THE JOURNAL OF PROSTHETIC DENTISTRY WINKLER ET AL

Comparison of four techniques for monitoring the setting kinetics of gypsum

Mark M. Winkler, DDS, MS, PhD,a Peter Monaghan, DDS, MS, PhD,b
Jeremy L. Gilbert, PhD,c and Eugene P. Lautenschlager, PhDd
Indiana University, Indianapolis, Ind.; The Ohio State University, Columbus,
Ohio; and Northwestern University, Chicago, Ill.

Statement of problem. Setting time of gypsum depends on the method of measurement.

Purpose. In this study, four methods for ascertaining the setting time of gypsum were compared.
Material and methods. Gypsum setting was evaluated with Gillmore needles, setting expansion,
scanning electron microscopy, and x-ray diffraction. Both die stone (Vel-Mix) and fast-setting dental stone
(Snap-Stone) were investigated. By using Gillmore needles, both initial set and final set were recorded.
Setting expansion was monitored until equilibrium was achieved. For scanning electron microscopy,
samples of mixed stone were immersed into liquid nitrogen to stop the reaction. The water was subse-
quently removed by freeze-drying. Samples were then examined in a scanning electron microscope. For
x-ray diffraction, the percentages of dihydrate and hemihydrate crystals were monitored every 1.5 minutes.
Results. Setting times for the fast-setting type III stone and for the regular-setting type IV stone were
approximately 3 and 10 minutes, respectively, with the Gillmore needle indentation tests; 10 and 20
minutes with scanning electron microscopy; and 20 and 60 minutes with x-ray diffraction.
Conclusion. Scanning electron microscopy, setting expansion, and x-ray diffraction indicated changes
that occurred at times after a clinically relevant set was obtained in the gypsum setting reaction. (J Prosthet
Dent 1998;79:532-6.)

CLINICAL IMPLICATIONS
Scanning electron microscopy, setting expansion, and x-ray diffraction all indicate
changes that occur at times after a clinically relevant set is obtained in the gypsum
setting reaction. If disturbed after this clinically relevant set but before final set as
determined in this study, the developing gypsum microstructure may be adversely af-
fected and result in weakened, easily abraded, or distorted gypsum models, which can
ultimately lead to a substandard clinical outcome for procedures that depend on these
casts.

G ypsum products are used in many dental mate-

rials. This exothermic reaction can be represented as seen
set dental stone, has been characterized in various ways
such as heat of reaction, changes in crystal structure,
in the following equation1-3: and speed of reaction. Both the initial and final forms,
hemihydrate and dihydrate, of this reaction have been
CaSO4 · ½ H2O + 3/2 H2O ⇔ CaSO4 · 2 H2O +
photomicrographed previously.1-4 Intermediate stages
3900 cal/g mole1
have recently been recorded with the scanning elec-
Hemihydrate + Water → Dihydrate + Heat tron microscope and a freeze-drying technique.5-7
(dental stone) (gypsum) Mahler and Asgarzadeh8 explained the discrepancy
The reaction of calcium sulfate hemihydrate, dental between the true volumetric contraction during setting
stone powder, plus water to calcium sulfate dihydrate, and the typical final volumetric expansion by the forma-
tion of pores within the stone. Lautenschlager and
Corbin9 found an exponential relationship between the
This research was supported in part by grant NIDR T32 DE07042 micropores and the amount of expansion and hypoth-
from the National Institute of Dental Research, Bethesda, Md.
a
Assistant Professor of Dental Materials, Department of Restorative
esized that these micropores must be created as a result
Dentistry, School of Dentistry, Indiana University. of the impingement of the growing crystals. In 1969,
b
Assistant Professor, Department of Restorative Dentistry, Section of Lautenschlager et al.10 reported on the proportioning
Restorative Dentistry, Prosthodontics, and Endodontics, College of calcium sulfate hemihydrate versus dihydrate x-ray
of Dentistry, The Ohio State University. diffraction peaks during the setting of dental stones. They
c
Associate Professor of Biological Materials, Division of Biological
Materials, Northwestern University.
found that the reaction was not nearly complete at the
d
Professor and Director, Division of Biological Materials, Northwest- final setting time determined by the Gillmore needle
ern University. and that this setting time became longer as the water/

