Evaluacion de Los Polimeros Gelantes
Evaluacion de Los Polimeros Gelantes
Evaluacion de Los Polimeros Gelantes
LO~
Abstract
Polymer gels are attractive tools for stopping unwanted fluid production from oil and gas wells, but conventional gelant
formulations become increasingly difficult to apply at higher formation temperatures. In contrast phenol-formaldehyde
cross-linked polymer gelants remain viable to at least 140°C. Choice of polymer composition permits control of the gelation
delay, and the gelants display good injectivity even when the residence time approaches the gelation time. Cross-linking
occurs over a wide range of pH and is insensitive to lithology. Although retention of the gelant components on formation
rock is negligible, phenol strongly partitions into oil phases contacted by the gelant. Experiments and numerical modelling
show that a preflush of phenol is an effective method of compensating for phenol partitioning in the formation. © 1997
Elsevier Science B.V. All rights reserved.
tion. Thus finding a gelant that satisfies all the influence cross-linking of p h e n o l - f o r m a l -
criteria may be difficult. dehyde/polymer gelants, and characterize their in-
As temperature increases, the difficulty in meet- jectivity and propagation in porous media. These
ing any of these characteristics increases. For exam- gelants emerge from these tests as attractive candi-
ple, the principal gelants currently employed in the dates for high-temperature field applications.
industry [formed with polyacrylamide and Cr(III)-
acetate cross-linker (Sydansk and Moore, 1992;
Sydansk, 1993)] undergo gelation in a few hours or 2. Methods and materials
less at 70°C and within a few minutes at 100°C
(Albonico et al., 1994). Moreover, Cr(III)-acetate is 2.1. Materials
strongly driven to precipitation and retention in
porous media at elevated temperature (Albonico et The commercial polymers employed in these stud-
al., 1994; Bartosek et al., 1994). Thus the placement ies were a low-molecular-weight (500,000 g/tool)
of Cr(III)-acetate/polymer gels a significant distance hydrolyzed PAAm, a p o l y ( a c r y l a m i d e - c o -
into matrix rock (e.g., > 1 m from the wellbore, as acrylamido-3-propane sulfonic acid) copolymer
some applications might require) will be difficult to ( P A A m - A M P S ) , and an acrylamide/N-vinyl
achieve in higher-temperature formations. Various pyrrolidone/acrylamido-3-propane sulfonic acid ter-
methods have been developed to address these draw- polymer (PAAm-VP-AMPS). The freshwater (FW)
backs. These include cooling the formation with a gelants were prepared with distilled water. The syn-
pre-flush (Sanders et al., 1994) and employing thetic seawater (SSW) employed contained 10,818
chelating ligands to delay gelation at high tempera- ppm Na +, 448 ppm Ca -'+, and 1,288 ppm Mg 2+, all
ture and strongly reduce Cr(III) retention (Lockhart furnished as their chloride salts. Standard reagent-
and Albonico, 1992; Albonico et al., 1993, 1994; grade phenol and formaldehyde, the latter furnished
Bartosek et al., 1994). Although these techniques as a 35% aqueous solution, were employed.
provide some useful extensions of the applicability
of Cr(III)-polymer systems, these gelants do not 2.2. Bulk gelation experiments
presently appear to provide a general solution to the
problem of water shut-off in formations above 100°C. Gelants were prepared as described previously
With regard to gel stability (characteristic 5 above), (Albonico et al., 1993), the phenol and formaldehyde
on the other hand, it is possible to formulate bulk being added last, followed by adjustment of the pH
gels with commercial polymers which exhibit long- by addition of diluted solutions of HC1 and NaOH
term stability in hard brines (Albonico and Lockhart, (0.01-1 M). Gelants were aged either in screw-cap
1993). More importantly, persistent permeability re- vials, in which case they were initially air-saturated,
duction of gel-treated porous media can be achieved or in sealed glass ampoules after careful degassing to
to at least 140°C (Bryant et al., 1996). remove oxygen.
Among the several other polymer gelants de- The influence of pH on gelation was studied by
scribed in the literature, those formed with acryl- preparing gelants covering the pH range of interest.
amide polymers or copolymers using a combination These were periodically removed from the oven
of phenol and formaldehyde as cross-linkers (Falk, (every 1-2 h the first day, then gradually less fre-
1984; Moradi-Araghi et al., 1988) are reported to quently as the pH stabilized) and the pH readjusted
undergo relatively slow gelation (from several days using a minimum of acid or base solution. By this
to a week or more) at 90-100°C. These long gelation method the pH of the gelants was maintained to
delays, coupled with the reasonable expectation that within 0.5 pH unit of the initial value. An effort was
the phenol and formaldehyde cross-linkers will not made to minimize dilution of the samples during the
interact strongly with formation rock, suggested to us adjustment of their pH; the variation in their final
that these gelants might be good candidates for some concentrations was less than 5%. The pH of the fluid
high-temperature applications. We report here stud- solutions and gels was monitored using a Metrohm
ies that provide new insight into the factors that EA 125 microelectrode.
