Site Visits Related to Diacetyl and Flavorings that Contain Diacetyl:

Food Manufacturing Facility A – 

A Coffee Roaster

Table of Contents

  TOC \o "1-5" \h \z \u    HYPERLINK \l "_Toc200188207"  1	Introduction	
 PAGEREF _Toc200188207 \h  1  

  HYPERLINK \l "_Toc200188208"  1.1.	Project Overview	  PAGEREF
_Toc200188208 \h  1  

  HYPERLINK \l "_Toc200188209"  1.2.	Food Manufacturing Facility A -
Overview	  PAGEREF _Toc200188209 \h  1  

  HYPERLINK \l "_Toc200188210"  1.3.	Quantifying Diacetyl Usage	 
PAGEREF _Toc200188210 \h  2  

  HYPERLINK \l "_Toc200188211"  1.4.	Plant Description	  PAGEREF
_Toc200188211 \h  3  

  HYPERLINK \l "_Toc200188212"  2	Process Descriptions	  PAGEREF
_Toc200188212 \h  4  

  HYPERLINK \l "_Toc200188213"  2.1.	Flavoring Worker	  PAGEREF
_Toc200188213 \h  5  

  HYPERLINK \l "_Toc200188214"  2.2.	Packaging Worker (flavored coffee)	
 PAGEREF _Toc200188214 \h  8  

  HYPERLINK \l "_Toc200188215"  2.3.	Sanitation Worker	  PAGEREF
_Toc200188215 \h  10  

  HYPERLINK \l "_Toc200188216"  2.4.	Maintenance Worker	  PAGEREF
_Toc200188216 \h  11  

  HYPERLINK \l "_Toc200188217"  3	Exposure Assessment	  PAGEREF
_Toc200188217 \h  12  

  HYPERLINK \l "_Toc200188218"  3.1.	Sampling and Analytical Methods	 
PAGEREF _Toc200188218 \h  12  

  HYPERLINK \l "_Toc200188219"  3.1.1.	Equipment	  PAGEREF _Toc200188219
\h  12  

  HYPERLINK \l "_Toc200188220"  3.1.2.	Side-By-Side Sampling	  PAGEREF
_Toc200188220 \h  14  

  HYPERLINK \l "_Toc200188221"  3.1.3.	Other Analytes	  PAGEREF
_Toc200188221 \h  15  

  HYPERLINK \l "_Toc200188222"  3.1.4.	Limit of Quantification	  PAGEREF
_Toc200188222 \h  15  

  HYPERLINK \l "_Toc200188223"  3.1.5.	Blank Samples	  PAGEREF
_Toc200188223 \h  16  

  HYPERLINK \l "_Toc200188224"  3.1.6.	Sources of Uncertainty	  PAGEREF
_Toc200188224 \h  16  

  HYPERLINK \l "_Toc200188225"  3.2.	Air Monitoring Results	  PAGEREF
_Toc200188225 \h  17  

  HYPERLINK \l "_Toc200188226"  3.2.1.	Diacetyl and Acetoin	  PAGEREF
_Toc200188226 \h  17  

  HYPERLINK \l "_Toc200188227"  3.2.2.	Other Analytes	  PAGEREF
_Toc200188227 \h  26  

  HYPERLINK \l "_Toc200188228"  4	Examples of Controls in Place	 
PAGEREF _Toc200188228 \h  26  

  HYPERLINK \l "_Toc200188229"  4.1.	Overview	  PAGEREF _Toc200188229 \h
 26  

  HYPERLINK \l "_Toc200188230"  4.2.	Controls Associated with Individual
Jobs	  PAGEREF _Toc200188230 \h  26  

  HYPERLINK \l "_Toc200188231"  4.2.1.	Flavoring Worker	  PAGEREF
_Toc200188231 \h  26  

  HYPERLINK \l "_Toc200188232"  4.2.2.	Packaging Worker	  PAGEREF
_Toc200188232 \h  27  

  HYPERLINK \l "_Toc200188233"  5	Discussion	  PAGEREF _Toc200188233 \h 
27  

  HYPERLINK \l "_Toc200188234"  5.1.	Relationship between Results,
Sources of Exposure, and Controls in Use	  PAGEREF _Toc200188234 \h  27 


  HYPERLINK \l "_Toc200188235"  5.2.	Other Opportunities for Exposure
Control	  PAGEREF _Toc200188235 \h  27  

 Butter Flavor and Food Flavors Containing Diacetyl

Site visit report: 

Food Manufacturing Facility A – A Coffee Roaster 

Introduction

Project Overview

Eastern Research Group, Inc. (ERG) conducted an industrial hygiene site
visit in January 2008 at Food Manufacturing Facility A, a coffee
roasting plant (hereafter referred to as Facility A), located in the
United States. The purpose of the site visit was to obtain information
regarding the use of flavor products containing diacetyl and commonly
associated flavor ingredients (e.g., acetoin) in the manufacture of food
and beverage products; gather information on the exposures of employees
at the facility to diacetyl and acetoin; and document the controls in
place to reduce exposure. In particular, the site visit was intended to
evaluate the exposures of workers who create, handle, or worked near
bulk flavored coffee products while they perform coffee grinding,
flavoring, packaging, housekeeping/sanitation, and routine maintenance
operations.

On January 16, 2008, two ERG industrial hygienists met with
representatives of Facility A to discuss the process of making flavored
coffees, the materials used in this process, and the job categories with
potential for exposure to food flavorings that contain diacetyl. A
walk-through of the facility was conducted as part of the meeting. On
January 17, 2008, ERG conducted short-term, partial-shift and full-shift
personal breathing-zone (PBZ) air monitoring and observed the work
practices of employees identified as having potential for exposure to
diacetyl. ERG also noted airflow patterns and exposure controls in use
at the facility. 

Food Manufacturing Facility A - Overview

Facility A roasts coffee beans to produce custom coffee blends. These
products leave the plant as packaged beans or ground coffee, which may
be unflavored or flavored. An overview of the general coffee roasting
process follows. 

The production steps for coffee involve:

Blending various types of unroasted, called “green,” coffee beans
(e.g., from various geographic locations) to achieve the specific
desired coffee taste characteristics. 

Roasting the blended beans to the desired roast color profile (e.g.,
“dark roast”).

Grinding the beans to a size compatible with the ultimate coffee brewing
machines in which the coffee will be used.

Packaging the whole beans or ground coffee. 

At each stage, quality is tested in an on-site laboratory ensure the
blending, roasting and grinding processes are meeting production quality
standards. No decaffeination is performed at this site (or anywhere in
the U.S.); instead, the facility purchases bulk quantities of
decaffeinated beans, which are processed using the same methods as other
beans.

To make a flavored coffee (e.g., “hazelnut”), a flavoring agent is
added to ground or whole beans shortly before the product is packaged.
Thus, the flavoring step is always the final action that employees
perform on the product before packaging the coffee. Earlier steps in the
production process, such as blending, roasting and grinding, do not
involve direct handling of food flavorings or flavored coffee.

Employment at Facility A falls between 50 and 200 employees, working on
two 8-hour shifts per day. The workforce is relatively stable, mature,
and experienced with high English proficiency among the employees with
whom ERG spoke. Coffee roasting plants are traditionally associated with
occupational hazards such as heat, grinding blades, machinery with
moving parts, transfer of heavy objects, and noise. Posted signs
indicated that the plant had reached significant milestones without
lost-time injuries. Basic occupational safety and health support for
this facility is provided by a collateral duty safety/environmental
manager, with more advanced services coordinated as needed through the
company’s main office.