532 THE JOURNAL OF PROSTHETIC DENTISTRY VOLUME 79 NUMBER 5

WINKLER ET AL THE JOURNAL OF PROSTHETIC DENTISTRY

powder ratio increased. Other studies reported that set-

ting time can be decreased by using either slurry water11
or dry dental stone powder12 in the gauging water.
The purpose of this qualitative study was to monitor
the crystal growth kinetics during the setting of both a
die stone and a fast-setting type III dental stone with
various methods. In addition to the Gillmore needle in-
dentation tests, x-ray diffraction and setting expansion
measurements were monitored to qualify morphologic
structures seen in images obtained of setting stone us-
ing SEM.
MATERIAL AND METHODS
Two dental gypsum products were examined. A com-
Fig. 1. Setting expansion apparatus. A, Wire cables; B, input
mon regular-setting type IV or die stone (Vel-Mix, Kerr
and output windings of LVDT; C, stylus of LVDT; D, gypsum
Manufacturing Co., Romulus, Mich.) and a fast-setting specimen with glass plate; E, rigid support.
type III dental stone (Snap-Stone, Whip Mix Corp., Lou-
isville, Ky.) were selected for comparison. The fast-
setting stone was chosen to demonstrate how the vari- to indent the surface, this was recorded as the final
ous setting time measurements change when the setting set.
time of the gypsum is greatly decreased. While not rec-
X-ray diffraction
ommended by the manufacturer, this fast-setting stone
was mechanically mixed to ensure not only a uniform Mixes of the two dental stones were prepared as pre-
mix but also to allow direct comparison with a die stone, viously described. Immediately after mixing, the setting
which is typically mixed mechanically, because mechani- dental stone was quickly loaded into the powder cham-
cal mixing decreases setting time. Samples of each gyp- ber of an x-ray diffractometer (XRG 3100 with APD
sum product were fabricated and studied in the follow- 3720 Controller, Phillips Electronic Instruments,
ing manners. Mahwah, N.J.). The diffractometer used a copper tar-
get operating at 40 kV and 15 mA and a Ni filter to
Setting expansion produce CuKα radiation. The 20.78 degrees - 2 Θ peak,
A small quantity, 100 g, of stone was incorporated attributed to the dihydrate and the 25.12 degrees - 2 Θ
into 24 ml of tap water. After the calcium sulfate hemi- peak, produced by the hemihydrate were scanned at 1.5
hydrate was completely wetted, the dental stone and minute intervals. Scanning of the peaks commenced
water combination were mechanically mixed under within 2 minutes from the beginning of mixing and con-
vacuum (Vac-U-Spat, Whip Mix Corp.) for 10 seconds. tinued for 1 hour. Scans were also made on the unreacted
The specimen holder was a metal cylinder, with dimen- powders and on the reacted materials 24 hours after
sions of approximately 15 mm in length and 31 mm in mixing began.
diameter, which was placed on a glass slab. A thin piece Scanning electron microscopy
of cover glass was placed on top of the mixture in the
cylinder, and a thin rod containing the core of the linear Additional mixes of each of the dental stones were
variable differential transformer (LVDT) was set atop prepared as described previously. Portions of this mixed
the cover glass (Fig. 1). As the gypsum specimen ex- mass were vibrated into several identical sample hold-
panded, the core moved upward in relationship to the ers. At selected times, ranging from 1 minute to 1 day,
housing of the LVDT and caused a change in the volt- after the incorporation of the powder into water, the
age passed from the input windings to the output wind- samples were freeze-dried as described earlier.5,7 Only
ings. The output of the LVDT was monitored with a the air-exposed surfaces, or the top surfaces, of the
strip chart recorder (model 1242, Soltec, San Fernando, samples were examined.
Calif.) for 1 hour. The output had been previously cali- All freeze-dried samples of dental die stone were kept
brated with a linear micrometer. in a desiccator until being prepared for the SEM. The
samples of dental die stone were sputter-coated with gold
Gillmore needle (Hummer II, Technics Inc., Alexandria, Va.) and then
When a small Gillmore needle, weighing ¼ pound examined in a SEM (S120, Cambridge Instruments Ltd.,
with a 1/ 12 inch diameter tip, failed to indent the sur- Cambridge, Mass.) in the backscatter mode. Operating
face of the setting expansion specimen, this was re- conditions for the SEM were 30 kV, a working distance
corded as the initial set. When a large Gillmore needle, of 16 to 17 mm, and a magnification of approximately
weighing 1 pound with a 1/24 inch diameter tip, failed 3000 for all micrographs.

