The data in Figures
1  and 2  (16  false negatives  and  6 false  positives,  respectively)  show the
uncertainty introduced when using either time period  for predicting occupational
exposure to radon concentration.

      Figure 3  shows  data from the Administration  building tested  during the
night  (unoccupied  time)  and day (occupied time).    Four  of  seven rooms showed
substantially greater radon concentration during the occupied time than during
the unoccupied time.  A replicate measurement showed a  similar pattern of radon
concentration response.

      Figure 4 shows data from the  Olmsted classroom building tested during the
night (unoccupied time) and day (occupied time).  Unoccupied time (night) radon
concentration  values  were   greater   than  the  occupied  time  (day)  radon
concentration values in this school.

      Figure 5  shows that  five of  six  rooms had a greater radon concentration
during  the occupied time than during the unoccupied time.  (The sixth room was
a second floor  6th  grade  room that had its windows  open during the day.  Note
that  typically  second  floor classrooms would not be  tested;  however,  the 6th
grade science teacher wanted to test that room.) A replicate measurement of all
rooms conducted in April 1989 during relatively warm weather was similar but with
less exaggeration of the differences.  The Kelley Elementary School has a boiler
located at one end of the building  with a tunnel that extends under some of the
rooms.   The temperature setback at the end of the day and the setup  at the
beginning of the day may  explain why the radon concentration increases during
the day (warm soil-gas laden  air  in the tunnel may rise  into the classrooms
above).

                       MEASUREMENTS IN MARYLAND SCHOOLS

      The radon measurements in the Maryland schools were made with EICs and a
CRM for comparison.  One school is  in Prince Georges County and the other is in
Washington County.  Both schools are currently being researched as part of EPA's
radon mitigation research program  in schools.

-------
METHODOLOGY

      The radon measurements were made in one  room in each of these two schools
and covered a  2-week period.   A CRM  collected hourly  radon measurements over
the entire 2-week period; one EIC was exposed during the entire 2-week period;
one EIC was exposed only during the day on Monday through Friday; and finally,
one EIC was exposed  only during  the  night on  Monday through Thursday.   School
personnel were instructed on opening and closing of the EICs and kept a log of
exposure times.  The measurements were made in late May 1989.

RESULTS AND DISCUSSION

      The results measured in the Prince Georges County classroom are displayed
in Figure 6.  Note that radon levels  rose sharply at night and decreased during
the daytime.  The actual radon exposure during the daytime was 2.1 and 2.5 pCi/L
for the CRM and EIC,  respectively.  The nighttime exposure for both devices was
4.4 pCi/L.  The  average radon level  for  the  entire  2-week  period was 4.1 and
4.2  pCi/L  for  the  CRM and  EIC,  respectively.   These  data  show  excellent
correlation between  the CRM and EIC.

      Comparing  the  daytime (only),   nighttime  (only),  and  cumulative (total)
exposures  during  the  2-week  period  indicates   that a   measurement  device
integrating radon  levels over 24-hour periods would overestimate actual radon
exposure (i.e., produce a false positive).

      Figure  7  shows CRM and EIC  data from an office  in a Washington County
school.  The  radon level measured during  the daytime  periods was 2.1 and 6.2
pCi/L for  the CRM and EIC, respectively.   The  nighttime  measurements for the
CRM and EIC,  respectively, were 2.0  and  4.1 pCi/L; and the  average radon level
for  the entire  2-week period  was  2.0  and 4.1  pCi/L for the CRM  and EIC,
respectively.

      The daytime  and nighttime comparisons for these  data  are  less conclusive
than those  shown  in  Figure  6, since  the  EIC measurements range  from 0.3 to 4.1
pCi/L greater  than the CRM  averages.   In any  event, had the radon measurements
been  collected with an integrating  device over  the  2-week period,  a false
negative would have  been produced.

                     MEASUREMENTS IN NORTH CAROLINA SCHOOL

      A school in Forsyth County that is currently part  of EPA's radon mitigation
research program was selected for  the EIC measurements in North Carolina.  The
classrooms  in this school are located in three pods connected to a central media
area.   There  are eight classrooms   in  each  pod.    Initial charcoal canister
measurements  during  the summer of 1989 with the HVAC system  off showed that all
of the 24 classrooms  in the pod area ranged from 3.5 to 10 pCi/L.  The building
has  a central  heating, ventilating,  and air-conditioning (HVAC)  system and
appears to operate at a very slight positive or neutral pressure relative to the
outdoors.

