Parameters Impacting the Emissions of Selected VOCs from
               the Toner for a Specific Photocopier
                                   by
                             D. Bruce Henschd
                         Risk Management Research Laboratory
                     U.S. Environmental Protection Agency
                       Research Triangte Park, NC  277 ] 1

               Roy C. Fort man n, Nancy F. Ro;idie, and Xiaoyu Liu
                       ARCAD1S Gersghty & Miller, Inc.
                             P.O. Box 13J09
                       Research Trlanftlp Park Nf  27709
                             tor presentation at
          Symposium on Engineering Solutions to Jndoor Air Quality Problems

                             sponsored by the
                     U. S. Environmental Protection Agency
                                 &nd the
                     Air & Waste Management Association
                               Raleigh, MC
                             July 17-19,2000

-------
Parameters Impacting tKe Emissions of Selected VQCs from
the Toner for a Specific  Photocopier
B- Bruce Tienschel
National Risk Management Research Laboratory, U.S. Environmental Protection Agency,
Research Triangle Park, NC  2771 1
Roy C\ FpttmanOj Nancy F. Roadie, arul Xmoyu Liu
ARCADIS Geraghty & Miller, Inc., P. 0. Box 13109, Research Triangle Park, NC 27709
ABSTRACT
A laboratory ihermat desorpiion apparatus was used to measure emissions from a number of
nominally identical photocopier toners - manufactured for use in a specific model copier- when
these toners were heated 10 fuse-r temperature (180 - 200 °C).  The objective was to determine
whether VOC emissions from a toner might be reduced th rough judicious selection of the process
and the raw polymer feedstock used in its manufacture. Tests were performed on: a series of
toner nnd fecdslock sampies obtained directly from a toner manufacturer, wtfematicatly varying
process and feedstock; and toner cartridges - from different lots (for which process and feedstock
were unknown) - purchased from local retailers.  The  results showed that the retailer toners
consistently Had up to 350% higher emissions of some major compounds, and up to 100% lower
emissions of others, relative to the manufacturer toners (p<0.05), probably due to differences in
process and/or feedstock. The manufacturer toners showed psspjirialiy nrs effect of process or
feedstock, probably because the two processes and two feedstocks used by the manufacturer were
not significantly differeni From each other. It is concluded that process and feedstock can have  a
significant effect on emissions of individual compounds, but it is not possible from this study to
make specific recommendations regarding how process or feedstock might be modified to
produc« lower-emitting toners for a given  topier.

INTRODUCTION
Dry-process photocopiers (and laser printers and  fax machines) are ubiquitous potential sources
of volatile organic compounds (VOCs), ozone, and participates in indoor air  The VOCs result
primarily from the organic polymers  present in the toners used in these machines, While low
VOC emissions can be observed even when the copier is off or idling, most emissions occur
during copying, when the toner  polymer is being melted in the high-temperature {user that fixes
the copied image onto the paper1 -.  Some VOCs (e.g.» acetaldehyde and hexanal) have been
observed in emissions from plain paper, in the absence of toner (unpublished data); but the toner
is the predominant source of most VGC$.

Polymers commonly present m commercial toners are copolymers of styrene and acrylates, the
resins most common for copiers utilizing hot-roll  (heat and pressure) fusers1 '• *  Upon  heating to
(user temperature, these polystyrene/polyacrylate resins have consistently been found1- *•*•'' to emit
styrene, ethylbenzene, xylenes, and acetophenone - all of which are IflhdeH HS hazardous air

                                         1

-------
 pollutants (HAPs)8 - as well as benzaldehyde, an array of other substituted benzenes, and a
 number of other VOCs. The precise composition of the proprietary styrene/acrylate copolymer
 that is used in a given toner product - and the nature of the pigment and the other additives in the
 toner - will vary between brands and models of copiers, depending upon machine requirements.
 Differences in the design and operating characteristics of different machines - e.g., fuser
 temperature - might also impact emissions. Even for a given model copier, emissions might be
 impacted by the toner manufacturing process and by the purity of the raw materials used in toner
 manufacture. Perhaps for these reasons, chamber tests on different copiers using polystyrene/
 polyacrylate toners have shown different VOC emission factors1*2.  And thermal desorption tests
 on toners from different machines have shown some variation in the specific VOCs that are
 released upon heating6.

 In prior testing1, three different lots of nominally identical toner - manufactured for one specific
 copier - were heated to 150 °C in a vial, and the headspace gas in the vial was analyzed for
 styrene, ethylbenzene, and xylenes.  For each compound, the headspace concentrations obtained
 with one of the toner lots were 50 to 80% lower than with the other two lots. The toner supplier
 indicated that the lower-concentration lot had been manufactured using an extrusion process, the
 process currently favored for new production facilities. By comparison, the other two had been
 manufactured using the older batch mixing approach (employing Banbury mixers) for melt-mixing
 the polymer,  pigment, and charge control additives to generate the final toner product.  It was
 postulated that the difference in manufacturing process might be responsible for the difference in
 headspace concentrations.  Since the three toner lots had  been manufactured  in plants that were
 widely separated geographically, the raw materials used in manufacture could well  have been
 obtained from different resin suppliers, and it is also possible that the concentration difference
 could be due in part to differences in the purity of the polymer feedstocks.

 The objective of the current study was to expand upon that prior work. In particular, it was
 desired to determine more rigorously the extent to which VOC emissions from a given toner -
 manufactured for a specific dry-process photocopier - might be reduced through judicious
 selection of the process and the polymer feedstock used in the manufacture of that  toner.

 TECHNICAL APPROACH AND PROCEDURES

 Basic Approach

 The Selected Pit otocopier
 The specific photocopier selected for study is a 70 copy/min monochrome unit using a heat and
 pressure fuser, widely marketed in.the U.S. The fuser nominally operates at 185 °C; a warning
 light illuminates if this temperature drops below 180 °C, and the machine shuts down if it exceeds
 200 °C. The  Material Safety Data Sheet for the toner indicates that it is 80 to 90% proprietary
 styrene/acrylate polymer (binder), 10 to 15% carbon black (pigment), 1 to 5% quaternary
 ammonium compound (charge control agent3'5), and 1  to 5% polyolefin (wax to reduce adhesion
to fuser rollers4'5). This is a typical composition for dry toners used with hot-roll Risers.