S.L. Bryant et al. / Journal of Petroleum Science and Engineering 17 (1997) 197-209 199
Filtration experiments on the partially gelled solu- solution was pre-filtered through a 5-1xm Millipore
tions were carried out by forcing the gelant through a filter, aged to 50% of its gelation time in bulk (24 h)
5-1xm Millipore or l-lxm Nuclepore filter at a con- at 120°C and then cooled before injection.
stant flow rate and monitoring for the rise in the One slim-tube experiment was carried out with a
injection pressure. Filtration of the gelants through 3 m × 0.8-cm i.d. stainless-steel tube dry-packed
the 1-~zm filter resulted in a noticeable decrease in with clean quartz sand (75-125 p~m) and fitted with
turbidity. The gelant solutions used in these experi- nylon endcap screens. After packing with sand, the
ments were filtered through 5-1xm Millipore filters slim-tube was coiled and saturated with SSW under
prior to being placed in the oven to cross-link. vacuum and placed in a forced air oven. The initial
permeability was 3300 roD. Phenol propagation in
2.3. Partitioning experiments the brine-saturated slim-tube was characterized at
120°C by injecting brine containing 1000 ppm phe-
To study the partitioning of phenol into crude oil, nol and 100 ppm of I tracer as KI. The phenol and
a 1000 ppm SSW solution of phenol and an equal I tracer were monitored simultaneously by UV at
weight of oil was mixed on a vortex mixer at ambi- 268 and 234 nm, respectively. The I absorption at
ent temperature until a fine emulsion was formed. 234 nm was corrected for the modest contribution of
The resulting emulsion was allowed to stand for phenol at this wavelength.
several minutes to an hour followed by centrifuga- The slim-tube was oil-saturated by flooding at a
tion to separate the phases. Control experiments differential pressure of 50 bar with 1 1 (16 PV) of a
showed that the equilibration of phenol between the light, paraffinic crude oil at ambient temperature.
two phases was essentially complete within the mini- This was followed by flooding with 1 1 of SSW at 50
mum time required for mixing and separating the bar and 120°C to establish residual oil saturation.
two phases. A similar procedure was employed for The p h e n o l / I - tracer flood was then repeated at
equilibrating 1000 ppm formaldehyde between SSW 120°C in order to determine both the water saturation
and the crude oils. Three light-to-medium crude oils and the phenol retention by the residual oil. Finally,
stored at ambient conditions were used. the slim-tube was flooded with a phenol-formal-
The quantity of phenol remaining in the aqueous dehyde/PAAm gelant at 120°C at an initial flow rate
phase after equilibration was determined by a UV of 30 ml/h. From the recorded pressure history, the
spectroscopic method, monitoring at 268 nm. Con- change in the apparent viscosity of the in-situ gelant
trol experiments showed that no components in the with time was calculated from Darcy's law.
crude oils contributed to the wavelength monitored. A second experiment was carried out with a
Formaldehyde analysis was carried out using a col- 6.6-m brine-saturated slim-tube. The gelant was in-
orimetric method based on chromotropic acid (Hauser jected at 0.08 ml/min at 140°C until the maximum
and Cummings, 1964). The detection limit of formal- injection pressure was reached (150 bar). At this
dehyde was better than 1 ppm. point the slim-tube was shut-in for 24 h and then
cooled to ambient temperature. The slim-tube was
2.4. Porous media experiments cut into pieces 10 cm in length, and their resistance
to flow was tested by applying a constant flow rate
General details of the equipment and procedures of brine (0.05 ml/min) for a few minutes.
used in the porous media experiments have been
given previously (Bartosek et al., 1994). Experi- 2.5. Numerical modeling
ments at elevated temperature were run with a back
pressure regulator at 2.5 atm. The effect of phenol partitioning into the residual
Gelant injectivity was measured at ambient tem- oil phase during injection of gelant into the reservoir
perature in Clashach sandstone cores (4-cm length was modeled using SQUEEZE~V (Sorbie, 1990; Sorbie
× 2.5-cm diameter) of 1 D permeability and in a et al., 1991). It was convenient to transform mathe-
Portland limestone core (10 cm × 2.5 cm, 160 mD) matically the phenol equilibrium with the immobile
using a constant flow rate of 30 ml/h. The gelant oil phase into an equivalent adsorption onto the
200 S.L. Br?,'ant et al. / Journal of Petroleum Science and Engineering 17 (1997) 197-209
reservoir rock. Partitioning of polymer and formal- of ~ 6 during the gelation process, independent of
dehyde into the oil phase and their adsorption onto their initial pH. The pH range over which these gels
reservoir rock were taken to be zero. Physical disper- form was better defined by preparing a number of
sion was simulated by numerical dispersion (1-2 phenol-formaldehyde/PAAm-AMPS gelants cov-
ft2/day). eting the interval from pH 1 to pH 8.5. By means of
frequent monitoring and correction of the pH with
acid or base, it was possible to maintain the gelants
3. Results and discussion to within 0.5 pH unit of the initial value through the
gelation point. Gelation occurred over the entire
3.1. Bulk studies interval examined, indicating that this cross-linking
chemistry is functional over a wide range of pH.