-------
METHODOLOGY

      Four of the eight classrooms in each of the three pods were measured for
1 week.  The data were collected in late October and early November 1989.  One
EIC was exposed  for  the entire  week;  one EIC was exposed only during the day;
and finally, one EIC was exposed only during the night.  School personnel were
instructed on opening and closing the EICs and kept a log of exposure times.

RESULTS AND DISCUSSION

      Figure  8  shows the  results  of  the  integrated,  daytime,  and nighttime
measurements in each of the classrooms.   (Note that data  for 11 of the 12 rooms
are presented since the data were not  recorded properly in the log sheet in one
of the classrooms.)   The  averaged data for each of  the  pods  are presented in
Figure 7.  The overall averages for the classrooms are relatively close during
all  three measurement  periods  averaging  6.0,  5.4, and  5.8  pCi/L for  the
integrated, daytime, and nighttime periods, respectively.

      Although these  averages  are  very close, some  trends can  be  observed in
Figure 8.  Looking at the  daytime  versus nighttime  radon levels for 10 of the
rooms (they are nearly the  same  for Room 36),  the nighttime level  is higher than
the daytime level  in 7  of the  10  rooms.   The averaged  data  in  Figure  9 also
follow this trend.   As mentioned  above,  the nighttime  average  was 0.4 pCi/L
greater than the  daytime average for all the classrooms. This is consistent with
the understanding that the building is  operated under a slight positive pressure.
However,  when radon levels build up overnight, they are not sufficiently diluted
during the daytime.   Inspection  of the HVAC system revealed that the outdoor air
controls for most of the multi-zone air handlers  appeared to be operating in an
economizer mode with no minimum outdoor air position.

      Except for one notable excursion (daytime levels in Room 25 were about 11
pCi/L  above  nighttime  levels),   the   integrated,  daytime,   and  nighttime
measurements in this school were relatively close.  Based on these results, if
only integrated screening measurements  had been  made in  this  school,  only one
false positive and no false negatives would have been produced.

                                 CONCLUSIONS

      The following  conclusions are based  on the EIC measurements  made in 10
Iowa,  Maryland,  and North  Carolina  schools  under occupied and  unoccupied
conditions.   Their  applicability to  other schools  will  depend  on  the  unique
characteristics  of each school.

      (1)  Although  there  was  generally  good  cooperation   from   the  school
           personnel  in the opening  and closing of the EICs and  logging of
           times,  this  method  is not flawless  and  introduces  additional
           uncertainty into the measurements.  Depending on the school,  it may
           be better to have one person open and close all the EICs.

      (2)  Building-specific responses  during occupied and unoccupied periods
           were  found  in  most  cases.    Some  schools  had  greater  radon
           concentrations  during  the  occupied  time  (day)   than  during  the

-------
           unoccupied  time  (night),  and  the  converse was   true  of  other
           buildings.   Consequently,  radon measurements that integrate levels
           over a 24-hour period have the potential to produce  false negatives
           or  false positives.
      (3)  Operating the building for 48 hours as it is operated during school
           hours  as recommended  by  EPA  should most  closely  correlate  with
           student exposure.

      (4)  Where a CRM was placed side-by-side with the EICs, the correlations
           were generally  good (Maryland schools).   However,  some  of the EIC
           measurements in Washington County were consistently higher than the
           CRM measurements during  the same period.

      Some potential advantages and  disadvantages of EICs are apparent from this
study and prior experience with EICs.  Advantages include:  (1) EICs are supposed
to be an integrating detector unlike charcoal which weights  the  later exposures
more than the  earlier, and (2) EICs can be opened and closed to start and stop
their integration  and enabling actual exposures  to be  estimated.   Potential
disadvantages  include:  (1) Since EICs are calibrated  based on an equilibrium
of radon and progeny, their response  to radon concentrations is not instantaneous
(this lag is  thought  to be  similar to  that of  the  CRM:  about  0.5 hr for
equilibrium with radon and about 4 hr for that radon to reach equilibrium with
progeny.  (2)  If the radon levels are varying significantly, the CRM might be
expected to follow changes better than an EIC that is opened  and closed each day
and has a start-up period each  day while  the CRM  is open  all the time.  (3) Can
school personnel be relied upon to record the daily openings accurately enough
to allow  the EIC exposure average to  be  computed?   (4) Will frequent handling
affect the EIC readings (this might be expected to increase the voltage losses
and lead to overestimates of radon levels)?  (5) Are  the changes  in school radon
levels so rapid when school starts in the morning and the HVAC systems are turned
on that this type  of measurement is  invalidated (due to the  decay delays in the
CRMs and EICs since they depend on an equilibrium  calibration)?   (6) If exposure
is to be  estimated for occupied time, how  many exposure  days must be averaged
to achieve enough EIC voltage drop to give an accurate estimate  for radon levels
around 4 pCi/L? Hopefully,  these questions will be addressed in future research.