 Toners Obtained from a Manufacturer
A cooperating toner manufacturer produced four nominally identical batches  of toner for the
 selected copier for this project, according to a 2x2 matrix: using two different manufacturing

-------
 processes; and using two different lots of polymer feedstock.  The two processes included:
 vented (or "devolatilizing") extrusion, in which there is a low-pressure relief port along the barrel
 surrounding the extruder screw, to allow removal of volatiles released during the extrusion
 process9'I0; and unvented extrusion, in which there is no such relief port.  (It had initially been
 hoped that toners might also be obtained using a third manufacturing process - batch mixing - but
 the manufacturer did not produce toners for the selected copier using that process.) The two lots
 of polymer feedstock - referred to as "Feedstock A" and "Feedstock B" - were presumably
 obtained from the same resin supplier, and were probably produced at about the  same time.

 During each of the four production runs, the manufacturer provided two or more toner cartridges
 that were filled at each of three different  times during the run.  Samples were also provided of the
 unprocessed feedstock that was being fed to the extruder.

 Using these samples, a systematic matrix of 44 thermal desorption tests was conducted to
 determine VOC emissions from the toners as a function of:  manufacturing process; feedstock;
 sampling time during the manufacturing run ("within-run variability"); toner cartridge, for
 cartridges filled at the same time during a run ("between-cartridge variability"); and sample
 location within a given cartridge ("within-cartridge variability"). This matrix also enabled
 comparison of the toner emissions against emissions from the corresponding unprocessed
 feedstocks.

 As it turned out, the toners produced by  vented extrusion were manufactured with the relief port
 operating at a vacuum only 2 Pa below atmospheric, due to operating problems.  This vacuum is
 so small that one might expect  only limited  extraction of volatiles from the vented process, and
thus limited differences in emissions from the toners produced by vented vs. unvented extrusion.
 Contacts with several individuals in the polymer industry - and available literature on polystyrene
 devolatilization" - suggest that vacuums more than 4 orders of magnitude greater than this might
 be required for effective removal of VOCs from this copolymer, based on  diffusion
considerations.

 Toners Obtained from Local Retailers
In addition to the toners obtained directly from the manufacturer, four cartridges of toner for the
selected copier were purchased from local retailers.  The manufacturer, the manufacturing
process, and the feedstock characteristics represented by these retailer toners were unknown.
Thus, it would not be possible to analyze these parameters and to rigorously explain any
differences in the emissions between these toners and the manufacturer toners. However, the
retailer toners provided an opportunity to demonstrate the magnitude of the difference that toner
manufacturer, process, and/or feedstock might make in the VOC emissions from  a given copier.

Two cartridges, representing two different manufacturing lots, were purchased from one retailer.
Two additional cartridges were obtained  from a second retailer; both cartridges from this second
retailer were from the same lot, different  from either of the lots obtained from the first retailer.

Thermal desorption tests were run on two samples from each of these cartridges.

-------
 Thermal Desorption Test Method

 The toner samples described above were tested using a thermal desorption procedure developed
 for this project.  In this procedure, a sample of the toner was ballistically heated to the fuser
 temperature operating range for the selected copier (180 to 200 °C), and the resulting VOCs
 captured either on Tenax* or on dinitrophenylhydrazine (DNPH)-impregnated silica gel.  Only the
 results with Tenax® are reported here.

 A thermal desorption approach was selected over a headspace procedure as the method for
 quantitative screening of the large number of toner samples in this study. Flow-through thermal
 desorption tests avoid vapor pressure constraints inherent in static headspace tests, and thus
 should more accurately reflect the total potential mass of the individual toner VOCs that might be
 emitted from a copier. Thermal desorption testing also reduces the risk of artifact formation, that
 might occur when samples are heated for the extended periods commonly employed in headspace
 testing. But it must be recognized that the mass emissions measured using this thermal desorption
 screening approach cannot currently be related to the emission factors that would be observed
 from an actual copier.

 The ballistic heating approach used here could not duplicate the temperature profile that would
 exist in a fuser, where the thin layer of toner on the paper is raised to temperature in milliseconds3.
 However, it might reveal some effects that the rapid heating could have on emissions. Initial
 efforts were devoted to demonstrating the reproducibility of this test method.

 The thermal desorption test  apparatus is illustrated in Figure 1, set up for sampling on Tenax®,
 The sample of toner or feedstock  powder (10.2 ± 0.4 mg) was held in place inside a 6 mm o.d.
 quartz tube using glass wool plugs, The tube was mounted inside two electrical heating elements,
 totaling 900 W, in an insulated oven.  Dry nitrogen carrier gas flowed through the tube (50
 mL/rnin), transporting the released VOCs out of the oven and into a sampling tube containing 250
 mg of 60/80 mesh Tenax* TA.  Nitrogen (N2) was used as the carrier, rather than air, to avoid
 possible oxygenation of the  VOCs. A thermocouple inside the tube measured the temperature at
the location of the sample,

During a test, N2 flow would be initiated, and the heaters turned on briefly, raising sample
temperature to 190 ± 4 °C in about 30 sec. Both heaters were turned off at the proper point
during this heat-up period (after 5-6 sec), to prevent the sample from overheating. Flow
continued for an additional 90 sec, during which time the sample cooled to 145 ± 3  °C. The
Tenax* sample tube was then taken off-line, after a total sampling period of 120 sec. It was not
possible to control the heating with sufficient sensitivity to hit the desired peak temperature of
 185 ° C on each run.  If the peak temperature reached in a given test fell outside the range of the
selected copier's fuser - i.e., below 180 or above 200 °C - that sample was discarded and re-run.