Bulk chemical studies on gelants serve to define The constant-pH experiments also yielded infor-
the range of compositions and the conditions under mation on the influence of pH on the gelation delay.
which gelation occurs. Appropriately designed exper- Polymer cross-linking was slowest at neutral pH and
iments can also yield insights into the factors that accelerated sharply at pH 1. The limited experimen-
control gelation delay, gel stability and the sensitiv- tation at alkaline pH suggests that cross-linking may
ity of the gelation chemistry to interference by also accelerate above pH 8. Most significantly, how-
fluid-fluid and fluid-rock interactions. In this part ever, the gelation delay varied by less than a factor
of the paper we report on experiments carried out in of 2 over the pH range of interest for field applica-
order to characterize the influence of pH on the tions, pH 4.5-8. This means that the inherent rate of
gelation reaction and to identify parameters that in- the polymer cross-linking reaction should be rela-
fluence the rate of gelation at elevated temperature. tively insensitive to acid-base interactions that occur
The reader is directed elsewhere (Albonico et al., within the formation, a most desirable property in a
1995) for studies defining the optimum phenol and gelant. We speculate that the slower gelation kinetics
formaldehyde concentrations for the polymers em- below pH 8 might be responsible for the earlier
ployed in the present studies. Those tests also estab- conclusion (Falk, 1984) that gelation requires an
lished that the phenol-formaldehyde cross-linking alkaline pH (preferably pH 8-9). Rheological mea-
system is relatively insensitive to the salinity of the surements (Albonico et al., 1995) confirmed that the
aqueous medium, and showed that the gel properties final mechanical properties of the gels were insensi-
do not depend strongly on the weight ratio of tive to pH.
phenol:formaldehyde; unless otherwise noted, a 1:1
weight ratio is employed in the studies reported 3.1.2. Gelation delay
below, It has been suggested (Falk, 1984; Moradi-Araghi
et al., 1993) that the gelation time of phenol-formal-
3.1.1. Gelation sensitir'i~ to pH dehyde gelants can be conveniently modified by
Many formation brines are buffered by the forma- varying the phenol and formaldehyde concentrations.
tion rock to a neutral or weakly acidic pH (Seright, We find, instead, that the gelation times of PAAm-
1992). It is reasonable to expect that the pH of a AMPS solutions vary by less than three-fold for
gelant introduced into such a formation will ap- phenol and formaldehyde concentrations from 800 to
proach the pH of the resident formation fluids unless 8000 ppm. The PAAm and PAAm-VP-AMPS
the gelant is strongly buffered or the formation has polymers behaved similarly. Considering the re-
been modified by an appropriate pre-flush. The re- stricted range of cross-linker concentrations that yield
ported requirement of an alkaline pH for gelation gels of good strength and stability for a given poly-
(Falk, 1984) was therefore perceived as a potential mer solution, these results suggest that the possibili-
limitation of phenol-formaldehyde gelants. Surpris- ties of programming the gelation time of phenol-for-
ingly, however, measurements on several phenol- maldehyde/polymer gelants by adjusting the cross-
formaldehyde gelants near the gel point revealed a linker concentration are, in fact, quite limited.
marked tendency of the solutions to evolve to a pH Temperature, on the other hand, influences the
S.L. Bo,ant et al. / Journal of Petroleum Science and Engineering 17 (1997) 197-209 201
gelation time quite strongly. The results for the three gelation delay for the systems in Fig. l correlates
acrylamide polymers plotted in Fig. 1 show that for a with several factors that would be expected to accel-
given composition, the gelation times decrease by a erate the cross-linking reaction: increasing acryl-
factor of 5-20 on passing from 90 ° to 120°C (see amide comonomer content of the polymer, increasing
also Moradi-Araghi et al., 1993). Except for the 5% polymer concentration, and increasing concentration
PAAm solution, cross-linking proceeds extremely of phenol and formaldehyde, although our previous
slowly at 60°C. This observation is consistent with results suggest that the latter factor has only a lim-
earlier indications that the lower useful temperature ited impact.