                                  REFERENCES

1.    United States Environmental Protection Agency, Office of Radon Programs,
      "Radon Measurements in Schools  -  An Interim Report." Washington,  D.C.
      20460,  EPA-520/1-89-010,  NTIS PB89-189-419, March 1989.

2.    United  States  Environmental   Protection Agency,  Office  of  Radiation
      Programs, "Indoor Radon and Radon  Decay  Product  Measurement Protocols,"
      Washington,  DC, February 1989.

-------
I
1
                           6   7   8  9   10  11   12
                            Urbandale school rooms
13  14
16

•
Weekend ^ Weekday

Weekend Weekday
Rn Rn
concentration concentration
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

Admin. Office
Admin. Office
Admin. Office
Admin. Office
Admin. Office
Admin. Office
Admin. Office
Admin. Office
High School
High School
Jensen
Jensen
Jensen
Jensen
Karen Acres
Rolling Green
Room pCi/L
Account. & Pay. 3.1
Business Man. Off. 3.6
Dale Y. 2.7
Lounge 1 .7
Maintenance 2.0
Shelby Off. 1.5
Storage Room 3.6
Women's Bathroom 2 4
104 4.0
606 2.6
310 2.0
312 < 0.5
317 
-------
        1   2 3  4  5  67  8  9 10 11 12 13 14 15 16 17  18 19 20 21 22  23
                            Urbandale school rooms
NO.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Building
High School
High School
High School
High School
High School
Middle School
Middle School
Olmsted
Olmsted
Olmsted
Olmsted
Olmsted
Olmsted
Olmsted
Rolling Green
Rolling Green
Rolling Green
Rolling Green
Rolling Green
Rolling Green
Rolling Green
Rolling Green
Rolling Green
• Weekend E
Room
101
106
305
602
614
Oust Break Rm
Vocal Music
22
23
Kitchen
Lounge
Marion S.
Mrs K
Music
Kitchen
Librarian
Lisa's Desk
McTaggert
Mrs. Ennen
Mrs 1.
Mrs O.
Mrs S.
Music Room
2 Weekday
Weekend
Rn
concentration
pCi/L
36
4.7
3.5
33
34
33
33
127
200
84
9.6
3.8
5.6
6.6
3.5
3.5
3.1
40
31
31
41
4.8
38
Weekday
Rn
concentration
pCi/L
1 7
1.3
09
21
30
06
<0.5
103
104
64
<0.5
31
21
1 5
16
25
1 3
<05
09
<05
09
1 t
1 0
Figure 2. Rooms  from buildings in the Urbandale  school system with
radon concentrations greater on weekends than    weekdays.

-------
                       Urbandale administrative rooms
                               Night
Day
NO.
1
2
3
4
5
6
7
Room
Accounts Payable
Business Management Off ice
Lounge
Maintenance
Shelby's Office
Maintenance Office
Outer Maintenance Office
Nighttime
Rn
concentration
pCi/L
2.0
2.5
2.9
5.0
5.2
7.1
6.0
Daytime
Rn
concentration
pCi/L
5.2
7.9
6.1
7.7
5.2
1.3
4.8
Figure 3. Night (unoccupied time)  and  day (occupied time) radon
concentrations in the Urbandale  school system Administration building.

-------
                       Olmsted school rooms
                            Night
E3 Day
No.
1
2
3
4
5
6
7
8
Room
22
23
Kitchen
Lounge
MarionS.
MarionS.
Mrs. K.
Music
Nighttime
Rn
concentration
pCi/L
6.2
11.2
9.3
3.0
2.7
1.9
3.3
2.3
Daytime
Rn
concentration
pCi/L
4.7
B.6
6.4
1.4
n.a.
n.a.
<0.5
1.3
Figure 4. Night (unoccupied time) and day (occupied time) radon
concentrations in the Urbandale school system's    Olmsted building.