Analytical Methods for Thermal Desorption Samples
The Tenax* sample tubes from the thermal desorption tests were analyzed using an Enviroehem
multiple tube desorber and an Enviroehem Model 85 Unacon concentrator/desorber, interfaced to
a Hewlett Packard 5890A gas chromatograph (GC) equipped  with a flame ionization detector

-------
   Figure 1. Thermal desorption test apparatus.
 THERMOCOUPLE -

STAINLESS STEEL
   FITTING WITH
        TEFLON
      FERRULES
                                  ^- INSULATION-
TONER
THERMOCOUPLE
6 mm x 250 mm
QUARTZTUBE
                                 ^
                                 ™	:	——	•—~-.	—	-~
                                            ~- 500 W HEATER
                                - SILANiZED GLASS WOOL
                             400 W HEATER
                                             INSULATION—7
                       x- 6mm TENAX TUBE
                            STAINLESS STEEL
                               FITTING WITH
                                   TEFLON
                                 FERRULES
                                     INSULATED BOX
                                     LIMIT
                                                                                      N2IN
                                                                                    50 ctrtmin
  (FID). Chromatography was performed with a DB-Wax column (30 m long, 0.53 mm i.d., 1 |im
  film thickness).

  For the toner tests reported here, this GC/FID system was calibrated for 11 individual VOCs,
  selected based upon separate GC/mass spectrometry (GC/MS) headspace analyses, as discussed
  below. Calibration was achieved using liquid standards loaded onto Tenax* by flash vaporization,
  at five concentrations spanning the range anticipated from the toner samples (with triplicate
  analyses at each concentration).

  Following the toner thermal desorption testing reported here, the GC/FID system was re-
  calibrated using standards that included the original 11 compounds plus 10 additional VOCs, The
  chromatograms from the current toner tests were subsequently re-evaluated to estimate whether
  any of the 10 additional compounds were present.

  Headspace Testing for Compound  Identification
  Prior to the thermal desorption testing, headspace screening tests were performed to qualitatively
  identify which individual VOCs were released from the heated toner samples. In tests with the
  manufacturer toners, 1 g of toner or feedstock was placed in a capped 20 mL sample vial

-------
 (containing air) and heated to 185 °C for either 20 or 75 minutes. At the end of that time, a 300
 p.L sample of the headspace gas was withdrawn and injected into a Hewlett Packard 5890/5970
 GC/MS (30 m Hewlett Packard 1301 column, 0.25 mm i.d., 0.25 [im film). Tests with the
 retailer toners were similar, except that 100 p,L of headspace gas was injected into a Hewlett
 Packard 6890/5973 GC/MS (30 m DB-Wax column, 0.25 mm i.d., 0.15 Jim film).

 Headspace testing was conducted on single samples of: four of the manufacturer toners
 (representing both manufacturing processes and feedstocks); each of the feedstocks; and each of
 the four retailer toners.

 The compounds identified in the manufacturer samples were used in selecting the 11 individual
 VOCs that were quantitated in the GC/FJD analyses on the Tenax® samples, discussed above.
 The compounds identified in the retailer samples contributed to the selection of the  10 additional
 compounds that were added to the subsequent standard.

 RESULTS

 Compound Identification  in Headspace Testing
 Table 1 lists the compounds that were identified by GC/MS in the headspace samples taken after
 20 min (corrected for background).  The table identifies only the compounds that had a matching
 quality greater than 85% when compared against the available computer library of mass spectra.
 In addition to these identified compounds, each sample contained two or more significant peaks
 that could not be identified.

 Essentially all of the toners and feedstocks  release ethylbenzene, xylenes, styrene, benzaldehyde,
 and acetophenone upon heating to 185 °C,  consistent with expectations and prior data1'2"6'7.  In
 addition, most of the toners release 10 or more additional compounds, with the identities of these
 additional  compounds sometimes differing between the manufacturer and retailer toners.  Some of
 these additional compounds were also detected by other investigators during extensive analysis of
 nine copier and laser printer toners6.

 It is unclear why the manufacturer toners from Feedstock A appear in Table 1  to contain so many
 fewer compounds than the toners from Feedstock B (or than the retailer toners).  The subsequent
 GC/FID chromatograms for the Feedstock  A toners,  in general, had the same number of peaks as
the Feedstock B toners, with similar retention times and area counts.

Source of the Observed Compounds
Most of the compounds observed from the manufacturer toners in Table 1  are also observed from
the corresponding feedstocks.  Thus, most  of the compounds seen in the manufacturer toners
 clearly result, at least in part, from impurities that had been present in the unprocessed
polystyrene/polyacrylate feedstock.  In some cases, the source of these feedstock impurities is
apparent.  In particular, ethylbenzene is the raw material used to produce styrene12,  so that low
levels of ethylbenzene and unreacted styrene monomer in the polystyrene could be expected.
Benzaldehyde and benzoic acid in the feedstock might result, in part, from  degradation of benzoyl
peroxide, sometimes used as an initiator in the polymerization of styrene13.

-------
Table 1.  Compounds tentatively identified in GC/MS headspace testing •  (>85% matching quality)
Manufacturer
Retailer Toners
Feedstock Toners Retailer 1 Retailer 1
Compound A BAB
Benzenes substituted with simple aliphatics
Benzene XX--
Ethylbenzene c X X X X
m,p-Xylenec X X X X
o-Xylenec X X X X
Isopropyl benzene X X - X
n-Propylbenzene - - - X
1 -Ethyl -2-methyl benzene - X
Cyclopropyl benzene - ...
Styrenec X X X X
a-Methylstyrene c X
1-Methylene propylbenzenc X X - X
Aldehydes and ke tones (often containing a phenyl group)
3,3,5-Trimelhylcyclohexanoncc X X X X
Benzaldchydec X X X X
Acetophenone c X X X X
2-Phenylpropenal XX--
Phenols
Phenol - - - X
2-(2-Propenyl) phenol - ...
Ether alcohols (often containing a phenyl group)
l,l-Dimc(hylcthoxy benzene - ...
3-Phenoxy-l-propanol - -
2-Plienoxy-l-propanol - ...
Aliphatic alcohols and di -alcohols
1-Butanol X X - X
2-Methyl-l-penlanol - ...
2-Ethyl-l-hexanol . . .
Propylene glj'col - ...
Long-chain aliphatics
Tetradecane - ...
Pentadecane -
Hexadecane - ...
Heptadecane - -
Carboxylic adds
Benzoic acid X X - X
Compounds with nvo aromatic rings
2-Ethenyl naphthalene - -
Biplienyl - ...
Bibenzyl - -
l,l-(l,3-Propanediyl)bis-benzene - - - X
1,3-Diphenyl cyclobutane - X - X
A Analysis on headspace from 1 g sample heated
B "X" = detected; "-" = not detected with >85%
Lot#l Lot #2