limit for this cross-linking system is 70-80°C (Falk,
1984; Moradi-Araghi et al., 1993). 3.2. Gelant propagation and injectiL,i~ in porous
Importantly, the results presented in Fig. 1 show media
that the gelation time is strongly dependent on the
type and concentration of polymer employed. The The experimentation on bulk solutions confirmed
low-molecular-weight PAAm compositions gel an that phenol-formaldehyde/polymer compositions
order of magnitude more rapidly than the acrylamide possess many of the attributes required of a gelant
copolymer gelants; gelation of the 5% PAAm solu- for application at high temperature: the gelation de-
tions takes place readily even at 60°C. While the lays are technologically useful and, to a reasonable
gelation delay of a given gelant will largely be degree, programmable to at least 140°C; the cross-
determined by the formation temperature, the marked linking chemistry is relatively insensitive to pH over
difference in the gelation delays obtained with differ- a wide pH range. The movement of a gelant through
ent polymer compositions increases the possibilities porous media, however, introduces a number of
of finding a suitable gelant for a given reservoir complex factors such as fluid-rock interactions,
treatment. Between the fast-gelling, low-molecular- which can progressively modify the chemical com-
weight PAAm and slower-gelling P A A m - V P - position of the gelant by adding ionic species to it,
AMPS compositions, phenol-formal- by changing its pH, or by retaining polymer or
dehyde/polymer gelants can be formulated to satisfy cross-linker. There will also be intimate contact be-
the requirements of most near-wellbore treatments tween the propagating gelant and the crude oil in
from 60° to 140°C and beyond. The variation in matrix rock. Filtration of partially cross-linked poly-
~5o l
", F--.o---PAAm 5% 1
~x - - - ~ - - - P A A m 3°1o I
~
o 125 ', - -,D - Terpolymer 1% I
Terpolymer 2 ~J
4
~ 100~
E
75
50
0.01 0,1 1 10 100 1000
mer from the gelant by the pore throats of the rock experiment using a 1 D Clashach sandstone core and
could affect the injectivity and placement of the a P A A m - A M P S gelant (1% in polymer, 1000 ppm
gelant. It is possible to evaluate the importance of each, phenol and formaldehyde) heated to 50% of its
such phenomena by means of porous media experi- gelation time are shown in Fig. 2. Whereas the
ments coupled with numerical simulation. In this part injection pressure at constant flow was steady for the
of the paper we will report on sandpack and core- uncross-linked gelant, it increased slowly but contin-
floods carried out in order to characterize the propa- uously for the turbid gelant, suggesting a small
gation and injectivity of phenol-formaldehyde degree of filtration. A similar experiment was run
cross-linked polymer gelants. using a P A A m - V P - A M P S gelant (l% in polymer,
4000 ppm each, phenol and formaldehyde) with a
3.2.1. Gelant turbidity 160 mD Portland limestone core. Once again, only a
While carrying out the studies reported in the first very slight rise in the injection pressure with injected
part of this paper, we observed that the phenol-for- volume was observed for solutions that had been
maldehyde gelants quickly became turbid, and had aged either to 25% or to 50% of their gelation time.
turned opaque by their gel point. Surprisingly, the Measurement of the turbidity of the gelant aged to
earlier literature does not refer clearly to this phase 25% of its gelation time indicated less than a 10%
separation process. Experiments on simple phenol- decrease after passage through the limestone core.