Quantifying Diacetyl Usage

Owned by a coffee manufacturer, Facility A produces over 15 million
pounds of coffee per year. Although some flavored coffee is made every
day, less than 10 percent of the coffee produced here is flavored, while
the remaining 90 percent is shipped unflavored. This is reportedly a
fairly typical ratio for the coffee industry, which has seen a slight
decline in interest in flavored coffees since the peak several years
ago. Between 2 and 10 pounds of flavoring agent are added to a batch of
flavored coffee, depending on the desired flavor and the size of the
batch, which in turn are based on customer requirements. Of the many
flavoring agents used at this facility over the course of a year, fewer
than two dozen contain diacetyl and roughly a quarter of those were used
less than four times in the 6 months prior to ERG’s visit. Some
diacetyl-containing flavors, however, are used nearly every day and
sometimes several times per day.

The facility management estimates that over a 6-month period this plant
used, on average, the equivalent of less than 1 pound (and probably
closer to a half pound) of diacetyl per day, always as an ingredient in
the numerous pre-blended flavors that the company purchases from food
flavor manufacturers (pure diacetyl is never used at this facility). The
diacetyl-content of flavorings used relatively frequently at this site
range from approximately 0.05 percent to slightly over 3 percent
diacetyl by weight. Three of the more frequently used flavorings contain
greater than 1.0 percent diacetyl, while most of the remainder include
diacetyl in the range of 0.1 to 1.0 percent. 

Information on diacetyl content was provided by the flavor manufacturers
after persistent requests from the coffee manufacturer in a process that
was described as “time consuming.” The coffee manufacturer does not
receive information on the content of acetoin or flavor components other
than diacetyl from the flavor manufacturers and contractual agreements
reportedly preclude discussion of these other ingredients.

Plant Description

The plant was built several decades ago in an industrial neighborhood as
a factory producing a different product. Typical of most coffee roasting
facilities, this one is divided into the following major areas: a
storage and blending area for green coffee beans, a roasting zone, an
area containing grinding equipment, and a packaging area. Separated from
the other spaces, discrete enclosed areas are designated as a flavor
room for flavoring coffee (plastic strips hanging in both doorways), an
equipment wash room (plastic strips hanging in doorway), a QA laboratory
(double set of solid latching doors), an employee cafeteria (latching
doors with inset louvers), a maintenance shop (open doorway), a
packaging supplies warehouse (latching doors), and offices (double set
of latching doors). 

The dedicated flavor room is used to blend flavors into coffee beans or
ground coffee. The room contains two machines of types typically used
for blending flavor into beans or ground coffee, one of which is on a
mezzanine. The room is approximately 20 feet by 30 feet in area with a
high ceiling. An additional level of free space, above the mezzanine
only, contains a fan that exhausts flavor room air directly outdoors
through that section of tall ceiling. The two flavor room cargo doors
are covered with hanging plastic strips, which ERG noted were gently
swaying outward from the flavor room, suggesting that, on this day, the
main direction of air flow was from the flavor room out into the
grinding/roasting area through the door on one side and into the flavor
storage area through the door on the other side of the room. Several
bulk containers of flavors are stored in the flavor room, while others
(both larger and smaller) are kept in the flavor storage area, which
covers several hundred square feet outside the flavor room. 

The packaging zone consists of numerous form/fill/seal packaging
machines covering a substantial area of the plant. Three packaging
machines dedicated to flavored coffee (beans or ground) are at one end
of the packaging zone.

In addition to the exhaust fan in the flavor room ceiling, there is a
canopy hood and associated work bench outside the flavor room. This work
station is reportedly used only rarely for special projects and is not
currently considered part of the manufacturing equipment.

General factory air is not tempered (cooled or heated, except by the hot
roasting equipment); however, air conditioning (A/C) is available for
the flavor room and is activated when indoor temperatures exceed 80
degrees F. Industrial exhaust fans (estimated diameter of 30 inches)
appear at regular intervals in exterior walls above head height. Even
when only a portion of these are in use, they create notable air
exchange and a sense of air movement in the plant (in the direction of
the warehouse). The exhaust fans tend to be used extensively as they are
the only method for cooling workers in warm weather.  

On the day of the site visit, measurements collected at eight different
times and locations showed that temperatures in the plant production
areas varied little and were cool at 56 to 60 degrees F, with relative
humidity ranging from 57 to 65 percent. During summer months, both
temperature and humidity could rise by 30 points in all production
areas, except the air conditioned flavor room.

Process Descriptions

Green bean handling, roasting, and most QA activities occur before
employees add flavoring to the coffee. Grinding is also performed on
unflavored beans. These activities largely occur 40 to 100 feet away
from the flavor room; however, one grinding machine is located adjacent
to the flavor room door. The same employees who perform grinding are
also responsible for adding/mixing flavor into the coffee as aseparate
process. Therefore, this process description begins with the flavoring
workers and continues with packaging workers, sanitation workers and
maintenance workers (mechanics). Both work shifts include employees in
each of these job categories. Unless otherwise noted, the numbers of
employees presented here are approximate plant totals, for both shifts
combined.

Quality control is also performed on flavored coffee, but this primarily
involves the QA technician evaluating the scent and flavor of a teacup
of the brewed coffee, as would a typical customer. Although many cups
may be evaluated over the course of the day, most are smelled and tasted
for only an instant. The flavored coffee evaluation session observed by
ERG involved less cumulative time and volume of coffee intake than a
consumer would experience in drinking one cup. QA technicians spend the
vast majority of their time testing unflavored or yet-to-be flavored
coffee and do not enter the flavor room. Flavored coffee only enters the
QA lab in small quantities for the taste testing sessions just
described. Therefore, QA technicians are not further addressed in this
report. 

Flavoring Worker

Employees in this job category rotate between two activities every few
weeks: 1) primarily grinding (with little or no flavor mixing) and 2)
primarily mixing flavors with coffee (with grinding duties between
flavor blending sessions). 

When primarily grinding, the employee operates industrial coffee
grinding equipment, obtains samples of ground coffee and conducts
quality checks in the laboratory several times per hour. The employee
might work at the grinding station near the door to flavor room or on
other grinding machines 40 or more feet away from the door. Employees
operating grinding machines occasionally pass through the flavor room
(10 seconds) as a shortcut to other parts of the plant.

All flavorings are mixed with coffee in the dedicated flavor room. The
flavor mixing task starts with the flavoring worker obtaining the
appropriate containers of pre-blended flavors from the open rack storage
area (open warehouse space adjacent to the flavor area). This employee
manually weighs flavor into an open transfer container on a scale at the
base of the mezzanine in the flavor room (thus nearly directly below the
exhaust fan, some 30 feet above), then adds the weighed material to the
flavoring equipment. During the weighing process, the 2-foot tall
open-top cylindrical transfer container is placed on a floor scale with
wall mounted display on a post 4 feet above the scale. The employee adds
flavor to the cylinder from hand-held packaging, leaning over the
transfer container and watching the scale display, so the employee’s
breathing zone is approximately 18 inches above the point where flavor
leaves the packaging and enters the cylinder. The tall cylinder
minimizes the distance that manually dispensed flavors freefall before
entering this transfer container. The cylinder depth also means that
until the container is partially filled, flavor must drop an additional
2 feet to the bottom of the cylinder.

The company treats some aspects of the coffee flavoring process as
proprietary; however, the text box below offers an example of how
flavored coffee is made. From an industrial hygiene perspective, the
process of handling diacetyl-containing flavorings at Facility A differs
from that described in the text box in that flavoring is the final step
before packaging of any product and, for some flavors, an automated
flavor metering system is associated with one of the flavoring machines.
Manual weighing is not required with the metering system, which draws
from large packages of the most commonly used flavors and meters
measured amounts of the flavor directly into the flavoring equipment. 

Other activities of the flavoring workers include material handling to
transfer totes or super sacks of room-temperature beans or ground coffee
into flavor mixing machines using a forklift or powered material
handling equipment. These workers also control the mixing equipment. 