-
X
X
-
X
X
X
X
X
-
-

-
X
X
-

X
X

X
X
X

- .
X
X
X

-
X
X
-

-

X
-
X
-
-
to 185°Cfor

.
X
X
-
X
X
X
-
X
-
-

-
X
X
-

X
-

X
X
X

X
-
X
X

-
-
X
X

-

-
-
-
-
-
Retailer 2, Lot #3
Can. #1 Cart. #2

.
X
-
-
X
X
-
-
X
-
-

-
X
-
-

X
-

X
X
X

X
-
X
X

-
.
X
-

-

•
-
-
-
-

.
-
X
-
X
X
-
X
X
-
-

-
X
X
-

X
X

X
X
X

-..
X
X
X

X
X
X
X

-


X
-
-
-
20 min in 20 mL vial.
matching quality.
c Selected as target compounds for quantitation in subsequent
GC/FID
analyses.


-------
 Benzene is observed from the feedstocks but not from the toners. Benzene is present during
 styrene manufacture12, and thus could be an impurity in the raw polystyrene that escapes from the
 hot mixture, or reacts, during melt-mixing.

 Some compounds in Table 1 could be degradation products resulting from the heating and
 shearing of the polymer during extrusion (or from the heating during the headspace testing here).
 Among the major byproducts reported from thermal-oxidative degradation of polystyrene13 are:
 styrene, benzaldehyde, and acetophenone (observed from essentially all samples in Table 1);
 phenol (observed from all of the retailer toners); and benzoic acid (observed from some
 manufacturer toners).  Some level of these degradation products in the toners might be expected.
 The hot toner mix can reach perhaps 150 "C in the extruder3'4- 5 under high shear forces, and it has
 been demonstrated that some styrene/acrylate copolymer degradation - in terms of reduced
 molecular weight, if not byproduct formation - does occur during extrusion14.

 The presence of these potential degradation products in the unprocessed feedstocks - which have
 not been heated to extrusion temperatures - raises the question regarding the extent to which
 these observed oxygenated compounds might be an  artifact. Perhaps they could be resulting from
 oxidative degradation occurring during the headspace test procedure, when the sample is heated
 for 20 min in a vial containing air. But the subsequent thermal desorption testing in this study - in
 which the samples were heated under N, ~ also resulted in the release of these oxygenated species
 from the feedstocks. Thus, it seems  clear that these compounds are not an artifact. They are, in
 fact, present in the raw polymer, perhaps created from the polymer during polymerization or, as
 discussed previously, from degradation of benzoyl peroxide initiator.

 Extraction Tests
 To further confirm whether the observed compounds are present even when the samples are not
 heated, three selected samples of the feedstocks and toners were extracted with methylene
 chloride. The extracts were analyzed by injection  into the Hewlett Packard 6890/5973 GC/MS.
 Probably because the sub-micron carbon black in the toners is an effective sorbent for organic
 compounds, recoveries of a surrogate standard (and, presumably, of the target analytes) in the
 extracts were low. As a result, only  qualitative results were obtained - and only for the
 compounds that were present  at the highest concentrations.

 These extraction results indicated that the major compounds observed during headspace tests on
the manufacturer toners - ethylbenzene, xylenes, styrene, benzaldehyde, acetophenone, and 3,3,5-
trimethylcyclohexanone - are detected in the extracts both from Feedstock A and from a toner
produced using that feedstock. Thus, at least some portion of the emissions of these compounds
from this toner is the result of impurities that were present in the unprocessed  polymer used to
manufacture the toner. These compounds are not solely the result of thermal degradation during
the extrusion process, or during heating in the laboratory.

Selection of Target Analytes for GC/FID Quantitation
Nine of the 11 compounds that were selected for quantitation during the GC/FID analyses on
thermal desorption Tenax*' samples are noted in Table 1. The other two target analytes are m-
diethylbenzene and 2,4-dimethylstyrene. These 11 compounds were selected based upon: the

                                            8

-------
 consistency with which they appeared in the headspace results for the manufacturer samples; their
 detection in studies by others; and the availability of standards.

 The 10 additional target analytes that were subsequently added to the GC/FID standard were:
 benzene; toluene; 1-butanol; n-propylbenzene; 2-ethyl-l-hexanol; 2,3-dihydro-2-methyl faran;
 hexadecane; phenol; 2-phenoxyethanol; and phenyl benzoate. These were selected based on the
 headspace results from the retailer toners, and based on further efforts to identify unknown peaks.

 Validation of Thermal Desorption Method

 Method Precision
 To verify the precision of the ballistic heating thermal desorption technique (and of the GC/FID
 analysis), ten different samples of the manufacturer feedstocks and toners were each divided into
 two or three  10 mg aliquots, with each aliquot being ballistically heated, sampled on Tenax*, and
 analyzed for the 11 target analytes.

 Eight of the target analytes were present at concentrations above the practical  quantitation limit
 (PQL) of the GC/FID system (5 to 6 |ig of compound per g of sample). Within each of the ten
 sets of replicated samples, the concentrations of each of these eight compounds varied with a
 percent relative standard deviation (% RSD) ranging from <1 to 17%.  The average % RSD,
 averaged over all of the compounds and all of the replicate sets, was 6%. This agreement is
 considered to be good, and the precision of the method is deemed to be acceptable.

 Role of Peak Temperature
 In one ballistic heating test, the sample inadvertently reached a peak temperature of 210 °C,
 exceeding the 200 °C maximum. The test was then repeated with a replicate sample, reaching a
 peak temperature of 196 °C.  The measured emissions of each of the eight target analytes (present
 above the PQL) were 17 to 36% greater in the test that reached 210 °C, suggesting the
 importance of controlling the peak temperature.

 The peak temperatures that were reached in this matrix of tests ranged from 181 to 198 °C.
Within this temperature range, there was no apparent correlation between the peak temperature
and  the emissions of any of the target compounds.

 Completeness of VOC Recovery from Sample
 Since the sample is at elevated temperature for only 2 minutes - and is at the peak temperature
only momentarily in this procedure - it is of interest to determine the extent to which volatiles
have been driven from the sample. Accordingly, following one test with one of the samples, the
quartz sample tube was removed from the oven and sealed while the oven cooled.  It was then re-
installed and run through the ballistic heating protocol a second time.