formaldehyde solutions showed that the separating Complementary results were obtained in filtration
phase was formed by reaction between the two experiments with Millipore filters. These experi-
cross-linkers. In the absence of polymer, the initial ments showed that a P A A m - A M P S gelant heated at
milky dispersion gave way to a viscous oil that 120°C to beyond 50% of its gelation time in bulk
settled to the bottom of the container and later passed through a 5-1xm filter without evidence of
solidified, presumably yielding a Bakelite-like mate- filtration. Some filtration was observed with a l q x m
rial (Moradi-Araghi et al., 1993). Evidently, the pres- filter, which provides an indication of the dimensions
ence of polymer in the gelant solution inhibits coagu- of the phase-separated material or cross-linked ag-
lation and settling of the dispersion before gelation gregates that were present in the partially cross-lin-
occurs. It remains unclear whether the condensed
phenol-formaldehyde oil phase plays a role in poly-
mer cross-linking or whether it is an inert side
product. 1 Gelant aged at 120°C Ibr
ked solution. Together these experiments suggest flow t ' = dXfD/dto, where x~ is the dimensionless
that separation of the phenol-formaldehyde reaction position of the phenol concentration front. Since K
products by filtration in formation rock is not likely and Sor are constant, v is also constant, and the time
to be important, at least up to a significant fraction of at which the phenol front arrives at some position x o
the gelation time in bulk. in the porous medium is simply:
arrival XD
. = -- (3)
3.2.2. Cross-linker propagation U
Co E
K- (1) • J
Cw m i
1.15 PV agrees well with the value of 1.13 PV crude oil ( K = 2.2) at a residual saturation of So~ =
calculated from Eqs. (2) and (3). The rapid rise of 0.3. Suppose that a gelant volume of 320 bbl (51 m 3)
the phenol curve to the injected level, once it began is injected. This is sufficient to fill the reservoir
to exit from the core, indicates that the partitioning aqueous phase volume around the well to a radius of
of phenol into the light crude oil was an equilibrium 20 ft. A numerical simulation of this treatment is
process. Phenol was recovered quantitatively during shown in Fig. 4. Dispersion has spread the concen-
a brine post-flush, establishing the complete re- tration fronts substantially during their advance into
versibility of the phenomenon. the reservoir. The phenol front lags those of formal-
dehyde and polymer, due to absorption of phenol
3.2.2.2. The influence of dispersion. Dispersion into the residual crude oil.
causes significant dilution at the leading edge of the Suppose now that gelation occurs only where the
injected gelant. Qualitatively this has the same effect gelant contains at least 80% of the injected concen-
as phenol partitioning: the volume of gel formed will trations of the polymer and of each of the two
be less than the volume of gelant injected. Moreover, cross-linkers. This threshold concentration for
an absorption process governed by a linear relation- formaldehyde and the polymer is at a distance of
ship such as Eq. (1) will be dispersive in the same 16.5 ft from the wellbore (Fig. 4). Thus the 35% of
way that a tracer concentration front is dispersive. the injected gelant volume that lies beyond this
Thus dispersion will spread the slower-moving phe- distance would not form gel, even in the absence of
nol front as well as the formaldehyde and polymer phenol partitioning. In this example phenol falls
fronts, exacerbating the problem of phenol retention below the 80% threshold concentration just 8.3 ft
discussed above. from the wellbore. The gelant solution lying within
To illustrate the magnitude of these effects, con- this distance of the wellbore, i.e., that which would
sider a treatment for a target interval that is 10 ft go on to gel during shut-in, is only 17% of the
thick, has 20% porosity and contains a medium volume injected.
100000
1 0 0 0 0 ~- - . _
- - ~ Polymer
0 \
I- " - Phenol
0
= ~
"13
[ Formaldehyde
l ",-,,! \
10 - - , '
' r , " ........ .... I ~
0 5 !0 15 20 25
3.2.3. Methods to improve propagation phenol would conveniently achieve the target con-
Dispersion and phenol partitioning obviously re- centration in the gelant at the desired depth in the
sult in wasted chemical. A more serious problem is formation. The polymer, formaldehyde and phenol
that gel will not form to the desired distance in the profiles for the two alternatives are shown in Fig. 5.
formation. One way to avoid the latter problem is Given the low cost of phenol, both alternatives
simply to inject more gelant, While over-sizing the appear to be economically sensible ways of achiev-
treatment addresses the issue of depth of penetration, ing a four-fold increase in the volume of gel formed.
it increases the economic penalty of production From Fig. 5, however, we see that both of these
downtime. Over-sizing also complicates the job de- approaches will generate high phenol concentrations
sign. A greater volume of viscous fluid must be in the gelant: with the pre-flush the phenol concen-
pushed into the formation within the time window tration will be as high as 5000 ppm, while in the
imposed by the gelation time, other case the gelant nearest the wellbore will con-
A better solution is to inject a preflush of phenol. tain 2.55% phenol. Experiments were therefore run
Suppose the preflush has the same phenol concentra- in order to determine whether gelation is affected by
tion as the gelant. Then the injected gelant will such high concentrations of phenol. FW and SSW
encounter oil with a phenol concentration already in solutions of PAAm (3%) and formaldehyde (1000
equilibrium with the gelant composition. Thus no ppm) gelled at 120°C in the presence of up to, but
stripping of phenol from the gelant will occur, and not beyond, 10,000 ppm of phenol. This suggests
the gelant components will propagate together until that the formaldehyde is tied up unproductively by
they reach the penetration depth of the preflush. phenol at high phenol concentrations, which is rea-
Since the preflush will have water-like viscosity and sonable in light of the hypotheses that have been
cannot form gel on its own, this option is opera- offered on the cross-linking mechanism (Moradi-
tionally simpler than the option of oversizing the Araghi et al., 1993). Other experiments showed that
gelant volume. It also means that much more of the the maximum amount of phenol for which gelation
injected gelant will form gel, reducing wasted chemi- occurs rises with increasing formaldehyde concentra-
cals. tion.