-------
                          Kelley elementary rooms
                              Night
Day
NO.
1
2
3
4
5
6
Room
1st Grade
3rd Grade
6th Grade
Kindergarten
Library
S.E. room
Nighttime
Rn
concentration
pCi/L
2.1
2.0
1.8
<0.5
1.4
1.0
Daytime
Rn
concentration
pCi/L
3.3
11.8
<0.5
3.8
5.5
1.8
Figure 5. Night (unoccupied time) and day  (occupied  time)  radon
concentrations in the Ballard school system's Kelley Elementary building

-------
   10
    8  -
    6  -
§
•o
    4  -
    2  -
Total  Period:
Weekday Period:
Weeknight Period:
     138
         142
146
                                    1989 JULIAN DAY
150
154
              Weekdays when EIC was  open
              Weeknights when EIC was  open
         Figure r.   CRM and EIC results  for  Washington County school

-------
  15
  12
hJ
•H

CX
I  6
                                      Time
                                     (days)
             Total Period:
             Weekday Period:
             Weeknight Period:
                  14.1
                   4.0
                   4.0
                                   CRM
                                 (pCi/L)
                    4.1
                    2.1
                    4.4
                                   EIC
                                 (pCi/L)
   4.2
   2.5
   4.4
     139
142
 145            148

1989 OULIAN DAY
151
154
            Weekdays  when  EIC  was open
            Weeknights  when  EIC was open
                Figure 6.   CRM and EIC results for Prince Georges  County  school.

-------
i
                                              27
                   31
      SB
37
SB
            INTEGRATED
ClassrooH Na.
    DAYTIME
EZ3 NIGHTTIME
       Figure 8. EIC measurements in North  Carolina school  under
                  occupied  and unoccupied  conditions.
i
            INTEGRATED
  Pod No.
   DAYTIME
                                                              ALL ROOMS
V771 NIGHTTIME
       Figure  9.  Summary  of EIC measurements in each  pod of
                  North  Carolina school.

-------
                                                         D-IX-6
                MEASURING RADON  IN THE WORKPLACE

                          Michael Boyd
              U.S.  Environmental Protection Agency
                  Office of Radiation Programs
                      Washington, DC 20460

                    Terry Inge, John MacWaters
                    Sandy Cohen and Associates
                        McLean,  VA 22101
                            ABSTRACT

     The Environmental Protection Agency (EPA) has issued
guidance for testing for radon in homes and interim guidance for
testing in schools.  Information on testing for radon in the
workplace is the next initiative and this paper describes the
current status of this effort.  The results of measurements made
in several buildings in the Washington, DC area are discussed.  A
discussion of preliminary guidance on radon survey design that
has been offered to Federal agencies is presented .

     This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review
policies and approved for presentation and publication.

-------
                           INTRODUCTION

     The discovery of dangerous levels of radon gas in homes in
the early 1980's prompted the Environmental Protection Agency
(EPA) to issue guidance for testing and mitigation of this
serious health threat.  Since most Americans spend a majority of
their time at home, EPA decided to concentrate its resources to
encourage people to test their homes for radon and to take
corrective action where appropriate.   In 1988, EPA was able to
extend its radon program to address the hazard of radon in
schools.  Since children may be at higher risk from exposure to
radon, this initiative was the most appropriate next step from a
public health standpoint.  As our understanding of the risk from
indoor air pollution has grown, the predominance of radon as the
most serious indoor air pollutant has become obvious.  It is for
this reason that EPA is now preparing guidance for testing for
radon in the workplace.  Congress has included radon in the
workplace as an area of concern.  Sections 118(k) and 403(b) of
the Superfund Amendments Reauthorization Act (SARA) instruct the
agency to undertake testing and research in this area.

     In addition, Section 309 of the Indoor Radon Abatement Act
(Public Law 100-551) requires the heads of all Federal
departments and agencies that own buildings to test a
representative sample of those buildings for radon and report
their results to EPA's Administrator.  A report to Congress on
these results is then due in October, 1990.  Preliminary
information on the design of an effective radon survey has been
provided to Federal agencies to assist them in completing this
task.  EPA expects that the information gained from these surveys
will significantly enhance the present understanding of radon
behavior in large buildings.