Only five of the target analytes were emitted above the PQL during the second heating.  The mass
of each compound released during the second heating was 13 to 34% of the mass emitted  during
the first heating. This result indicates that some modest amount of the volatiles still remain after
the first heating, or that additional volatiles are created by thermal degradation upon re-heating.

-------
Effect of Air vs. N2 Carrier Gas
The ballistic heating tests were conducted with N2 as the carrier to avoid formation of oxygenated
species through oxidative degradation of the polymer.  Two toner samples were tested using
purified dry air as the carrier, to  assess the extent of additional degradation that might occur
during the tests if oxygen (0,) were present. The results showed that - with isolated exceptions -
there was no statistical difference in the emissions of any of the compounds caused by the use of
air as the carrier (p<0.05). Even in the isolated cases where there was a difference, the emissions
in air were only about 5 to 20%  higher. Thus - for the temperature profile seen by these samples
-the presence of O2 does not seem to create significant additional emissions from polymer
degradation in this test system.

Analysis of Thermal Desorption Results

Manufacturer Toners and Feedstocks
The results indicate that there is  no effect on the observed emissions resulting from:  the sampling
time during the manufacturing run; the toner cartridge, for cartridges filled at the same time
during the run; or the sampling location within a given cartridge. Thus, the discussion here
focuses solely on the effects of the manufacturing process and the feedstock.

The emissions from the thermal desorption tests on the manufacturer samples (expressed as
micrograms of compound emitted per gram of sample) are presented in Table 2.

As shown, 8 of the 11 target analytes consistently appear at concentrations above the PQL in all
of the samples; the other 3 targets are present, but below the PQL.  Of the 10 additional analytes
that  were added later, only n-propylbenzene could be identified in these samples, at low levels.
The eight major target compounds represent about 50% of the total VOC (TVOC)  mass emitted
from the  toners (estimated using the response factor for ethylbenzene), and about 60% of the
TVOC mass from the feedstocks. There were also four major unidentified compounds
consistently present. Using the ethylbenzene response factor, these compounds accounted  for an
estimated additional 30% of the toner TVOC mass, and 20% of the feedstock TVOC mass.

As shown, the % RSD for each compound in each sample set is usually less than 15%, which is
good.

The means in Table 2 were statistically compared using the two-sided t-test with pooled standard
deviations.  This  analysis yields the following results, with p < 0.05.

   «   The manufacturing process has no effect on toner emissions in this study.  For toners
       produced using either feedstock, the mean emission  of each compound is statistically the
       same regardless of whether the toners were produced by unvented or vented  extrusion.
       Since vented extrusion is specifically intended to reduce the volatiles content in the
       polymer product9'10'",  it  is suspected that the lack of an effect seen here for vented
       extrusion is the result of the negligible vacuum applied in producing the vented toners
       available for this study.
                                           10

-------
Table 2.  Summary of thermal desorption test results with the manufacturer toners.
Values in micrograms of VOC emitted per gram of solid sample (average ± standard deviation, % RSD in parentheses)
Feedstock A Feedstock B
Unprocessed Toner Produced by
Compound*
Ethylbenzene
p-Xylene
m-Xylene
o-Xylene
Styrene
3,3,5-Trimethyl-
cyclohexanone
Benzaldehyde
Acetophenone
Y, of all quantitated
VOCs
TVOC (as ethyl-
benzene) B
TVOC accounted for
Feedstock Unvented Extrusion
37 ± 5(13)
32 ± 4(12)
91 ± H (12)
52 ± 7(14)
155 ±19 (12)

18 ± 4 (22)
9± 1 (15)
9± 1 (13)

410±53(13)

672±113(17)
61%
33 ± 0.8 (2)
27 ±0.7 (3)
83 ± 2 (3)
53 ± 1 (3)
123 ± 4 (4)

26 ± 1 (4)
17 ±0.9 (5)
21 ± 2 (7)

388 ± 13 (3)

812 ±68 (8)
48%
Toner Produced by Unprocessed Toner Produced by
Vented Extrusion
32 ± 0.6 (2)
27 ± 0.6 (2)
83 ± 2 (2)
53 ± 1 (2)
123 ± 3 (2)

27 ±0.8 (3)
18 ±0.6 (3)
23 ± 0.4 (2)

390 ± 7 (2)

840 ± 49(6)
47%
Feedstock Unvented Extrusion
39 ± 4(10)
32 ± 3(10)
96± 10(10)
56 ± 6(10)
148 ± 11 (8)

22 ± 2(10)
10 ±0.9 (9)
10 ±0.9 (9)

418 ± 39(9)

705 ± 57(8)
59%
30 ± 2 (7)
25 ± 2 (8)
76 ± 6 (8)
48 ± 4 (8)
103 ± 12(12)

22± 3(13)
14± 2(13)
17± 3(17)

339 ±33 (10)

647 ±95 (15)
53%
Toner Produced by
Vented Extrusion
32 ± 2 (6)
26 ± 2 (7)
83 ± 5 (7)
53 ± 4 (7)
115± 10 (9)

25 ± 3(11)
16± 2(12)
20 ± 2(11)

375 ±30 (8)

741 ±99(13)
51%
Estimated mass in
  4 unknown peaksc  124 ±29 (24)    243 ± 35 (14)
TVOC accounted for
  (incl. 4 unknowns)     80%

No. of samples           5
                                       78%

                                        7
262 ± 20(8)


    77%

     4
                                                                       123 ±  9 (7)    171 ±38(22)
77%

 2
79%

 5
205 ± 44 (22)


    79%

     5
  A  Commonly detected, but below the practical quantitation limit (PQL):  diethylbenzene; a-methylstyrene; and 2,4-dimethylstyrene.
     Tentatively identified in all manufacturer samples, below the PQL:  n-propylbenzene.  The PQL is 5 to 6 [ig/g.
  B  jyQ£ js jegne{j as ^g tota| YOC mass eluting after 2.0 min on this column, computed using the response factor for ethylbenzene.
  c  Computed using the response factor for ethylbenzene. The peaks elute at 13.4, 28.3,  29.9, and 30.4 min on the DB-Wax column.