Increasing the phenol concentration in the gelant These results rule out the use of a high phenol
is also a viable means of pushing the desired phenol concentration in the gelant to compensate for the loss
concentration deeper into the reservoir. This is not of phenol into the immobile oil phase. The use of a
because a front of high phenol concentration travels phenol pre-flush, on the other hand, appears to be a
more rapidly than a front of low concentration; Eq. viable means for maximizing the conversion of the
(2) shows that the velocity is independent of concen- gelant to gel within the formation.
tration. Rather, a higher injected concentration im- This analysis assumes that the viscous (of order
poses a higher concentration gradient and thus in- 10-100 cP) gelant does not mobilise any residual
creases the dispersive/diffusive flux. oil. Any oil swept out of the treated zone ahead of
Numerical simulation was again used to evaluate the gelant would no longer act as a phenol sink, thus
these alternatives. The target was to achieve 800 increasing phenol penetration. The possibility of mo-
ppm phenol in the gelant at a distance of 16.5 fl from bilisation depends upon the producing conditions for
the wellbore (this corresponds to the concentration the well and the treatment design, so it is not consid-
that phenol would have had at this distance if there ered further here.
were no partitioning). For the increased-concentra-
tion alternative, this requirement is met with a phe- 3.2.4. Gelant injecticiO'
nol concentration of 2.55% in the gelant. The possi- The gelation reaction changes an ordinary viscous
ble solutions for the preflush alternative are defined solution into an elastic solid which cannot be pumped
by two variables, the concentration and volume of through porous media at practical pressures. The
the phenol pre-flush. Assuming that a short pre-flush evolution of the effective viscosity of the gelant
and moderate phenol concentration would be pre- during this reaction imposes obvious constraints on
ferred in the field, we find that a 2.5-m 3 slug 10% in the placement of a treatment. To assess its injectiv-
206 S.L. B ~ a n t et al. / Journal of Petroleum Science and Engineering 17 ( 1997) 197-209
ity, a gelant 1000 ppm each in phenol and formal- this point, gelant in the last quarter of the slim-tube
dehyde and 3% in PAAm (MW 5 X 105) was pumped had been in the sand pack for at least 11 h, the
continuously but at different flow rates through a gelation time in bulk. The steady increase in the
3-m slim-tube packed with quartz sand and contain- apparent viscosity during this stage of the experi-
ing 20% of residual oil at 120°C. At the initial ment reflects the increasing volume of gelant in the
injection rate (30 m l / h ) the residence time of the slim-tube approaching the gel point.
gelant in the sand pack (1.6 h) was much less than A second slim-tube experiment was carried out in
the gelation time measured in bulk at this tempera- order to confirm that gel forms at the leading front of
ture (11 h). As shown in Fig. 6 the apparent viscosity the gelant. This was accomplished using a 2%
increased as the gelant displaced the brine from the PAAm-VP-AMPS SSW gelant (containing 4000
slim-tube, reaching a steady value of just under 20 ppm each of phenol and formaldehyde) with a bulk
cP. (The apparent viscosity is calculated directly gelation time of 17 h at 140°C. The flow rate was
from Darcy's law and the observed pressure drop, chosen so that after 17 h of injection the gelant
ignoring the presence of the oil phase. It thus in- would have arrived approximately half-way through
cludes the effect of any permeability reduction due the 6.6-m brine-saturated slim-tube. The experiment
to filtration or blockage.) Reducing the flow rate to was terminated after 9.5 h of injection when the
12 m l / h and then to 6 m l / h increased the residence injection pressure had reached 150 bar. At this point,
time to 37% and 75% of the gelation time, respec- the calculated gelant front (neglecting dispersion)
tively, yet the gelant remained pumpable, with no had penetrated to a depth of 2.33 m. The 10-cm
significant change in the apparent viscosity. How- lengths of the slim-tube (see Section 2.4) between
ever, when the injection rate was reduced to 1.2 the inlet and 2.3 m from the inlet resisted flow at
ml/h, the residence time of the gelant in the sand pressure drops up to 200 bar. Gelled sand was
pack was much longer than the gelation time. After extruded from the sections between 2.3 and 2.4 in
6-h injection at this rate the pressure limit of the from the inlet at 55 bar. Free-flowing sand was
equipment was reached and injection was halted• At extruded from the sections beyond 2.4 m at low
100000 ~-
E
Q.
t-
10000
1
t Phenol (prefiush then . . "
normal gelant) n"
"I --
.O
i
1000 ~ " "
¢
0
¢
0 ",', Formaldehyde
\
100 q
-,-Q \
,i! :, -,-..-~
5 10 15 20 25
too T
i
9o
l
go
30 12 mL/h 6 mL/h 12 mL/h
A
e~ ' mL/t 41h 82h 41 h
o 70 16 t
60
0
U
"g 50.-
40.