The flavoring worker also moves bulk containers of finished flavored
coffee to a short-term holding area or to the packaging line. These bulk
containers consist of a large (estimated typical half-ton size) woven
super sack lined with a large plastic bag and suspended on a rolling
support. Once the sack is filled, the mouth of the plastic liner is
loosely folded over the top. The plastic liner effectively limits
employee exposure to flavored beans once the liner is folded over to
cover the top. While the sack fills and until the liner top is folded,
an area of approximately 20 square feet of flavored beans is open to the
air. 

The same flavoring worker cleans the flavor transfer container in the
washroom and rinses one of the two flavor machines in place in the
flavor room. Both the container and machine are rinsed with hot water
from hoses fitted with spray nozzles and ERG observed steam rising (in
the cool facility) from both rinsing operations. Rinse water from the
weigh container is poured across the washroom floor and eventually runs
down a drain. Rinsing the flavor machine (on the flavor room mezzanine)
involves the employee leaning into the equipment so chest and head are
near the plane of the mixer cavity opening (so the employee can see
inside – the operators do not enter the mixer). The employee reaches
in and directs the spray nozzle around the mixer interior. To capture
rinse water from the mixing machine, the operator drains the water
through a port

into a large tub and transports it to the wash room, where the tub is
emptied into a floor drain, wetting a 10-foot by 15-foot area of the
floor. ERG observed that together the container and machine rinsing take
less than 10 minutes.

The other coffee flavoring machine is cleaned by running some unflavored
coffee through it. The resulting coffee waste is accumulated in a 30
gallon covered container (plastic snap-on lid) in the flavor room until
the container is full, then it is discarded (removed from the facility).
The flavoring worker also hoses and mops the flavor room floor for 2 to
5 minutes after each multiple-batch flavoring session, which reportedly
may last from 20 minutes to 8 hours depending on the number and types of
batches mixed. 

Employees are instructed to minimize time in the flavor area. Only the
individual performing the coffee flavoring is usually present in the
flavoring room while coffee flavoring is in progress and even this
employee may step out of the room for part of the mixing time. ERG
observed that for up to 50 percent of the flavoring session the employee
may be elsewhere, performing tasks such as operating a forklift,
selecting flavors from the storage area, moving coffee containers, or
coordinating schedules. Thus, during a 45 minute flavoring session, the
employee may spend less than 20 minutes in the flavor room. Based on an
initial hazard evaluation, the company does not require employees to
wear respiratory protection in the flavor room, but does require safety
glasses and hearing protection (as a precaution).

During periods when only unflavored coffee is produced, the flavoring
workers perform routine coffee grinding (described previously). When
assigned to flavoring duty, these employees might spend a half hour or
less per day flavoring coffee, or they might work in the flavoring room
for their entire shift. The flavoring work has strong seasonal
variation, with flavored coffee production higher in the late summer and
early fall. Production on the date of the air monitoring was typical of
the winter and spring seasons, but was light compared to the months
proceeding the winter holiday season.

ERG monitored the activities of three flavoring workers, two of whom
each performed one multi-batch flavoring session (each making three
consecutive batches of flavored coffee): 

In the morning, Flavoring Worker #1 conducted one approximately
30-minute flavoring session, producing three batches of flavored ground
coffee and/or beans by running both coffee flavoring machines
simultaneously. Two of these flavors contained diacetyl, as
approximately 0.5 percent and 2.7 percent by weight. According to
information received by Facility A from the flavor manufacturer, the
third flavor contained no diacetyl. Flavoring worker #1 ground beans and
performed QA evaluations of ground beans before and after the flavoring
session.

In the afternoon Flavoring Worker #3 spent 45 minutes producing three
consecutive batches of flavored beans, two of which contained diacetyl
as an ingredient of the flavor (as approximately 0.05 percent and 1.4
percent of the two flavors). A third flavor used in the afternoon
reportedly contained no diacetyl. This employee also performed QA
evaluations and ground unflavored beans for the remainder of the evening
work shift.

The third employee, identified as Flavoring worker #2, spent the entire
shift grinding unflavored beans and performing QA steps to produce
unflavored ground coffee. This employee worked a split or bridge shift
(overlapping the morning and afternoon shifts) and was operating the
grinder outside the flavor room during the afternoon flavoring session,
but was not present in the facility during the morning session.

Packaging Worker (flavored coffee)

Approximately a half dozen of the several dozen packaging workers at
this plant are responsible for monitoring the three packaging lines used
for flavored coffees. The operator receives flavored beans or ground
coffee in bulk containers from the flavor room or short-term holding
area soon after the coffee is flavored. The coffee is automatically
transferred from the bulk container into hoppers that gravity-feed the
packaging equipment below. Standing at the control position, the
operator maintains product flow rate through the packaging machine and
monitors stocks of packaging materials. This job, however, frequently
requires the operator to move around the machine to adjust process
materials and finished packages. In general, the normal operation of
this “form/fill/seal” packaging equipment minimizes operator
exposure to the product being packaged by keeping the product enclosed
(in a hopper, in a delivery tube, in a packaging material sleeve) at all
times. In the equipment observed by ERG, there was no product “free
fall” through open air and there was little air displaced by this
filling process. 

Form/fill/seal packaging machines typically function in the following
manner:

The product falls from a hopper or source container through a tube
inserted into a continuous sleeve of packaging material. 

The sleeve is constantly formed in the packaging machine by folding the
sides of a continuous plastic sheet upon itself. The plastic (Mylar,
film, or other material) sheet is dispensed from a roll.

As the sleeve forms, a measured amount of product is dropped into it.

The machine automatically lowers and seals the sleeve to create a pouch
of measured product. Meanwhile a new segment of sleeve is being formed
and filled above.

Once separated from each other by the machine, the pouches form
individual bags or packaging units of product. 

Occasional equipment malfunctions can cause product spills, which in
this case, the packaging worker cleans with a broom and dust pan.

On this date the operator being monitored oversaw equipment that filled
typical 1-pound bags of ground coffee that included a flavor containing
approximately 0.5 percent added diacetyl by weight; thus the flavored
coffee consisted of less than 0.01 percent diacetyl.  Process equipment
did not operate smoothly and the maintenance worker (also monitored)
spent approximately a half hour in the morning making adjustments to the
machine and the hopper/feed equipment above, from which coffee grounds
flow into the machine. Despite time out for adjustments, the amount of
coffee ultimately packed was reportedly similar to that produced on a
normal day. The packaging worker was monitored in the morning only (249
minutes) to free the monitoring equipment for sampling in other
locations. In the afternoon this operator packed coffee that had been
flavored that morning (also 0.5 percent diacetyl in the flavoring agent)
and this employee’s exposure to flavoring ingredients during the
unsampled period was likely to have been generally similar to the
sampled period.

ERG noted that when functioning well, the packaging process released
little if any residualproduct and even the adjustment/repair work
performed by the operator and maintenance personnel generated minimal
dust. A cumulative total of a few ounces of spilled product was observed
in the flavored coffee packaging area during the entire day of the site
visit.

Sanitation Worker

The facility employs one full time sanitation worker who is responsible
for cleaning all surfaces except in the immediate area of process
equipment, which is typically cleaned by the operator. Another worker
helps with this job as needed on the second shift. Coffee manipulation,
particularly green bean handling and roasting, can generate dust. The
grinding process generates coffee grounds, but reportedly creates
minimal fine particles. 

Because all green bean handling, roasting, and grinding are performed on
unflavored coffee, and because flavored coffee is a small portion of
this facility’s business, ERG estimates that general dust cleaned by
the sanitation worker in most areas of the plant probably contains
little or no flavoring agents. Any dust and spilled coffee that contain
flavoring agent are most likely to be concentrated in zones near the
flavor mixing and flavored coffee packaging areas. Although the areas
immediately under and around equipment are cleaned by the equipment
operators, the sanitation worker cleans adjacent thoroughfares and
common areas. 