-------
   »   The feedstock has a small but statistically significant effect on the emissions from toners
       produced by unvented extrusion.  Emissions of each compound, and of TVOCs, average 8
       to 20% higher with the unvented toners made from Feedstock A, compared to unvented
       toners from Feedstock B. Feedstock has no statistically discernable effect on toners from
       vented extrusion.

   •   No difference can be discerned between the emissions from the two feedstocks used in this
       study. Thus, differences between the unprocessed feedstocks would not seem to be the
       sole explanation for the small differences between the unvented toners made from them,
       cited above.

   *   With a few exceptions, each unprocessed feedstock generally has 15 to 30% higher
       emissions of ethylbenzene, p-xylene, and styrene, compared to the toners manufactured
       using that feedstock. This observation would be consistent with the thesis that some
       fraction of these compounds - present as impurities in the feedstock - is driven off during
       the extrusion process.

   «   Each unprocessed feedstock consistently has  about 30 to 60% lower emissions of
       benzaldehyde and acetophenone, compared to the toners manufactured using that
       feedstock.  In addition, Feedstock A has about 30% lower emissions of 3,3,5-trimethyl-
       cyclohexanone, compared to the toners from  that feedstock. These observations would be
       consistent with the thesis that some amount of these oxygenated compounds is created by
       thermal-oxidative degradation of the polymer during the extrusion process,

   •   Feedstock A has about 50% less mass in the four unknown peaks than do the toners made
       from this feedstock, suggesting that these compounds also are created, in part, during
       extrusion. There is insufficient statistical power to discern whether these four unknowns
       are different between Feedstock B and its toners.

Retailer Toners
Table 3 presents the thermal desorption emissions from the four retailer toners, and compares
them against the results from the manufacturer toners. The manufacturer toner results are
presented according to the feedstock from which the toner was made, since feedstock seemed to
have a somewhat greater effect on emissions than did the manufacturing process.

As shown, three compounds consistently observed at significant concentrations from the
manufacturer toners - p-xylene, 3,3,5-trimethylcyclohexanone, and acetophenone - were below
the PQL in all of the retailer toners.  Thus, only 5 of the 11 target analytes were present above the
PQL in the retailer toners, accounting for only about  20 to 30% of the TVOC mass.

Of the 10 additional analytes that were added later, 5 could be tentatively identified in all of the
retailer chromatograms, at levels above the PQL: n-propylbenzene, 2-ethyl-l-hexanol, phenol, 2-
phenoxyethanol, and phenyl benzoate. These five compounds - plus four other significant,
unidentified peaks that were consistently present from all of the retailer toners - accounted for an
estimated 35 to 40% of the TVOC mass.

                                           12

-------
     Table 3. Thermal desorption test results comparing manufacturer toners vs. retailer toners.
u>
                    Values in micrograms of VOC emitted per gram of solid sample (average ± standard deviation, % RSD in parentheses)

                    	Manufacturer Toners	    	Retailer Toners
                    Toners P
Compound*

Ethylbenzene
p-Xylene
m-Xylene
o-Xylene
Styrene
3,3,5-Trimethyl-
  cyclohexanone
Benzaldehyde
Acetophenone

Y, of all quantitated
     VOCs

TVOC (as ethyl-
     benzene)

TVOC accounted for

Est. mass in other
  major peaks D

TVOC accounted for
  (incl. other peaks)

No. of samples
Toners Produced
from Feedstock AB
32 ±0.8 (2)
27 ± 0.7 (2)
83 ± 2 (2)
53 ± 1 (2)
123± 4 (3)
.26 ± 1 (4)
18 ±0.9 (5)
22 ± 1 (7)
Toners Produced
from Feedstock B B
31 ± 2 (7)
26 ± 2 (8)
79 ± 6 (8)
51 ± 4 (9)
109 ± 12(11)
23 ± 3(14)
I5± 2(13)
!9± 3(16)
Retailer l,Lot#l,
Cartridge # 1
I6± 2(10)
c
7 ±0.7(11)
13± 2(11)
48 ± 5(10)
r
47 ± 5(11)
c
Retailer l,Lot#2,
Cartridge #1
20 ± 2(12)
c
c
7 ± 0.6 (9)
28 ± 4(12)
c
68 ± 10 (9)
c
Retailer 2.
Cartridge #1
27 ± 1 (4)
c
6 ± 0.3 (5)
9 ± 0.4 (4)
64 ± 3 (4)
c
43 ± 2 (5)
c
Lot #3
Cartridge #2
25 ± 4(17)
c
6 ± 1 (20)
10 ± 1 (9)
60 ± 10(16)
c
40 ± 8(19)
c
                           389 ± II  (3)


                           822 ±61  (7)

                              47%    .


                           250 ±31  (12)


                              78%
357±35(10)


694 ±104 (15)

   52%


188 ±43 (23)


   79%

    10
143 ± 15(11)


592 ±75 (13)

    24%


234 ±34 (14)


    64%

     2
138±18(13)      163 ±10 (6)   155±26(17)


515 ±67 (13)      771 ±104(14) 683 ±153(23)

    27%              21%         23%


200 ±31 (16)      275±18(6)  248 ± 64 (26)
    66%

     2
57%

 2
59%

 2
       A  Commonly detected in both manufacturer and retailer toners, but below the PQL: diethylbenzene; a-methylstyrene; and 2,4-dimethylstyrene,
          Tentatively identified in all manufacturer samples, below the PQL:  n-propylbenzene.  Tentatively identified in all retailer samples, above the
          PQL: n-propylbenzene; 2-ethyl-l-hexanol; phenol; 2-phenoxyethanol; and phenyl benzoate. The PQL is 5 to 6 p.g/g.
       B  Includes toners produced using both non-vented and vented extrusion.
       c  Detected at levels below the PQL.
       D  For manufacturer toners,  includes the four unknown peaks cited in Table 2.  For retailer toners, includes:  five peaks which have been
          identified (n-propylbenzene, 2-ethyl-l-hexanol, phenol, 2-phenoxyethanol, and phenyl benzoate) but which were not included in the
          calibration standard; plus four unknown peaks (eluting at 7.7, 22.1, 24.9, and 31.0 min).  Mass computed using ethylbenzene response factor.