0 s ~0 ~s 2o 25 30 3s 40
Time (hours)
Fig. 6. Apparent viscosity (as defined in text) of phenol-formaldehyde/PAAm gelant injected continuouslyinto a 3-m sandpack at 120°C at
decreasing flow rates. The labels for each section give the flow rate and correspondingresidence times. The apparent viscosity of the gelant
does not increase from its initial value until the residence time of the gelant in the sandpack exceeds the gelation time, which is 11 h for this
formulation.
pressure. Thus the maximum depth of effective gel formations. Cross-linking occurs over a broad pH
formation corresponds to the calculated depth of range, and the gelation delay is relatively insensitive
gelant penetration. This desirable feature of the phe- to this parameter. In contrast, gelation delay depends
nol-formaldehyde system is not restricted to sand- strongly upon temperature and the nature and con-
packs. Gelant injected into a Portland limestone core centration of the polymer used. By selecting between
formed gel throughout the core length. The insensi- different polymer compositions, reasonable flexibil-
tivity to lithology is a significant advantage, as it ity in the gelation time can be achieved over the
simplifies treatment design and reduces the chances temperature range from 60 ° to at least 140°C. The
that unknown or unexpected variations in formation turbidity that results from reaction between the phe-
mineralogy could compromise the treatment. nol and formaldehyde cross-linkers has limited or
The above slim-tube experiment also demon- negligible effects on gelation, gelant propagation or
strates how effectively these gels reduce permeabil- gelant injectivity. Indeed, the gelants can be injected
ity. Typical post-gelation permeabilities are tens of through 3-m sand packs at 140°C even for residence
microdarcies, regardless of the lithology or the initial times approaching the gelation time. Porous media
permeability of the medium. These values are com- studies show that the propagation and the gel perfor-
parable to those found for other strong gels (Seright, mance of this system are not sensitive to lithology.
1993). Hence the phenol-formaldehyde system The solubility of phenol in crude oil is the primary
should have the same capacity for stopping fluid disadvantage of this system. Simulations show that
flow in high-temperature formations that other sys- the simplest method to overcome this problem is to
tems offer in less harsh environments. inject a slug of phenol before the gelant.
4. Conclusions 5. Notation
Cw = concentration of phenol in brine (mol/1) Pet. Eng. Int. Symp. on Oilfield Chemistry, San Antonio, TX,
K = phenol partitioning coefficient between Feb. 14 17, Pap. SPE 28983.
Bartosek, M., Mennella, A., Lockhart, T.P., Causin, E., Rossi, E.
brine and crude oil (dimensionless) and Passucci, C., 1994. Polymer gels for conformance treat-
Sot = residual oil saturation (dimensionless) ments: propagation of Cr(III) cross-linking complexes in
r = velocity of phenol concentration front porous media. Presented at the Soc. Pet. Eng./U.S. Dep.
relative to velocity of inert aqueous tracer Energy 9th Syrup. on Improved Oil Recovery, Tulsa, OK,
t,~ = velocity of inert aqueous tracer through a Apr. 17-20. Pap. SPE/DOE 27828.
Borling, D.C., 1994. Injection conformance control case histories
porous medium
using gels at the Wertz Field CO 2 tertiary flood in Wyoming.
xD = dimensionless distance within porous Presented at Soc. Pet. Eng./U.S. Dep. Energy 9th Symp. on
medium Improved Oil Recovery, Tulsa, OK, Apr. 17-20, Pap.
x~ = dimensionless position of phenol concen- SPE/DOE 27825.
tration front within porous medium Bryant, S.L., Rabaioli, M.R and Lockhart, T.P., 1996. Influence
of syneresis on permeability reduction by polymer gels. SPE
tD = dimensionless time (pore volumes in-
Production and Facilities, (Nov. 1996), pp. 209-215.
jected) Falk, D.O., 1984. Process for selectively plugging permeable
t~rri,,~
D = dimensionless time at which phenol con- zones in a subterranean formation. U.S. Patent 4,485,875
centration front reaches given position (Dec. 4, 1984).
within a porous medium Hauser. T.R. and Cummings, R.L., 1964. Increasing sensitivity of
3-methyl-2-benzothiazolone hydrazone test for analysis of
aliphatic aldehydes in air. Anal. Chem., 36: 179.