The sanitation worker constantly cleans floors with broom, dry mop, and
dustpan, at times using a sweeping compound (“Oil Dry – Quick Sorb
Sweeping Compound” was used outside the office suite). The worker also
dusts surfaces with a brush or broom. Although plant processes
constantly generate dust, perpetual cleaning keeps the plant generally
clean and free of gross accumulations. When spills occurred, ERG
observed that the dry debris were usually swept up within the hour, if
not sooner. The general impression on the day of the site visit was of a
fairly clean manufacturing facility.

The sanitation worker was monitored only for the first half of the day
to free monitoring equipment for additional samples elsewhere in the
facility. During the monitored period, the sanitation worker swept and
dry mopped in and around the core area of the facility, including in the
flavor storage area, employee cafeteria, office areas, restrooms, and
green bean storage area. This worker spent the remainder of the shift
cleaning floors around the inside perimeter of the facility, where
potential sources of exposure might have been more closely related to
packaging lines rather than in the morning, when the worker spent time
in the flavor storage area. Reportedly, the sanitation worker spends
only about 1 percent of work time in the flavor room, which is cleaned
after each flavoring session by the individual who mixes flavor into the
coffee.

Maintenance Worker

Five fulltime maintenance workers adjust and repair process equipment
throughout the plant. The maintenance worker interviewed stated that
maintenance work is rarely required in the flavor room and most
maintenance efforts are focused on packaging equipment, which has more
moving parts, is more complex, and is present in greater numbers than
other equipment at this plant. When not adjusting fixed-position
equipment, maintenance workers work in a maintenance shop (adjacent to
the green bean storage area), which is equipped with typical shop
equipment and exhaust ventilation fans set in the wall.

The maintenance worker who was monitored spent part of the morning in
meetings, then made adjustments to the same flavored coffee packaging
line on which the monitored packaging worker worked. Later in the day,
the maintenance worker repaired another of the plant’s flavored coffee
packaging lines, approximately 50 feet away from the first one. The 344
minutes monitored are generally representative of the unsampled portion
of the shift.

Table 1 summarizes the job categories and employee activities with
potential for exposure to flavors that contain diacetyl. 

Table 1.

Job Categories and Activities with Potential for Exposure to Flavors
That Contain Diacetyl  

Food Manufacturing Facility A – A Coffee Roaster

January, 2008

Job Title	Activities/Potential Sources

Flavoring Worker	Rotates every couple weeks between primarily grinding
coffee and primarily mixing flavors with coffee. 

Grinding: Operates coffee grinding equipment, obtains and processes QA
samples, may work near or walk through flavor room.

Mixing flavors with coffee: In dedicated flavor room, weighs
commercially prepared flavor and adds flavor to coffee mixing equipment,
or activates bulk flavor metering system. Operates mixing equipment.
Rinses container and equipment with hot water. Performs grinding for
remainder of shift.

Packaging Worker (flavored coffee)	Operates automated packaging
equipment, adjusts product flow rate, and monitors stocks of packaging
materials. Cleans up minor spills of flavored coffee or beans.  

Sanitation Worker	Constantly sweeps floors and dusts surfaces throughout
plant with broom, dustpan, brush, and at times sweeping compound. Does
not clean the flavor room or process equipment.

Maintenance Worker	Adjusts and repairs process equipment throughout
plant, including flavored coffee packaging machines, or works on parts
in maintenance shop.



Exposure Assessment

Sampling and Analytical Methods

Equipment

ERG used SKC Universal Flow Sampling Pumps calibrated at a flow rate of
1.5 liters/minute, combined with SKC adjustable-flow quad tube holders
drawing a maximum total of 500 ml/minute. The pumps were set up per
manufacturer’s instructions for use with low flow sampling media. Each
port of the adjustable-flow tube holders was set in advance to the
designated air flow level (see Table 2) and checked for calibration
levels immediately before sampling and again at the end of the period
using a Dry-Cal DC-1 primary air flow meter (Flow Cell Serial Number:
S2348; Base Serial Number: B2440) manufactured by BIOS International,
Inc.; or DC-Lite Electronic Flowmeter (5 ml/min to 5000 ml/min), model
412360, manufactured by SKC, Inc., Eighty-Four, Pennsylvania.



Table 2.  Sampling and Analytical Methods

Low Flow Tube Holder Port

Analyte	Sampling & Analytical Method	Media	Air Volume	Air Flow Rate
Sample Duration	Notes/Special Handling 

Port A

Diacetyl/Acetoin

(Experimental)	Based on

OSHA PV2118, but  with both increased air flow rate and extended
sampling time	Silica gel 

200/400 mg 

SKC 226-10-3

pair in series	Not defined	150 to 200 ml/min

 	2 hours

(doubled for extended time)	Two “jumbo” tubes in series

Port B

Diacetyl/Acetoin

(Extended)	OSHA PV2118 with extended sampling time	Silica gel 

200/400 mg

SKC 226-10-3

pair in series	6 L 	50 ml/min

 	2 hours

(doubled for extended time)	Two “jumbo” tubes in series; larger
sample volume acceptable at lower temperature and humidity. Increased
air volume collected with longer sampling times.

Port C

Benzaldehyde

(as vapor)

DISCONTINUED*	OSHA 68	XAD-2 tube

75/150 mg

SKC 226-117 	12L	50 ml/min	4 hrs	Acetaldehyde may also be analyzed upon
request. 

Port D (partial shift)

Total Volatile Organic Compounds -TVOC

DISCONTINUED*	OSHA 07	Anasorb - CSC 50/100 mg 

SKC 226-01

 	10 L	50 ml/min

 	4 hours	Cumulative total of analytical peaks



Port D (short-term)

Total Volatile Organic Compounds

DISCONTINUED*	OSHA 07	Anasorb - CSC, 50/100 mg 

SKC 226-01

 	10 L	150 to 200 ml/min

 	30 minutes 	Cumulative total of analytical peaks



Notes:

L means liters.

ml/min means milliliters per minute.

SKC# indicates the brand and model number of sampling media.

* Monitoring for benzaldehyde, acetaldehyde, and TVOCs was discontinued
after the first two site visits (to Food Manufacturing Facilities A and
B)



Complying with food manufacturing industry policies against exposed
glass in food handling areas, ERG maintained all sampling media enclosed
in appropriate size plastic tube covers (SKC models 222-3-1 L, XL or XD)
attached to the tube holders. To change sorbent tubes the ERG team
removed the entire sampling train (or in some cases just the tube holder
and associated covered media) from the employee and carried the sampling
equipment to a designated area before removing tube covers, replacing
sorbent tubes and covering the new tubes. The sampling train could then
be returned to the employee in the food handling area. To improve air
flow calibration reliability for later samples, the flow rate through
pumps, holder, and tube sets were individually checked before these were
placed on an employee. All analyses were performed by the OSHA Salt Lake
Technical Center (SLTC) Laboratory using the methods presented in Table
2.

ERG obtained temperature and humidity readings and, where feasible, air
velocity measurements using a VelociCalc model 9545-A thermoanemometer
(air velocity meter), manufactured by TSI, Inc., Shoreveiw, Minnesota.
Both the flow meters used to calibrate air sampling pumps and the air
velocity meter were factory calibrated within the previous 12 months. 

Drop-shipped rental sampling equipment arrived incomplete and so
partial-shift samples were obtained for employees considered most likely
to have similar exposures in the morning and afternoon. This permitted
ERG to collect separate samples for additional employee activities in
the afternoon (e.g., during afternoon flavor mixing). For the same
reason, two samples originally anticipated to be STELs (samples 8 and
12) and one area sample (sample 7) were obtained using only the higher
air flow rate instead of both flow rates.