-------
 Statistical comparison of the means in Table 3 indicates the following, with p < 0.05.

   »   All of the retailer toners have emissions of ethylbenzene, xylenes, styrene, and
       acetophenone that are significantly lower than those from the manufacturer toners (with
       either feedstock). Retailer ethylbenzene emissions range from 15 to 50% lower; styrene
       emissions from 40 to 75% lower, xylene emissions from 70 to 100% lower; and
       acetophenone emissions (which are below the PQL in the retailer toners) roughly 75%
       lower.

   *   All of the retailer toners have emissions of benzaldehyde that are significantly higher - 125
       to 350% higher - than those from the manufacturer toners.

   •   TVOC emissions from the two Retailer 1 toners (Lot #1 and Lot #2) are lower than the
       TVOC mass from the manufacturer Feedstock A toners (by  about 25 to 35%).  For the
       other retailer toners and the manufacturer Feedstock B toners, the difference between the
       retailer and manufacturer TVOC emissions is generally not statistically significant.

   •   The emissions of o-xylene and styrene from one retailer toner (Retailer I/Lot #2) are
       statistically lower than the emissions of those compounds from the other three retailer
       toners.  But with that exception, there is no statistical  difference between the four retailer
       toners, within the statistical power of this analysis (two samples per toner). There is no
       difference between toners:  for the individual target analytes (except as noted for Lot #2),
       or for the estimated mass in other major  peaks; or for TVOC mass.  ~

The manufacturer, the process, and the feedstock characteristics represented by the four retailer
cartridges are unknown. It  is known only that the two cartridges from Lot #3 had the same
manufacturer, process, and  feedstock. Thus, it is uncertain whether the three retailer lots might
have been produced by the same manufacturer and process, and whether the similarities between
the toners might result for that reason.

As discussed previously in connection with the headspace results,  ethylbenzene, xylenes, and
styrene appear to result in large part from impurities in the raw polymer feedstock. The lower
emissions of these compounds from the retailer toners could be suggesting that:  a) the feedstocks
used to produce the retailer toners had lower concentrations of these compounds; and/or b) the
manufacturing process for these toners enabled more complete devolatilization of the melt (e.g.,
through effective venting, higher temperature, or longer residence time).

Since benzaldehyde and acetophenone originate in part as impurities in the feedstock polymer, the
higher emission of benzaldehyde (and the lower emission of acetophenone) from the retailer
toners could be suggesting differing concentrations of these compounds in the feedstocks used  to
produce those toners. But both of these compounds can also be created by thermal-oxidative
degradation of the polymer  during toner manufacture, as indicated by the results from the
manufacturer toners and feedstocks.  Thus, the higher emissions of benzaldehyde could also be
suggesting that the process  used to manufacture the retailer toners promotes a greater degree of
                                           14

-------
degradation producing this by-product (e.g., through higher temperature or longer residence
time).

The retailer toners also show significant emissions of another degradation product - phenol - at
concentrations on the order of (but less than) the benzaldehyde levels.  By comparison, the
manufacturer toners do not emit meaningful levels of phenol.  Thus, to the extent that process-
induced thermal-oxidative degradation is contributing to the increased level of oxygenated species
in the retailer toners, it would be doing so in a manner that tends to produce benzaldehyde and
phenol, and not acetophenone.  The degradation process with the manufacturer toners, by
comparison, clearly would be favoring 3,3,5-trimethylcyelohexanone and acetophenone to the
same degree as benzaldehyde, and would not be tending to produce phenol.

CONCLUSIONS

  1.  From comparison of the manufacturer and retailer toners tested here, it is clear that
      nominally identical toners - manufactured to meet the fuser specifications for a single
      photocopier - can have significantly different emissions of individual VOCs when heated
      in the laboratory. Emissions of a given compound can vary by a factor of 2 or more
      between toners.

  2.  Even when there are significant differences in emissions of individual VOCs between
      toners, it might not be possible to recommend  one as a clearly preferable low-emitting
      product. Comparison of the manufacturer and retailer toners indicates that - while the
      retailer toners had much lower emissions of some compounds (ethylbenzene, xylenes,
      styrene, acetophenone) - they had higher emissions of other compounds (benzaldehyde,
      phenol). (All of these compounds, except benzaidehyde, are HAPs.) And the difference
      in TVOC emissions between the two toner sets is modest at best, and often not
      statistically significant.

  3.  The differences in emissions between the manufacturer and retailer toners are almost
      certainly due to differences between the manufacturing processes and/or the feedstock
      polymers used in the two cases.  But without information on the retailer process(es) and
      feedstocks, the specific factors creating the differences could not be identified in this
      study.

  4,  Because the specific factors creating the emission differences between the manufacturer
      and retailer toners could not be identified, it is  not possible from this study to make
      specific recommendations regarding how process or feedstock might be  modified in order
      to produce lower-emitting toners for a given copier.

  5.  The tests on the manufacturer toners showed that vented extrusion did not produce toners
      having lower emissions than did unvented extrusion; but this result was probably obtained
      because only negligible vacuum (2 Pa) was applied during vented extrusion.  These tests
      also showed that the feedstock had only modest, if any, impact on toner emissions; but the
      feedstocks were almost identical, creating this  result.

                                           15

-------
   6.   The tests on the manufacturer toners and feedstocks demonstrated that essentially all of
       the compounds observed in the toner emissions result, at least in part, from impurities that
       are present in the feedstocks to begin with.  The tests also demonstrated that the
       concentrations of some species can be increased during the extrusion process, presumably
       by oxidative degradation of the polymer. These observations can be used to postulate
       explanations for the differences in emissions between the manufacturer and retailer toners.

ACKNOWLEDGMENTS
       The authors gratefully acknowledge the dedicated assistance of Angelita S. Ng and Bill T.
Preston of ARCADIS Geraghty & Miller, in conducting the extensive chemical analyses.
Appreciation is also expressed to the cooperating manufacturer who provided the systematic
series of toner and feedstock samples.