Lockha~, T.P. and Albonico, P., 1992. A new gelation technology
Acknowledgements for in-depth placement of Cr + 3/polymer gels in high-temper-
ature reservoirs. Presented at Soc. Pet. Eng./U.S. Dep. Energy
8th Symp. on Enhanced Oil Recovery, Tulsa, OK, Apr. 22-24,
It is a pleasure to acknowledge Giuseppe Calloni Pap. SPE/DOE 24194.
for his contributions to several of the experiments Moradi-Araghi, A., Beardmore, D.H. and Stahl, G.A., 1988. The
application of gels in enhanced oil recovery: theory, polymers
described in this paper, Alberto Malandrino for the
and cross-linker systems. In: G.A. Stahl and D.N. Schulz
numerical simulations and Paola Albonico for the (Editors), Water-Soluble Polymers for Petroleum Recovery.
bulk gelation experiments. This research was sup- Plenum, New York, NY, pp. 299-312.
ported by Agip SpA. Parts of this paper were pre- Moradi-Araghi, A., Bjornson, G. and Doe, P.H., 1993. Thermally
sented at the 1995 SPE Oilfield Chemistry Sympo- stable gels for near-wellbore permeability contrast corrections.
Soc. Pet. Eng. Adv. Technol. Ser., 1: 140-145.
sium, San Antonio, Texas, as SPE 28983.
Sanders. G., Chambers, M. and Lane, R., 1994. Successful gas
shutoff with polymer gel using temperature modeling and
selective placement in the Prudhoe Bay field. Presented at
69th Soc. Pet. Eng. Annu. Tech. Conf. Exhibit., New Orleans,
References LA, Sept. 25-28, Pap. SPE 28502.
Seright, R.S., 1988. Placement of gels to modify injection profiles.
Albonico, P. and Lockhart, T.P., 1993. Divalent ion-resistant Presented at Soc. Pet. Eng./U.S. Dep. Energy 6th Enhanced
polymer gets for high temperature applications: syneresis in- Oil Recovery Syrup., Tulsa, OK, Apr. 17-20, Pap. SPE/DOE
hibiting additives. Presented at Soc. Pet. Eng. Int. Symp. on 17332.
Oilfield Chemistry, New Orleans, LA, Mar. 2-5, Pap. SPE Seright, R.S., 1992. Impact of permeability and lithology on gel
25220. performance. Presented at Soc. Pet. Eng./U.S. Dep. Energy
Albonico, P., Burrafato, G., Di Lullo, A. and Lockhart, T.P., 8th Symp. on Enhanced Oil Recovery, Tulsa, OK, Apr. 22-24,
1993. Effective gelation-delaying additives for Cr+3/polymer SPE/DOE 24190.
gels. Presented at Soc. Pet. Eng. Int. Symp. on Oilfield Seright, R.S., 1993. Effect of rock permeability on gel perfor-
Chemistry, New Orleans, LA, Mar. 2-5, Pap. SPE 25221. mance in fluid-diversion applications. In Situ, 17: 363-386.
Albonico, P., Bartosek, M., Lockhart, T.P,, Causin, E. and Rossi, Seright, R.S. and Martin, F.D., 1991. Impact of gelation pH, rock
E., 1994. New polymer gels for reducing water production in permeability, and lithology on the performance of a
high temperature reservoirs. Presented at Eur. Prod. Oper. monomer-based gel. Presented at Soc. Pet. Eng Int. Syrup. on
Conf. Exhibit., Aberdeen, Mar. 15-17, Pap. SPE 27609. Oilfield Chemistry, Anaheim, CA, Feb. 20-22, Pap. SPE
Albonico, P., Bartosek, M., Malandrino, A., Bryant, S. and Lock- 20999.
hart, T.P., 1995. Studies on phenol-formaldehyde crosslinked Sorbie. K.S., 1990. The SQUEEZE IV User Manual. Dep. Pet.
polymer gels in bulk and in porous media. Presented at Soc. Eng., Heriot-Watt Univ., Edinburgh.
S.L. Bryant et al. / Journal of Petroleum Science and Engineering 17 (1997) 197-209 209
Sorbie, K,S., Yuan, M.D., Todd, A.C. and Wat, R.M.S., 1991. carboxylate gels for near wellbore matrix treatments. Soc. Pet.
The modelling and design of scale inhibitor squeeze treat- Eng. Adv. Technol. Set,, 1: 146-152.
ments in complex reservoirs, Presented at Soc. Pet. Eng. Int. Sydansk, R.D. and Moore, P., 1992. Gel conformance treatments
Symp. on Oilfield Chemistry, Anaheim, CA, Feb. 20-22, Pap. increase oil production in Wyoming. Oil Gas J. (Jan.), 20:
SPE 21024. 40-45.
Sydansk, R.D., 1993. Acrylamide-polymer/chromium(llI)-