     A study was undertaken in the winter of 1989 to perform
intensive measurements in several buildings in the Washington, DC
metropolitan area.  Permission was obtained from the General
Services Administration  (GSA) to use three of their buildings for
this study.  The buildings were chosen to represent differing
ages, construction types, and sizes.  This work provided
confirmation of several phenomena observed in homes including
diurnal cycling of radon concentrations and dependance of radon
levels on the operating parameters of the ventilation system.
Information on radon movement throughout upper floors of
buildings was not obtained because the observed radon
concentrations were too low.  The three GSA buildings, all
located in downtown Washington, DC, had ground floor averages
below 1 picocurie per liter  (pCi/L) which made detection on upper
floors impossible.  This study is continuing this year.

-------
                      MATERIALS AND METHOD

     Federal agencies were required to submit their radon survey
designs to EPA by July,  1989.  The surveys are limited to radon
screening measurements.   Although follow-up measurements and
mitigation are not addressed, these components will almost
certainly be undertaken when elevated concentrations of radon are
reported.  Most of the agencies have chosen to perform 3-month
measurements using alpha track detectors placed during the winter
heating season.  The next most popular device is the diffusion
barrier charcoal canister which may be deployed for 2 to 7 days.
Electret ion chambers and continuous monitors are also being
considered by some agencies.  EPA's policy is that any
measurement device listed in the Indoor Radon and Radon Decay
Product Measurement Protocols (EPA 520/1-89-033, March 1989) is
acceptable for making screening measurements except that grab
sampling may not be used alone.  Grab sampling can be a useful
confirmatory measurement made in conjunction with a measurement
from some other approved device.

     As an independent initiative, EPA is investigating radon
behavior in large buildings in preparation for issuing interim
guidelines for measuring radon in the workplace.  This effort
will include a variety of building types and sizes.  The phase of
this project that was completed in 1989 included initial
screening measurements made with 2-day open faced charcoal
canisters.  Continuous radon measurements were made using
continuous radon gas monitors.  The data collected in the 3 GSA
buildings was supplemented with the results of extensive 3-raonth
alpha track measurements made by GSA as part of its radon program
begun in 1989.  Other data collected included differential
pressure measurements, meteorological data, and information on
the layout, type, and operation of heating, ventilation, and air
conditioning (HVAC) systems.

     This project is continuing this season.  The prospect for
locating Federal buildings with elevated radon levels is improved
because of the availability of the data from GSA's radon program.
It is anticipated that seasonal and annual measurements will be
taken as part of this study.

-------
                      RESULTS  AND DISCUSSION

EXPERIMENTAL DATA

     Data from the 6SA buildings include charcoal canister
measurements and alpha track results from all 3 buildings and
continuous monitoring data from one building.  A summary of the
ATD results is presented in Table 1.

                             Table 1
                   ATD Results in GSA Buildings

Radon Concentration        Number of Readings at Each Level
	fpCi/Ll	       Blda. 1	Blda. 2	Blda. 3


     0 - 0.5               12             50             77

   0.6 - 1.0               13             26             30

   1.1-1.5                3             14              1

   1.6 - 2.0                021

     The charcoal canister results in the three buildings gave
readings that were for the most part in the 0-0.5 pCi/L range.
A few readings were obtained in the 0.6-2 pCi/L range with the
highest of these readings being located on the lowest level of
the building.  One reading of 3.3 pCi/1 was obtained in a
crawlspace under one of the buildings, confirming the generally
low source potential for radon in the GSA buildings that were
tested.

     Continuous monitoring of radon in Building 3 exhibited a
diurnal pattern of radon concentrations.  During the weekdays, a
rise in radon concentration was noted beginning around 6 pro.  By
2 am, the radon concentration was at its maximum where it
remained at this plateau or oscillated until 6 am.  At 6 am on
weekdays, the radon concentration dropped rapidly to under 1
pCi/L by 10 am where it remained until 6 pm.   This pattern is
explained by the operation of the HVAC system.  The air handlers
in the area under study are turned on at 6 am and shut off at 6
pm on weekdays.  In spite of a slight negative pressure created
by running the system, the air handlers are responsible for
decreasing radon concentration because of fresh air dilution and
mixing.  On the weekends, when the HVAC system is not operating,
the peak concentration remains relatively stable until Monday
morning when the system is restarted.

-------
FEDERAL GUIDANCE

     Based on the preliminary results of this study and input
from technical experts, information was distributed to Federal
agencies to assist them in preparing their radon survey designs
for complying with the Indoor Radon Abatement Act.  It was
decided to recommend that all ground-contact occupiable rooms be
tested.  In addition, it was suggested that a few detectors be
placed on upper levels in stairwells, near elevators, and beside
service shafts and ductwork.  Radon is principally a soil gas
problem and testing ground contact rooms should indicate the
highest radon concentrations in the building.  Movement of radon
throughout a building should be indicated by the upper level
locations that were specified.  Private water supplies are also
required to be tested which increases our confidence that we can
identify a radon problem if one exists.