Side-By-Side Sampling

Use of the SKC quad adjustable-flow tube holders permitted up to four
low-flow samples to be obtained simultaneously using one air sampling
pump. Air sampling ports on the quad tube holders were labeled A through
D. For evaluating diacetyl (CAS 431-03-8) in Port B, ERG used jumbo
silica tubes in series following a modified version of the OSHA
provisional sampling method (OSHA PV2118), which allows for an increased
sampling time compared to the unmodified method. In order to maintain a
low-as-possible limit of detection, in several cases ERG extended the
sampling period beyond the recommended 2 hours, but maintained the
recommended air flow rate of 50 ml/minute. Anecdotal information
suggested that a previous investigator had success with this extended
time method, providing the humidity and temperatures were relatively low
and the sample duration was maintained well below 8 hours. 

In an attempt to further reduce the diacetyl sample limit of detection,
ERG also used an informal experimental procedure to test a higher air
flow rate in some samples. Using sampling media in Port A identical to
that in Port B, ERG adjusted the air flow rate to between 150 and 200
ml/minute. Therefore, for most diacetyl samples, a second side-by-side
sample was obtained at the higher-than-recommended flow rate. These
results, which are designated with sample numbers ending in “A” (for
Port A) should be considered experimental. 

Sorbent tubes were used as a pair in series in an attempt to detect
analyte breakthrough if analyte was captured on the second “backup”
tube. None of these backup tubes contained detectable diacetyl or
acetoin. ERG did, however, notice a trend (not uniformly consistent)
toward higher values reported for the non-experimental sample results
(Port B). 

Under the PV2118 method the SLTC laboratory is able to simultaneously
analyze acetoin (acetyl methyl carbanol, CAS 513-86-0) from the same
media as is used for diacetyl, so the side-by-side sampling method
produced a result for each of these analytes.

Other Analytes

Ports C and D were designated for testing the value of monitoring
airborne benzaldehyde and total volatile organic compounds (TVOC) in the
facility using standard methods available for typical industrial
monitoring through the OSHA SLTC laboratory, as listed in Table 2. OSHA
and ERG originally wondered whether these and related analytes (such as
organic acids) might be used as alternatives to diacetyl as markers to
evaluate beneficial exposure controls (e.g., engineering, work practice
controls). However, the analytical methods as tested did not add
practical value. Furthermore, food manufacturing facilities were unable
to discuss the presence of these other ingredients in food flavors
because this is considered protected information. A brief summary of the
benzaldehyde results is provided in the section of this report on Air
Monitoring Results, Other Analytes.

Limit of Quantification

The SLT laboratory usually reports sample results that are less than the
limit of quantification (LOQ), as “less than the LOQ” (or <x ppm,
where x is the LOQ). The LOQ is a statistically derived value below
which results are often judged unacceptably imprecise due to limitations
of the analytical process. ERG, however, is interested in quantifying
exposures in the same range as the LOQs for OSHA’s available
analytical method and anticipated that a substantial number of results
obtained at this site could be in that range. Therefore, instead of
reporting results as simply “<LOQ”, the laboratory reported the
numeric value that the analytical equipment provided, even if it was
below the LOQ. ERG provides these results in Table 3, believing they
offer more information about the approximate airborne concentration than
would a blanket description of “<LOQ.” ERG acknowledges, however,
that the lower the result, the less precise that value is likely to be.
Individual values surrounding the LOQ might not be reproducible,
although together, they may suggest trends that should be confirmed
later using more precise techniques. With the results in Table 3, ERG
indicates whether the result is above or below the LOQ and reports the
LOQ as a ppm value for each sample, calculated from the reporting limits
of 0.468 μg per sample for diacetyl and 1.008 μg per sample for
acetoin.

Results for samples in which no analyte was detected are reported as ND
(“none detected”). In calculating averages, ERG takes ND to be equal
to 0 parts per million (ppm), but acknowledges that this practice also
introduces a small but undefined degree of imprecision. The laboratory
did not provide a minimum limit of detection (LOD), since the actual
measured values were reported down to the level at which no analyte was
detected.

Blank Samples

ERG submitted two blank samples for every 10 air samples. Manipulated
with clean hands in a designated conference room within the office
suite, the blanks were handled in the same manner as the air samples,
except that air was not drawn through them. The laboratory reported
measurable background levels in the blanks for this site and the average
background values of 0.769 μg of diacetyl and 0.972 μg of acetoin (1.6
and 1 times the LOQ respectively) were subtracted from sample results.
No measurable background levels of diacetyl and acetoin were reported
for blanks of the same lot number used during a similar site visit to
another facility. 

Sources of Uncertainty

There was poor agreement between the results obtained through
side-by-side sampling at the two different flow rates (nominally 50
ml/minute and 150 to 200 ml/minute), probably due to the following
reasons:

Most of the reported results are in the low parts per billion level –
where, with the currently available sampling and analytical methods for
diacetyl, the effects of normal random variations can appear to make
small differences between results appear more meaningful than they are. 

Large percentage of results at or below the LOQ, with no result more
than 10 times the LOQ. The uncertainty of results increases as their
value decreases. This effect is substantial for results less than the
LOQ.

Other possible factors include:

Possibility of sample loss at the higher flow rate or during longer
sampling times associated with the experimental sample method.

Normal variations in industrial processes (a few inches difference in
sampling position can affect the results).

Higher-than-expected blank sample values.

Air Monitoring Results

Diacetyl and Acetoin

All samples were analyzed for diacetyl and acetoin; however, OSHA has
not published an exposure limit for either diacetyl or acetoin. For this
reason, all results are presented relative to each other, rather than
relative to a fixed standard. 

Table 3 (at the end of this section) provides details of all personal
breathing zone (BZ) samples obtained at this site. ERG collected seven
partial-shift BZ samples of 50 to 279 minutes duration using the
modified OSHA PV2118 method. Table 4 summarizes these personal BZ
results. All BZ partial-shift TWA results were 0.054 ppm diacetyl or
less for the period sampled, with this highest result associated with a
flavoring workerwho was grinding beans outside the flavor room door (no
acetoin detected in this sample). For acetoin no BZ partial-shift result
exceeded 0.039 ppm, a value associated with the sanitation/housekeeping
worker (no diacetyl detected in this sample).

Three area samples of 69 to 218 minutes duration were also collected and
these results appear in Table 5. Figure A diagrams the time frame of
each sample (personal and area) in relation to the morning and afternoon
flavor mixing sessions.  

Additionally, ERG obtained two other BZ results using only the higher
airflow rate method, both associated with flavoring workers flavoring
coffee. The first, a 31-minute STEL of 0.008 ppm diacetyl and 0.116 ppm
acetoin (the highest value reported at this facility) associated with
flavor mixing activities of Flavoring Worker #1. Interestingly, the
other result, a 114-minute sample for Flavoring Worker #3, who performed
a similar coffee flavoring procedure, was notably lower at 0.001 ppm
diacetyl and no acetoin detected. Table 6 summarizes these results,
along with an area sample result also obtained only with the higher
airflow method. The details associated with these values (sample volume,
LOQ) appear in Tables 3 and 5.

These sample results are characterized by generally low values (in the
parts per billion range) that show few consistent trends. Diacetyl
results tend to be somewhat higher for employees in the flavoring worker
(0.011 to 0.054 ppm) and packaging worker (0.032 ppm) job categories
than for the sanitation worker (ND) or maintenance worker (0 to 0.003
ppm). Interestingly, this trend is reversed for acetoin – for which
the sanitation worker (0.039 ppm) and maintenance worker (0.035 ppm)
were associated with higher levels. Area sample results also tend to be
associated with higher acetoin values (0.034 to 0.077 ppm) and lower
diacetyl values (0 to 0.035). An interesting similarity exists between
acetoin results for the sanitation worker (0.039 ppm) and for the flavor
storage area (0.034 ppm), near which this employee spent much of the
sampling period – but the correlation is not evident for the
associated diacetyl results.