REFERENCES
   1.   Northeim, C.M.; Sheldon, L.S.; Whitaker, D.A.; Hetes, R.G.; Calcagni, J.A. Indoor Air
       Emissions from Office Equipment:  Test Method Deve/o/mient and Pollution Prevention
       Opportunities;  U.S. Environmental Protection Agency:  Research Triangle Park, NC.
       1998; EPA-600/R-98-080 (NTIS PB9S-165137)."
  2.   Brown, S.K. Indoor Air 1999,9,259-267.
  3.   Scharfe, M.E.; Pai, D.M.; Gruber, R.J.   In Encyclopedia of Materials Science and
       Engineering (\7ol. 2); Bever, MB., Ed.;  Massachusetts Institute of Technology; Pergamon
       Press: New York, 1986; pp 1503-1508.
  4.   Gruber, R.J.; Ahuja, S.K.; Seanor, D.A.  In Encyclopedia of Polymer Science and
       Engineering (Vol. 17), rev. ed.; John Wiley & Sons: New York, 1989; pp 918-943.
  5.   Gruber, R.J.; Julien, P.C. In Handbook of Imaging Materials; Diamond, A.S., Ed.;
       Marcel Dekker, Inc.: New York, 1991; pp  159-200.
  6.   Wolkoff, P.; Wilkins, C.K.; Clausen, P.A.; Larsen, K. Indoor Air 1993, 3, II3-123.
  7.   Leovic, K.W.; Whitaker, D.A.; Northeim, C.M.; Sheldon, L.S. J. Air & Waste Manage.
       Assoc. 1998, 48, 915-923.
  8.   U.S. Congress Public La\v 101-5-19 - Nov.  15,  1990: Amendments to the Clean Air Act:
       Title III -  Hazardous Air Pollutants. U.S. Government Printing Office:  Washington, D.C.
       1991; 104 Stat. 2531-2537.
  9.   Rauwendaal, C.  Polymer Extrusion (3rd ed.)\ Hanser Publishers: New York, 1994; pp
       387-402.
 10.   Stevens, M.J. Extruder Principles and Operation,  Elsevier Applied Science Publishers:
       New York, 1985; pp 109-115.
 11.   Biesenberger, J.A.; Todd, D.  In Devolatilization of Polymers;  Biesenberger, J.A., Ed.;
       Macmillan Publishing Co.: New York, 1983; pp 3-31.
 12.   James, D.H.; Gardner, J.B.; Mueller, E.G. In Concise Encyclopedia of Polymer Science
       and Engineering; Kroschwitz, J.I., Ed.; John Wiley & Sons: New York, 1990; pp 1110-
       1113.
 13.   Maecker,  N.L.; Armentrout, D.N. In Concise Encyclopedia of Polymer Science and
      Engineering; Kroschwitz, J.I., Ed.; John Wiley & Sons: New York, 1990; pp 1133-1134.
 14.   Forgo, G.; Ragnetti, M.; Stubbe, A.  J. Imaging Science and Tech. 1993, 37, 176-186.
                                          16

-------
 N RMRL- RTF- P- 518
      TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
I. REPORT NO.
     EPA/600/A-00/053
                           2.
                                                      3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
 Parameters Impacting the Emissions of Selected
  VOCs from the Toner for a Specific Photocopier
                                                      5. REPORT DATE
                            6. PERFORMING ORGANIZATION CODE
7.AUTH0R(s)D. B. Henschel (EPA); and R. C. Fortmann,
 N.F. Roache, and X.Liu (ARCADIS)
                                                      8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORQANIZATION NAME AND ADDRESS
 ARCADIS Geraghty  and Miller, Inc.
 P. O.  Box 13109
 Research Triangle Park, North Carolina 27709
                                                      10, PROGRAM ELEMENT NO.
                             11. CONTRACT/GRANT NO.
                              68-D4-0005
12. SPONSORING AGENCY NAME AND ADDRESS
 EPA, Office of Research and Development
 Air Pollution Prevention and Control Division
 Research Triangle Park, NC 277U
                             13. TYPE OF REPORT AND PERIOD COVERED
                             Published paper; April 2000
                             14. SPONSORING AGENCY CODE
                              EPA/600/13
is. SUPPLEMENTARY NOTES APPCD project officer is D. Bruce Henschel, Mail Drop 54, 919/
 541-4112. For presentation at Symposium on Engineering Solutions to Indoor Air
 Quality Problems. Raleigh,  NC,  7/17-19/00.	  	                    	
is. ABSTRACT
              paper gives results of the measurement of emissions — using a labora-
 tory thermal desorption apparatus- -from a number of nominally identical photocop-
 ier toners (manufactured for use in a specific model  copier) when the toners were
 heated to fuser temperature (180-200  C).  The objective was to determine if VOC
 emissions from a toner can be reduced by judiciously  selecting the process  and the
 raw polymer feedstock used in its  manufacture. Tests  were performed on toner and
 feedstock samples obtained directly from a toner manufacturer, systematically var-
 ying process and feedstock,  and on toner cartridges from different lots (for which
 process and feedstock were unknown)  purchased from local retailers.  Results show-
 ed that the retailer toners consistently had up to 350% higher emissions of some ma-
 jor  compounds, and up  to 100% lower emissions of others, relative to the manufac-
 turer toners  (p<0.05),  probably due to differences in process and/or feedstock. The
 manufacturer toners  showed essentially no effect of process or feedstock,  probably
 because the two processes and two feedstocks used by the manufacturer were not
 significantly  different from each other. It is  concluded that process and" feedstock
 can have a significant effect on emissions of  individual compounds, but it is not pos-
 sible from this study to recommend changes for lower- emitting toners.
17.
                             KEY WORDS AND DOCUMENT ANALYSIS
                DiSCRIPTORS
                                          b.lDENTIFIERS/OPEN ENDED TERMS
                                          c. COSATI Field/Group
 Pollution
 Pho toe opyi ng
 Emission
 Organic Compounds
 Volatility
                 Pollution Control
                 Stationary Sources
                 Photocopiers
                 Toners
                 Volatile Organic Com-
                   pounds (VOCs)
                 Tnrinnr Ai y	
13B
14E
14G
07C
20M
18. DISTRIBUTION STATEMENT

 Release to Public
                 IS. SECURITY CLASS (This Report)
                 Unclassified
21. NO. OF PAGES
    16
                 20. SECURITY CLASS (This page)
                 Unclassified
                                          22. PRICE
EPA Form 2220-1 (9-73)

-------