MAY 1998 533

THE JOURNAL OF PROSTHETIC DENTISTRY WINKLER ET AL

Fig. 2. Setting expansion curves for regular-setting type IV

stone (dotted line) and for fast-setting type III stone (solid line).

RESULTS
Gillmore needle
The fast-setting gypsum product set too rapidly to
accurately obtain an initial set, but the final set was mea-
sured as 2.5 minutes. For the regular-setting type IV
gypsum, the initial and final setting times were 7.5 min-
utes and 9.25 minutes, respectively.
Setting expansion
The percentage expansion (% setting expansion) was
calculated using the equation:
Fig. 3. X-ray diffraction data. A, Regular-setting type IV stone.
% setting expansion = [Vertical travel ÷ Specimen B, Fast-setting type III stone. (Dihydrate peaks = solid circles;
height] × 100 Hemihydrate peaks = open circles; GI = Gillmore needle [ini-
tial set]; GF = Gillmore needle [final set]; P = photograph [final
where the height of the specimen and vertical travel
set]; S = setting expansion peak.)
of the stylus of the LVDT were measured to the nearest
0.01 mm. The expansion curves are presented in
Figure 2. Both materials appeared to initially contract. marked GI and GF represent the initial and final setting
Then, at approximately their Gillmore needle final set times, respectfully, with the Gillmore needles. The dashed
times, they began to expand and release heat (which was line marked P designates the photographic setting time,
sensed but was not quantified). Later the specimens the time corresponding to the first photomicrograph that
cooled and shrank, and by 1 hour had reached a con- did not appear to be different from final setting as shown
stant height, which represented approximately 0.15% in the 24-hour photomicrograph. The dashed line
expansion for the fast-setting stone, and 0.18% for the marked S designates the setting expansion setting time,
regular-setting die stone. Of course, the curve minimum the time at which the setting expansion peaks. For each
and maximum occurred in much shorter time spans for stone type, the intensity due to the hemihydrate peak
the fast-setting stone. diminishes with time and the intensity of the dihydrate
peak increases with time. The fast-setting stone seems
X-ray diffraction
to have reacted almost completely by approximately 20
Figure 3 shows the changes in intensity of the hemi- minutes, whereas the regular-setting stone seems to have
hydrate and dihydrate peaks of the regular-setting die fully reacted by about 60 minutes.
stone (A) and the fast-setting stone (B) over time. Stan- Figures 4 and 5 are photomicrographs of the two
dard deviations for each data point were less than 10% stones for intermediate stages relating to the final set-
of the mean values and for reason of clarity were not ting times (10 minutes and 20 minutes) determined with
included. The hollow markers represent the intensity due the methods.7 A full set of photomicrographs detailing
to the hemihydrate peaks and the solid figures represent the setting reaction of both a regular-setting (type IV)
the intensity due to dihydrate peaks. Dashed lines in Fig- die stone and a fast-setting (type III) dental stone can
ure 3 represent various setting times. Dashed lines be found in this earlier article.

534 VOLUME 79 NUMBER 5

WINKLER ET AL THE JOURNAL OF PROSTHETIC DENTISTRY

Fig. 4. Scanning electron photomicrograph of regular-setting Fig. 5. Scanning electron photomicrograph. A, Fast-setting type
type IV stone. A, At 10 minutes and B, 20 minutes after mix- III stone at 3 minutes; B, 10 minutes after mixing.7 (Original
ing.7 (Original magnification ×3000.) magnification ×3000.)