One surprising observation is that partial-shift flavoring worker
acetoin exposure levels tend to be slightly higher during coffee
grinding compared to flavor mixing. In fact, the highest value for this
job category (0.054 ppm) is associated with 131-minute sample for an
employee working outside the flavor room door, who performed only
grinding during the sampling period. It is not possible from these
results to determine the reason for this effect, although one
consideration is possible sample loss during a brief period of elevated
humidity created when employees who performed flavoring used hot water
to rinse flavoring equipment. The OSHA PV2118 method may be affected by
high humidity; however, the temperatures and humidity were not
particularly high on the day of this sampling. A second possibility is
that vapors are carried from the flavor room to the grinding area by the
observed air flow pattern, in which case the grinding station near the
flavor room door would be the position most affected. ERG notes that the
highest result for acetoin in the flavor room (0.077 ppm) occurred not
during flavor mixing, but during the 146-minute period at the beginning
of the day, while the room was unoccupied; however, the corresponding
diacetyl concentration (ND) does not support this as the source of
grinder exposure during the early morning hours. 

In all cases the designation of “higher” or “lower” is relative
only to the other values recorded for this facility. In general, most of
the results are close enough to (or below) the LOQ and otherwise
affected by possible sources of uncertainty that even the relative
differences observed should be considered tentative. 

Table 3.  Individual Air Monitoring Results for Food Manufacturing
Facility A – Coffee Roaster



Sample

No.1	

Sample Type	

Sample

Duration

(minutes)	

Sample Volume

(liters)	Diacetyl Results	Acetoin results





Diacetyl 

Conc.2

(ppm)	

LOQ3

(ppm) 	

Result  above LOQ ?	Acetoin 

Conc. 2

(ppm) 	

LOQ4

(ppm) 	

Result less than LOQ ?

Flavoring Worker #1

A-2A	TWA 1	278	44.3	0.027	0.003	Yes	0.015	0.006	Yes

A-2B	 “  “	278	14.7	0.011	0.009	Yes	0.009	0.019	< LOQ

A-11A	TWA 2	131	19.9	0.018	0.007	Yes	0.001	0.014	< LOQ

A-11B	“  “	131	7.5	0.054	0.018	Yes	ND*	0.037	< LOQ

A-8A	STEL	31	5.9	0.008	0.023	< LOQ	0.116	0.047	Yes

A-8B	(B sample not collected)	NA**	NA	NA	NA	NA	NA	NA	NA

Flavoring Worker #2

A-10A	TWA	223	44.6	0.014	0.003	Yes	ND	0.006	< LOQ

A-10B	“  “	223	13.1	0.028	0.010	Yes	ND	0.021	< LOQ

Flavoring Worker #3

A-12A	TWA	114	21.9	0.001	0.006	< LOQ	ND	0.013	< LOQ

A-12B	(B sample not collected)	NA	NA	NA	NA	NA	NA	NA	NA

Packaging Worker

A-4A	TWA	249	49.8	ND	0.003	< LOQ	ND	0.006	< LOQ

A-4B	“  “	249	14.6	0.032	0.009	Yes	0.027	0.019	Yes

Sanitation/Housekeeping Worker

A-1A	TWA	252	152	ND	0.003	< LOQ	ND	0.007	< LOQ

A-1B	“  “	252	57.2	0.001	0.009	< LOQ	0.039	0.019	Yes

Maintenance Worker

A-3A	TWA 1	294	43.5	0.001	0.003	< LOQ	0.004	0.006	< LOQ

A-3B	“  “	294	15.8	0.003	0.008	< LOQ	0.003	0.018	< LOQ

A-13A	TWA 2	50	7.4	ND	0.018	< LOQ	0.002	0.038	< LOQ

A-13B	“  “	50	2.7	ND	0.049	< LOQ	0.035	0.104	< LOQ

1          Samples are listed by worker. Some workers were sampled more
than one time; as a result some of the sample numbers are out of
sequence. Samples with numbers ending in A were obtained at a nominal
sampling air flow rate of 150 to 200 milliliter (ml)/minute and are
considered experimental (EXP). Those with numbers ending in B were
obtained at the recommended nominal flow rate of 50 ml/minute.       

2	Sample concentrations are based on the period monitored (i.e., the
sample duration).  Samples were reported based on any amount of analyte
that was found; reporting

 	limits were not considered.

3	The reporting limit (limit of quantification) for diacetyl is 0.468
micrograms (μg) per sample.

4	The reporting limit (limit of quantification) for acetoin is 1.008 μg
per sample.

< LOQ = sample concentration was detected below the limit of
quantification for the specific analyte.

*An ND result was reported by the laboratory when there was no
indication of the analyte in the sample. 

** An NA indicates that no sample was obtained or analyzed (this is a
placeholder row).

Abbreviations

Conc. =	Concentration.

LOQ = Limit of quantification.

NA = Not Applicable.

ND = Not detectable.  

TWA = Time weighted average (for period sampled).

STEL = Short term exposure limit.







Table 4.  Summary of Employee Activities and Average/Time-Weighted
Average Air Monitoring Results 

for Food Manufacturing Facility A – Coffee Roaster



Sample

No.	Sample

Duration

(minutes)	

Diacetyl

Conc.1  (ppm)	

Acetoin

Conc. 1 (ppm)	

Employee Activities

Flavoring Worker #1

A-2B  	278	0.011

	0.009*

	Mixed flavor with beans and ground coffee to produce 3 batches in 30
minutes. Two of the three flavorings added to coffee contained 0.5 and
2.7 percent diacetyl, the third flavor contained no diacetyl. Cleaned
equipment and swept room. Grinding beans and processing QA samples for
the remainder of the sampling period.

A-11B  	131	0.054

	ND*

	Grinding beans and processing QA samples after lunch.

TWA for 2B & 11B	409	0.024	0.006*	As above.

Flavoring Worker #2

A-10B 	223	0.028

	0 (ND)*	Grinding beans in a grinder adjacent to the flavor room door
(just outside the flavor room), including period of the afternoon
flavoring session performed by Flavoring Worker #3. Collected QA Samples
and analyzed them in the laboratory. Was not involved directly in
flavoring on this day.

Packaging Worker

A-4B	249	0.032

	0.027

	Operated flavored coffee packaging equipment, bagging ground coffee
produced with a flavoring agent containing 0.5 percent diacetyl in
flavor (resulting in <0.01% diacetyl in ground coffee product). Swept
area around packaging machine.  Packaged another batch of the same
product (same diacetyl content) during the unsampled portion of the
shift.

Sanitation/Housekeeping Worker

A-1B	252	0 (ND)*

	0.039

	Cleaned facility with broom, brush and dustpan. Moved throughout the
facility, including in work in the flavor storage area.  Cleaned indoor
perimeter of plant during remainder of shift.

Maintenance Worker

A-3B	294	0.003*

	0.003*

	In morning, worked in maintenance shop, attended meeting in conference
room, adjusted packaging equipment (including flavor coffee packaging
used by the monitored packaging worker).

A-13B	50	0 (ND)*	0.035*

	In afternoon, worked on flavored coffee packaging equipment for 50% of
this sampling period (different packaging line than the monitored
packaging worker).

TWA for 3B & 13B	344	0.003*	0.008*	As above.

1	Based on results presented in Table 3.

*  	  Indicates that the result (or for composite samples, at least one
of the averaged results) is below the limit of quantification (LOQ) and
therefore of limited precision.    The associated LOQs are presented in
Table 3.

TWA means time-weighted average for the period monitored.

STEL means short term exposure level.

An ND result was reported by the laboratory when there was no indication
of the analyte in the sample. For averaging purposes ERG records this as
0 ppm.