     Federal agencies were advised to perform screening
measurements during the winter heating season.  In warm climates
where there is no heating season, it is recommended that testing
be done during the period of minimal outside air intake to
minimize dilution flow.  For colder climates, the stack effect is
assumed to be the principal driving force for radon entry.  The
stack effect occurs during the colder months when the effect of
rising warm air creates a negative pressure in the lowest levels
of a building relative to the air outside the building.  This
negative pressure causes radon-rich soil gas to be sucked into
the building.  In commercial buildings, large negative pressures
can also result from unbalanced HVAC systems.  The assumption
that winter is the time to test is being investigated as part of
EPA's ongoing research efforts.

     The suggested time to test is during the work week under
normal building conditions.  The HVAC system should be operated
normally. While screening conditions are chosen to produce
results higher than the annual average, there is no benefit to
testing under conditions or during periods that do not reflect
conditions present when the building is occupied.

     In the situation where an agency decides not to test all of
its buildings, EPA's recommendation is to test fewer buildings
according to the above guidelines rather than to test more
buildings with only a few detectors per building.  While the
latter option may produce a statistically valid estimate of the
percentage of Federal buildings with elevated levels of radon, it
cannot certify that any individual building is safe.  Experience
has shown that even adjacent rooms can have dramatically
different radon levels.

     Finally, all Federal departments and agencies were strongly
advised to put in place a quality assurance program with a strong

-------
quality control component.  For radon measurements made with
passive detectors, a good quality control program will include an
appropriate number of duplicate and blank detectors as well as a
few spiked detectors.  Duplicate detectors are collocated and
provide an indication of the precision of the analytical
laboratory.  Blank detectors serve as a check on the background
of the detectors.  Spiked detectors, those exposed to a known
concentration of radon, provide an indication of accuracy or
bias.

                           CONCLUSIONS

     Ongoing research has tended to support assumptions about the
importance of building conditions to the potential for radon
entry.  Research to confirm that winter testing will yield
maximum values is continuing.  Experience has shown us that a
balanced HVAC system, particularly one that contributes to a
slightly positive differential pressure in rooms in ground
contact, will minimize or eliminate radon entry.

     Yet to be determined is the extent of radon movement
vertically within large buildings.  Testing has been encouraged
in stairwells, near elevators, and near service shafts.  It is
hoped that the extent to which radon may be a problem on upper
levels of buildings can be determined from the current phase of
EPA's work in this area.

     Because the Congress required Federal agencies to sample a
representative number of their buildings this year, EPA will have
a large amount of data to assist in confirming the preliminary
guidance for testing in the workplace.  Interim guidelines for
measuring radon in the workplace will likely be available by the
end of this year.

-------
                                                                   D-IX-7
                         THE SCHOOL EVALUATION PROGRAM
                  by:    Eugene  Fisher
                         Florence Blair
                         U.  S. Environmental  Protection Agency
                         Office  of Radiation  Programs
                         Washington,  D.  C.    20460

                         Terry Brennan
                         Camroden Associates
                         Oriskany,  New York  13424

                         William Turner
                         Harriman Associates
                         Auburn,  Maine  04210
                                   ABSTRACT

      A pilot program  to provide  classroom and field training to school
facility operators was  implemented by  the  U.  S.  Environmental Protection
Agency's Office of Radiation  Programs  in 1989.   This program consisted of two
phases.  The first phase developed and delivered a three day workshop in
Nashville, Tennessee.   As  a result of  the  workshop,  a second phase was
initiated.  The second  phase  investigated  several school buildings with
elevated indoor radon  levels  in the Western United States.   Radon entry
mechanisms were identified.   Measurements  to  evaluate soil  depressurization as
a radon control method  were made  and HVAC  systems were characterized.
Measurements were made  to  evaluate HVAC modification as a radon control
method.  Building shell tightness measurements were  made and information was
collected to judge the  suitability of  potential  sites for additional EPA
sponsored "hands on" school training.   Physical  and  institutional problem
areas were identified.
ft U.S. SOVEMfCNT PRINTING OFFICE:  1990 748-010/25004

-------