For regular-setting die stone, the reaction of water Figure 5 illustrates the fast-setting type III gypsum
with the hemihydrate crystals resulted in the nucleation product for 3 and 10 minutes after mixing.7 The reac-
of numerous particles that appeared to nucleate on or tion of water with the hemihydrate crystals at 1 minute
near the original powder particles and grow until at 10 resulted in needle-like crystals, approximately 5 to 10
minutes (Fig. 4, A); the developing crystals were rather µm in length. These crystals were already slightly larger
thick but not near their final length. The large (5 to 10 than the initial prismatic hemihydrate crystals, before
µm in diameter) initial cuboidal hemihydrate particles any water was added. By 3 minutes (Fig. 5, A), the
were still evident. By 20 minutes (Fig. 4, B), the devel- needles were much more numerous and a bit thicker,
oping particles grew sufficiently large to vaguely resemble whereas the original particles appeared to be nearly con-
the familiar crystal pattern of set gypsum; however, the sumed. At 10 minutes (Fig. 5, B), the needles were less
large original particles were still visible. They have al- numerous but definitely thicker. The microstructure
most entirely disappeared by 60 minutes, whereas the observed at 10 minutes was essentially unchanged at 24
growing dihydrate crystals have interdigitated with one hours.
another in a mesh. Besides the usual elongated prismatic
DISCUSSION
crystals, plate-like formations can be seen. Little differ-
ence can be discerned between the microstructures at The gypsum structures studied continued to change
20 minutes, 60 minutes, and at 24 hours, for the freeze- well after the stones had hardened and the Gillmore
dried, air-exposed gypsum surfaces. Changes in struc- needle final set was achieved (Fig. 2). As shown in the
ture were still seen well after the final set obtained with same figure for both materials, the final Gillmore set
the Gillmore needle. appeared to occur at the end of contraction and at the