Table 5.  Other Air Monitoring Results for Food Manufacturing Facility A
– Coffee Roaster



Sample

No.	Sample

Duration

(minutes)	

Diacetyl

Conc.1  (ppm)	

Acetoin

Conc. 1 (ppm)	

Employee Activities

Flavoring Worker #1

STEL

A-8A

	31	0.008*

	0.116	Short-term sample obtained during morning flavoring session.
Mixed flavor with beans and pre-ground coffee to produce 3 batches in 30
minutes. Two of the three flavorings added to coffee contained 0.5 and
2.7 percent diacetyl, the third flavor contained no diacetyl.

Flavoring Worker #3

A-12A	114	0.001*	0 (ND)*	Weighed flavor for and then mixed three batches
of flavored whole beans (two of the flavors contained diacetyl at 0.05
and 1.3 percent respectively, the third contained no diacetyl). Cleaned
weighing container, mixing equipment, and flavoring room. Process took
approximately 15 minutes per batch (45 minutes total flavoring
activities). Ground beans and processed QA samples for the remainder of
the shift (not monitored). This sample was originally anticipated to be
a STEL sample, but flavoring activities took longer than anticipated.
This sample represents setup, a complete triple flavoring cycle using
only the flavor mixing machine on the mezzanine, followed by cleanup.

Area Sample (flavor storage area, 4-feet above ground, on top of tote,
all morning)

A-7A	96	0.007*	0.036	Same as sample A-12A, without the final cleanup
activities.

1	See Tables 3 and 6 for sample details.

*  	  Indicates that the result (or for composite samples, at least one
of the averaged results) is below the limit of quantification (LOQ) and
therefore of limited precision.    The associated LOQs are presented in
Table 3.

STEL means short term exposure level.





Table 6.  Area Air Monitoring Results for Food Manufacturing Facility A
– Coffee Roaster



Sample

No.1	

Job Title or 

Area Sample Location	

Sample

Duration

(minutes)	

Sample Volume

(liters)	Diacetyl Results	Acetoin results





Diacetyl 

Conc.2

(ppm)	LOQ3

(ppm) 	Result  above LOQ ?	Acetoin 

Conc. 2

(ppm) 	LOQ4

(ppm) 	Result  above LOQ ?

Area sample (unoccupied flavor room,  6 feet above ground, on top of
computer console, morning, prior to flavoring activities)

A-5A

146	23.1	0.01	0.006	Yes	0.024	0.012	Yes

A-5B

146	8.5	ND	0.016	< LOQ	0.077	0.033	Yes

Area sample (occupied flavor room,  Same location as sample 5, starting
with morning blending and continuing to, but not including afternoon
blending) 

A-9A

218	34.5	0.011	0.004	Yes	0.010	0.008	Yes

A-9B

218	12.7	ND	0.010	< LOQ	0.044	0.022	Yes

Area sample (occupied flavor room, top of blending equipment on
mezzanine during period including afternoon preparation, flavoring
activities, and cleanup)  

A-7A	

	96	14.4	0.007	0.009	< LOQ	0.036	0.019	Yes

A-7B

NA	NA	NA	NA	NA	NA	NA	NA

Area sample (flavor storage area, 4-feet above ground, on top of tote,
all morning)

A-6A

216	31.8	ND	0.004	< LOQ	0.006	0.009	< LOQ

A-6B

216	12.2	0.035	0.011	Yes	0.034	0.023	Yes

Notes

1	Samples are listed by area. Some areas were sampled more than one
time; as a result some of the sample numbers are out of sequence.       


2	Sample concentrations are based on the period monitored (i.e., the
sample duration).  Samples were reported based on any amount of analyte
that was found; reporting

 	limits were not considered.

3	The reporting limit (limit of quantification) for diacetyl is 0.468
micrograms (μg) per sample.

4	The reporting limit (limit of quantification) for acetoin is 1.008 μg
per sample.

< LOQ = sample concentration was detected below the limit of
quantification for the specific analyte.

*   An ND result was reported by the laboratory when there was no
indication of the analyte in the sample.

** An NA indicates that no sample was obtained or analyzed (this is a
placeholder row).

Abbreviations

Conc. =	Concentration.

LOQ = Limit of quantification.

NA = Not Applicable

ND = Not detectable  

STEL = short term exposure limit



  Figure A.  Timeline of Sample Collection Periods (Grouped by
Employee or Area) and Flavor Mixing Sessions

 for Food Manufacturing Facility A – Coffee Roaster



 	 	Time  of day - a.m.	Time of day - p.m.

 	 	7	 	8	 	9	 	10	 	11	 	12	 	1	 	2	 	3	 	4

Flavoring Worker #1 (day shift)	 	Sample 2	 	 	 	 	 	 	 	 	 
 

	 	 	 	 	 	 	 	 	8	 	 	 	 	 	 	 	 	 	 	 

	 	 	 	 	 	 	 	 	 	 	Sample 11	 	 	 	 	 

Flavoring Worker #2 (split/bridge shift)	 	 	 	 	 	 	 	 	 	 
 	Sample 10

Flavoring Worker #3 (evening shift)	 	 	 	 	 	 	 	 	 	 	 	 
 	 	 

Sample 12

Packaging Worker (day shift)	Sample 4	 	 	 	 	 	 	 	 	 	 	 

Sanitation/housekeeping Worker (day shift)	Sample 1	 	 	 	 	 	 	 
 	 	 	 

Maintenance (day shift)	 	Sample 3	 	 	 	 	 	 	 	 	 

	 	 	 	 	 	 	 	 	 	 	 	 	 	13	 	 	 	 	 

Area - Flavor Room	 	Sample 5	 	 	 	 	 	 	 	 	 	 	 	 	 	 

	 	 	 	 	 	 	 	 	Sample 9	 	 	 	 	 

	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 

Sample 7

Area - Flavor Room	 	 	 	Sample 6	 	 	 	 	 	 	 	 	 

 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 	 

 	 	 	 	 	 	 	 	 	Flavor

 mixed	 	 	 	 	 	 	 	Flavor

 mixed	 



Other Analytes

Experimental monitoring for benzaldehyde was conducted at the same time
as each diacetyl/acetion sample. No benzaldehyde was detected in any
sample. The laboratory reporting limit (limit of quantification) for
benzaldehyde is 3.99 μg per sample. ERG had no information indicating
whether benzaldehyde was present in the flavors used that day. 

Examples of Controls in Place

Overview

Most controls were either administrative, secondary benefits of
workplace conditions, or built into process equipment. The following
existing controls (intentionally or unintentionally) may help reduce the
diacetyl and acetoin exposure of all employees at the plant.

This facility relies primarily on dilution ventilation, with notable
capacity for air exchange through exhaust fans in walls and exterior
cargo doors that are occasionally opened. The whole plant often has
significant air exchange, in part to help improve employee thermal
conditions. Due to cold weather, only about 25 percent of these fans
were operating on the day of the site visit. Even at this level of fan
operation, ERG had the impression of consistent air motion throughout
the plant. Only the roasting equipment is fitted with air quality
controls. All other fans exhaust directly outdoors.

Indoor temperatures and stored flavoring agents were cool (less than 60
degrees F for most of the shift) in this unheated plant on the day of
the air monitoring. This may have influenced diacetyl and acetoin
exposure levels, which might be higher during the hot summer months. 

An aggressive housekeeping program – ALL production workers spend some
portion of their shift cleaning – usually more than once per day. Full
time sanitation worker brushes and dusts (most dust estimated to contain
little or no flavor).

The surface area of exposed flavored coffee products is kept to a
minimum. Once flavored, coffee is contained in a plastic-lined super
sack and packaged as soon as possible.

Controls Associated with Individual Jobs

Flavoring Worker

An exhaust fan is present in the elevated section of the flavor room
ceiling, although this fan was not in use on the day of air monitoring.

Air conditioning is available in flavor room, set to prevent temperature
from exceeding 80 degrees F in the warm season.

Administrative process and work practices encouraging employees to
minimize time in flavor room. Instead they perform other related tasks
elsewhere during part of the coffee flavoring process.

Automated flavor dispensing system eliminates some manual weighing and
pouring. The use of large packages of flavor reduces the frequency with
which employees handle/transfer dispensing system components. This plant
purchases commonly used flavors in bulk packages that contain the amount
of flavor usually used in about 2 weeks.

Packaging Worker

Automated form/fill/seal packaging minimizes product handling and
maintains product relatively well enclosed throughout the packaging
process.

Discussion

Relationship between Results, Sources of Exposure, and Controls in Use

The active air exchange and cool temperatures in this plant, coupled
with the brief periods of exposure to flavors and low exposed surface
area of flavored coffee, may all contribute to the fact that all
partial-shift results were less than 0.054 ppm for diacetyl and 0.039
ppm for acetoin on this date. Nevertheless, flavoring workers are more
likely than other employees to experience exposures above the LOQ, both
during flavor mixing and when performing grinding outside the flavor
room door. Other trends were not clearly established with these results.

These results are likely representative of typical work during cool
weather and during the winter and spring manufacturing phase of the
cycle. Both the warmer temperatures and increased production of flavored
coffees in late summer and fall mean that exposure levels could be
higher during those periods than were measured during this site visit.  

Other Opportunities for Exposure Control 

If additional exposure control is desired for the jobs and situations
observed at Firm A, the following options might offer an additional
level of exposure management:

Ensure the container for accumulated ground coffee waste from cleaning
the flavor mixing equipment is kept tightly sealed.

Consider making periodic checks of the seals on bulk containers where
automatic dispensing equipment attaches, to ensure the container seal is
tight. Some were observed to be deteriorating or loose during the site
visit. Flavor odors were strong near some bulk flavor containers (these
were not necessarily flavors containing diacetyl).

During rinsing activities, consider carefully pouring or using a
connecting hose to direct rinse water that contains residual flavor
directly down drain. This step would reduce the size of the puddle (hot
water that contains flavor) on the washroom floor. Another possible
option is to use a cold water rinse prior to hot water wash to reduce
the quantity of flavor components available to volatilize on contact
with the hot water.

Room air appeared to flow from the flavor room out to the grinding area.
More consistent use of the ceiling exhaust fan might reverse that trend.


Air sampling results suggest that acetoin (and possibly other) vapors
build-up in the flavor room during periods of disuse. Steps that reduce
vapors escaping from bulk flavor containers, increase air exchange
during unoccupied periods, and exhaust the flavor room air (such as
turning on the exhaust fan for a few minutes before entering) would
limit airborne acetoin concentrations as employees enter. Additionally,
there could be benefits associated with steps that minimize airflow from
the flavor room to the grinding area. ERG notes, however, that a
complicating factor is that potential exposure reduction due to the
cooling influence of the air conditioning in the flavor room could be
rendered less effective if warm air is pulled in from the grinding room.


ERG used a simple air direction test and determined that the cafeteria
is under slight negative pressure compared to surrounding spaces –
plant air enters the cafeteria through door louvers and each time the
door is opened. One option would be to increase flow of outside air into
the cafeteria so that the space is under positive pressure compared to
the surrounding production space. 

 For example, if in one day a facility produced four batches of coffee,
each flavored with 8 pounds of a hypothetical but realistic flavoring
containing 1.7 percent diacetyl, the daily total diacetyl use would be: 
4 batches x 8 pounds x 0.017 fraction diacetyl = 0.54 pounds of
diacetyl, or about a half a pound.

 For readers who are not familiar with form/fill/seal packaging, see
background information, photos, general operation information and future
trends on form/fill/seal packaging equipment (generally similar but not
necessarily identical to that used at Food Manufacturing Facility A) at
the following sources. 

Description:   HYPERLINK
"http://en.wikipedia.org/wiki/Vertical_form_fill_sealing_machine" 
http://en.wikipedia.org/wiki/Vertical_form_fill_sealing_machine .

Video clip:   HYPERLINK "http://www.youtube.com/watch?v=cZZ0F4Tz5X0" 
http://www.youtube.com/watch?v=cZZ0F4Tz5X0 . 

Discussions:   HYPERLINK
"http://www.fresco.com/green_mountain_coffee.html" 
http://www.fre捳⹯潣⽭牧敥彮潭湵慴湩损景敦⹥瑨汭―猨
敥瀠潲散獳搠獥牣灩楴湯椠⁮潢瑴浯琠楨摲漠⁦桴⁥牡
楴汣⥥※–奈䕐䱒义⁋栢瑴㩰⼯睷⹷楣捲敬慰正条湩⹧
潣⽭牄卹汯摩牐摯捵獴栮浴≬Ġ栔瑴㩰⼯睷⹷楣捲敬慰
正条湩⹧潣⽭牄卹汯摩牐摯捵獴栮浴ᕬമ഍̍഍ഄ̍഍
ഄግ䅐䕇†㈔ᔸ

@

ª

 

t

Ô

	

	

ᔛ煨蕯ᘀ奨嵄　ቊ洀H渄H甄Ĉ⌀

-



 

:

;

<

=

>

?

@

A

B

^

_

`

a

e

f

ˆ

‰

Š

¤

¥

¦

§

¨

©

ª

«

¬

È

É

Ê

Ë

ᔛ煨蕯ᘀ奨嵄　ቊ洀H渄H甄Ĉᘑ奨嵄洀H渄H甄Ĉ⌀Ë

Ï

Ð

á

â

ã

ý

þ

ÿ

-



 

!

"

>

?

j;

ᔀ煨蕯ᘀ奨嵄　ቊ唀Ĉ䡭Ѐ䡮Ѐࡵ⌁?

@

A

G

H

Q

R

S

m

n

o

q

r

s

t

u

v

’

“

”

•

›

œ

±

²

³

Í

Î

Ï

Ñ

Ò

Ó

Ô

Õ

Ö

ò

j/

ÿࡵ⌁ò

ó

ô

õ

û

ü

j

hID

hID

hC

hC

hC

>

>

>

>

㐀ۖĀ̊Q┃曾Ĵ

  h[C

h[C

h[C

>

㐀ۖĀ̊Q┃曾Ĵ

hrH

 h<

h<

h<

" híQ

kd

ਁᴃ愀϶Ġ

摧䎧Z

ਁᴃ愀϶Ġ

" híQ

摧ᱧ

'g

Ffä

 h<

h<

h<

Site Visits Related to Diacetyl –Food Manufacturing Facility A        
                                               

Example of How Flavored Coffee Is Made

Flavors are typically added to roasted beans before they are ground. The
beans are placed in a large mixer which is specially designed to gently
tumble the beans without causing them damage. Examples of this type of
mixer include ribbon blenders, drum rotators, and candy pan coaters. The
flavors are usually introduced via a pressurized spray mechanism which
breaks the oils into tiny droplets which allows for better mixing. Oils
must be added to the beans very gradually to guard against areas of
highly concentrated flavor called hot spots. The beans are agitated for
a set amount of time to ensure the flavor is evenly spread. This process
may take 15-30 minutes, depending on the batch size and mixing
characteristics of the oil. When the beans are properly coated, they
take on a glossy finish that indicates a uniform distribution of oils.

It is also important to note that, instead of flavoring whole beans,
flavors in dry form can be blended with ground coffee. In such cases,
the flavors are encapsulated in starch or some other powdered matrix.
There is enough moisture in the coffee to promote transfer of flavor and
color from the encapsulated flavors to the coffee grounds in about 24
hours after mixing.

Source: Flavoring Coffee, from How Products Are Made    HYPERLINK
"http://www.madehow.com/Volume-3/Flavored-Coffee-Bean.html" 
http://www.madehow.com/Volume-3/Flavored-Coffee-Bean.html 

 