MAY 1998 535

THE JOURNAL OF PROSTHETIC DENTISTRY WINKLER ET AL

time of the onset of the setting expansion. Although the obtained, SEM, setting expansion, and x-ray diffraction
resistance to indentation and the expansion due to im- all indicate changes that occurred at longer times in the
pingement of neighboring crystals may be closely re- gypsum setting reaction sequence.
lated, an indenter of dissimilar weight and/or another
configuration probably would have halted penetration
CONCLUSIONS
at a different time. The Gillmore needle indentation test The photomicrographically documented transforma-
may best determine clinical handling times but does not tion from calcium sulfate hemihydrate crystals to an in-
accurately predict the end of the setting reaction. Well terdigitating mesh of dihydrate crystals was compared
after the Gillmore needle final set, the SEMs indicated with setting expansion, x-ray detected percentage of crys-
continuing growth of structures, the LVDT indicated tal structure, and Gillmore needle final set. Dimensional
onset and furthering of expansion, and the x-ray moni- changes agreed most closely with the x-ray diffraction
tored definitive changes in increasing dihydrate and de- data that denoted the percentage of dihydrate crystal
creasing hemihydrate crystallographic structures. formation during setting. Final setting times, when
In the situation of the LVDT, contraction and expan- ranked in order from shortest to longest, were Gillmore
sion during setting results from a combination of pow- needle indentation, scanning electron microscopy, set-
der wetting, crystal growth, expansion from exother- ting expansion, and x-ray diffraction.
mic heat of reaction, and contraction due to cooling to If the setting gypsum is disturbed in any way after the
room temperature. As the heat of reaction dissipates and clinically relevant set before final set, as determined by
as crystal formation lessens, LVDT setting expansion the parameters in this study, the still developing micro-
decreases until finally a small contraction results when scopic structure of the gypsum may be adversely affected
the specimens return to room temperature. However, and result in problems such as weakened, easily abraded,
the net dimensional change is still expansion (Fig. 2). or distorted casts. The use of such casts to fabricate res-
On the basis of the x-ray diffraction data (Fig. 3), the torations or appliances for patient therapy may result in
conversion of calcium sulfate hemihydrate to dihydrate a decreased quality of care and/or a reduced service life.
continues not only beyond the Gillmore final set but
also even beyond what appears to be the final structural REFERENCES
change as determined from the SEMs. The setting ex- 1. Phillips RW, Moore BK. Elements of dental materials for dental hygienists
and assistants. 5th ed. Philadelphia: WB Saunders; 1994. p. 40-56.
pansion data correlate best with the x-ray diffraction data.
2. Anusavice KJ. Phillips’ science of dental materials. 10th ed. Philadelphia:
Both show activity until approximately 20 and 60 min- WB Saunders; 1996. p. 185-209.
utes, respectively, from the start of mixing. Although 3. Craig RG. Restorative dental materials. 10th ed. St Louis: Mosby–Year Book;
1997. p. 333-46.
the x-ray and SEM data are obtained only from the sur-
4. Phillips RW, Ito BY. Factors affecting the surface of stone dies poured in
face layers of the specimens, the setting expansion comes hydrocolloid impressions. J Prosthet Dent 1952;2:390.
from the entire bulk. The interior may become hotter 5. Winkler MM, Lautenschlager EP, Gilbert JL. Scanning electron microscopy
of setting gypsum. [Abstract 2073.] J Dent Res 1990;69(special issue):368.
and set faster than the surface. In any case, there are
6. Winkler MM, Monaghan P, Lautenschlager EP, Gilbert JL. Scanning elec-
thermal and degree of set gradients, making the results tron microscopy of a fast-setting gypsum. [Abstract 1332.] J Dent Res
dependent on the overall specimen shape and the sur- 1991;70(special issue):432.
7. Winkler MM, Monaghan P, Gilbert JL, Lautenschlager EP. Freeze-drying and
roundings that allow the heat to dissipate.
scanning electron microscopy of setting dental gypsum. Dent Mater
Photomicrographs of both stones revealed that most 1995;11:226-30.
observable changes occurred early in the setting reac- 8. Mahler DB, Asgarzadeh K. The volumetric contraction of dental gypsum
materials on setting. J Dent Res 1953;32:354-61.
tion. Only approximately 70% of the transformation from
9. Lautenschlager EP, Corbin F. Investigation of the expansion of dental stone.
hemihydrate to dihydrate according to x-ray diffraction J Dent Res 1969;48:206-10.
occurred at the time no further visible changes were seen 10. Lautenschlager EP, Harcourt JK, Ploszaj LC. Setting reactions of gypsum
materials investigated by x-ray diffraction. J Dent Res 1969;48:43-8.
on the SEMs. However, these SEMs were more qualita-
11. Brukl CE, McConnell RM, Norling BK, Collard SM. Influence of gauging
tive than quantitative. No attempts were made to re- water composition on dental stone expansion and setting time. J Prosthet
peatedly view the same sample or region, and no quan- Dent 1984;218-23.
12. von Fraunhofer JA, Spiers RR. Accelerated setting of dental stone. J Prosthet
titative measurements were taken, so there may have been
Dent 1983;859-60.
additional continued activities, such as crystal volume
increase, which simply were not detectable by such quali- Reprint requests to:
DR. MARK M. WINKLER
tative SEM observations.
DEPARTMENT OF RESTORATIVE DENTISTRY
Hence, four methods were used to track the setting SCHOOL OF DENTISTRY
of gypsum products. Because four different qualities of INDIANA UNIVERSITY
ROOM 112A
the gypsum setting process were monitored, and each
1121 W. MICHIGAN ST.
test used different sizes, shapes and locations (surface INDIANAPOLIS, IN 46202
vs. interior), exact correlation of final setting times with
Copyright © 1998 by The Editorial Council of The Journal of Prosthetic Den-
respect to method should not be expected. However, it
tistry.
should be clear that long after a clinically relevant set is 0022-3913/98/$5.00 + 0. 10/1/88848

536 VOLUME 79 NUMBER 5